Contents

Sysman API

Contents

Sysman API

oneAPI Level Zero Specification - Version 1.8.0

Common

Common Enums

zes_structure_type_t

enum zes_structure_type_t

Defines structure types.

Values:

enumerator ZES_STRUCTURE_TYPE_DEVICE_PROPERTIES

zes_device_properties_t

enumerator ZES_STRUCTURE_TYPE_PCI_PROPERTIES

zes_pci_properties_t

enumerator ZES_STRUCTURE_TYPE_PCI_BAR_PROPERTIES

zes_pci_bar_properties_t

enumerator ZES_STRUCTURE_TYPE_DIAG_PROPERTIES

zes_diag_properties_t

enumerator ZES_STRUCTURE_TYPE_ENGINE_PROPERTIES

zes_engine_properties_t

enumerator ZES_STRUCTURE_TYPE_FABRIC_PORT_PROPERTIES

zes_fabric_port_properties_t

enumerator ZES_STRUCTURE_TYPE_FAN_PROPERTIES

zes_fan_properties_t

enumerator ZES_STRUCTURE_TYPE_FIRMWARE_PROPERTIES

zes_firmware_properties_t

enumerator ZES_STRUCTURE_TYPE_FREQ_PROPERTIES

zes_freq_properties_t

enumerator ZES_STRUCTURE_TYPE_LED_PROPERTIES

zes_led_properties_t

enumerator ZES_STRUCTURE_TYPE_MEM_PROPERTIES

zes_mem_properties_t

enumerator ZES_STRUCTURE_TYPE_PERF_PROPERTIES

zes_perf_properties_t

enumerator ZES_STRUCTURE_TYPE_POWER_PROPERTIES

zes_power_properties_t

enumerator ZES_STRUCTURE_TYPE_PSU_PROPERTIES

zes_psu_properties_t

enumerator ZES_STRUCTURE_TYPE_RAS_PROPERTIES

zes_ras_properties_t

enumerator ZES_STRUCTURE_TYPE_SCHED_PROPERTIES

zes_sched_properties_t

enumerator ZES_STRUCTURE_TYPE_SCHED_TIMEOUT_PROPERTIES

zes_sched_timeout_properties_t

enumerator ZES_STRUCTURE_TYPE_SCHED_TIMESLICE_PROPERTIES

zes_sched_timeslice_properties_t

enumerator ZES_STRUCTURE_TYPE_STANDBY_PROPERTIES

zes_standby_properties_t

enumerator ZES_STRUCTURE_TYPE_TEMP_PROPERTIES

zes_temp_properties_t

enumerator ZES_STRUCTURE_TYPE_DEVICE_STATE

zes_device_state_t

enumerator ZES_STRUCTURE_TYPE_PROCESS_STATE

zes_process_state_t

enumerator ZES_STRUCTURE_TYPE_PCI_STATE

zes_pci_state_t

enumerator ZES_STRUCTURE_TYPE_FABRIC_PORT_CONFIG

zes_fabric_port_config_t

enumerator ZES_STRUCTURE_TYPE_FABRIC_PORT_STATE

zes_fabric_port_state_t

enumerator ZES_STRUCTURE_TYPE_FAN_CONFIG

zes_fan_config_t

enumerator ZES_STRUCTURE_TYPE_FREQ_STATE

zes_freq_state_t

enumerator ZES_STRUCTURE_TYPE_OC_CAPABILITIES

zes_oc_capabilities_t

enumerator ZES_STRUCTURE_TYPE_LED_STATE

zes_led_state_t

enumerator ZES_STRUCTURE_TYPE_MEM_STATE

zes_mem_state_t

enumerator ZES_STRUCTURE_TYPE_PSU_STATE

zes_psu_state_t

enumerator ZES_STRUCTURE_TYPE_BASE_STATE

zes_base_state_t

enumerator ZES_STRUCTURE_TYPE_RAS_CONFIG

zes_ras_config_t

enumerator ZES_STRUCTURE_TYPE_RAS_STATE

zes_ras_state_t

enumerator ZES_STRUCTURE_TYPE_TEMP_CONFIG

zes_temp_config_t

enumerator ZES_STRUCTURE_TYPE_PCI_BAR_PROPERTIES_1_2

zes_pci_bar_properties_1_2_t

enumerator ZES_STRUCTURE_TYPE_DEVICE_ECC_DESC

zes_device_ecc_desc_t

enumerator ZES_STRUCTURE_TYPE_DEVICE_ECC_PROPERTIES

zes_device_ecc_properties_t

enumerator ZES_STRUCTURE_TYPE_POWER_LIMIT_EXT_DESC

zes_power_limit_ext_desc_t

enumerator ZES_STRUCTURE_TYPE_POWER_EXT_PROPERTIES

zes_power_ext_properties_t

enumerator ZES_STRUCTURE_TYPE_OVERCLOCK_PROPERTIES

zes_overclock_properties_t

enumerator ZES_STRUCTURE_TYPE_FABRIC_PORT_ERROR_COUNTERS

zes_fabric_port_error_counters_t

enumerator ZES_STRUCTURE_TYPE_ENGINE_EXT_PROPERTIES

zes_engine_ext_properties_t

enumerator ZES_STRUCTURE_TYPE_RESET_PROPERTIES

zes_reset_properties_t

enumerator ZES_STRUCTURE_TYPE_DEVICE_EXT_PROPERTIES

zes_device_ext_properties_t

enumerator ZES_STRUCTURE_TYPE_DEVICE_UUID

zes_uuid_t

enumerator ZES_STRUCTURE_TYPE_POWER_DOMAIN_EXP_PROPERTIES

zes_power_domain_exp_properties_t

enumerator ZES_STRUCTURE_TYPE_MEM_TIMESTAMP_BITS_EXP

zes_mem_timestamp_bits_exp_t

enumerator ZES_STRUCTURE_TYPE_MEMORY_PAGE_OFFLINE_STATE_EXP

zes_mem_page_offline_state_exp_t

enumerator ZES_STRUCTURE_TYPE_FORCE_UINT32

Common Structures

zes_base_properties_t

struct zes_base_properties_t

Base for all properties types.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_base_desc_t

struct zes_base_desc_t

Base for all descriptor types.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_base_state_t

struct zes_base_state_t

Base for all state types.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_base_config_t

struct zes_base_config_t

Base for all config types.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_base_capability_t

struct zes_base_capability_t

Base for all capability types.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

Driver

Driver Functions

zesInit

ZE_APIEXPORT ze_result_t ZE_APICALL zesInit(zes_init_flags_t flags)

Initialize ‘oneAPI’ System Resource Management (sysman)

  • The application must call this function or zeInit with the ::ZES_ENABLE_SYSMAN environment variable set before calling any other sysman function.

  • If this function is not called then all other sysman functions will return ZE_RESULT_ERROR_UNINITIALIZED.

  • This function will only initialize sysman. To initialize other functions, call zeInit.

  • There is no requirement to call this function before or after zeInit.

  • Only one instance of sysman will be initialized per process.

  • The application must call this function after forking new processes. Each forked process must call this function.

  • The application may call this function from simultaneous threads.

  • The implementation of this function must be thread-safe for scenarios where multiple libraries may initialize sysman simultaneously.

Parameters

flags – [in] initialization flags. currently unused, must be 0 (default).

Returns

zesDriverGet

ZE_APIEXPORT ze_result_t ZE_APICALL zesDriverGet(uint32_t *pCount, zes_driver_handle_t *phDrivers)

Retrieves sysman driver instances.

  • A sysman driver represents a collection of physical devices.

  • Multiple calls to this function will return identical sysman driver handles, in the same order.

  • The application may pass nullptr for pDrivers when only querying the number of sysman drivers.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • pCount – [in,out] pointer to the number of sysman driver instances. if count is zero, then the loader shall update the value with the total number of sysman drivers available. if count is greater than the number of sysman drivers available, then the loader shall update the value with the correct number of sysman drivers available.

  • phDrivers – [in,out][optional][range(0, *pCount)] array of sysman driver instance handles. if count is less than the number of sysman drivers available, then the loader shall only retrieve that number of sysman drivers.

Returns

zesDriverGetExtensionProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesDriverGetExtensionProperties(zes_driver_handle_t hDriver, uint32_t *pCount, zes_driver_extension_properties_t *pExtensionProperties)

Retrieves extension properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDriver – [in] handle of the driver instance

  • pCount – [in,out] pointer to the number of extension properties. if count is zero, then the driver shall update the value with the total number of extension properties available. if count is greater than the number of extension properties available, then the driver shall update the value with the correct number of extension properties available.

  • pExtensionProperties – [in,out][optional][range(0, *pCount)] array of query results for extension properties. if count is less than the number of extension properties available, then driver shall only retrieve that number of extension properties.

Returns

zesDriverGetExtensionFunctionAddress

ZE_APIEXPORT ze_result_t ZE_APICALL zesDriverGetExtensionFunctionAddress(zes_driver_handle_t hDriver, const char *name, void **ppFunctionAddress)

Retrieves function pointer for vendor-specific or experimental extensions.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDriver – [in] handle of the driver instance

  • name – [in] extension function name

  • ppFunctionAddress – [out] pointer to function pointer

Returns

Driver Enums

zes_init_flags_t

enum zes_init_flag_t

Values:

enumerator ZES_INIT_FLAG_PLACEHOLDER

placeholder for future use

enumerator ZES_INIT_FLAG_FORCE_UINT32

Driver Structures

zes_driver_extension_properties_t

struct zes_driver_extension_properties_t

Extension properties queried using zesDriverGetExtensionProperties.

Public Members

char name[ZES_MAX_EXTENSION_NAME]

[out] extension name

uint32_t version

[out] extension version using ZE_MAKE_VERSION

Device

Device Functions

zesDeviceGet

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceGet(zes_driver_handle_t hDriver, uint32_t *pCount, zes_device_handle_t *phDevices)

Retrieves sysman devices within a sysman driver.

  • Multiple calls to this function will return identical sysman device handles, in the same order.

  • The number and order of handles returned from this function is NOT affected by the ::ZE_AFFINITY_MASK, ::ZE_ENABLE_PCI_ID_DEVICE_ORDER, or ::ZE_FLAT_DEVICE_HIERARCHY environment variables.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDriver – [in] handle of the sysman driver instance

  • pCount – [in,out] pointer to the number of sysman devices. if count is zero, then the driver shall update the value with the total number of sysman devices available. if count is greater than the number of sysman devices available, then the driver shall update the value with the correct number of sysman devices available.

  • phDevices – [in,out][optional][range(0, *pCount)] array of handle of sysman devices. if count is less than the number of sysman devices available, then driver shall only retrieve that number of sysman devices.

Returns

zesDeviceGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceGetProperties(zes_device_handle_t hDevice, zes_device_properties_t *pProperties)

Get properties about the device.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pProperties – [in,out] Structure that will contain information about the device.

Returns

zesDeviceGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceGetState(zes_device_handle_t hDevice, zes_device_state_t *pState)

Get information about the state of the device - if a reset is required, reasons for the reset and if the device has been repaired.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pState – [in,out] Structure that will contain information about the device.

Returns

zesDeviceReset

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceReset(zes_device_handle_t hDevice, ze_bool_t force)

Reset device.

  • Performs a PCI bus reset of the device. This will result in all current device state being lost.

  • All applications using the device should be stopped before calling this function.

  • If the force argument is specified, all applications using the device will be forcibly killed.

  • The function will block until the device has restarted or an implementation defined timeout occurred waiting for the reset to complete.

Parameters
  • hDevice – [in] Sysman handle for the device

  • force – [in] If set to true, all applications that are currently using the device will be forcibly killed.

Returns

zesDeviceResetExt

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceResetExt(zes_device_handle_t hDevice, zes_reset_properties_t *pProperties)

Reset device extension.

  • Performs a PCI bus reset of the device. This will result in all current device state being lost.

  • Prior to calling this function, user is responsible for closing applications using the device unless force argument is specified.

  • If the force argument is specified, all applications using the device will be forcibly killed.

  • The function will block until the device has restarted or a implementation specific timeout occurred waiting for the reset to complete.

Parameters
  • hDevice – [in] Sysman handle for the device

  • pProperties – [in] Device reset properties to apply

Returns

zesDeviceProcessesGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceProcessesGetState(zes_device_handle_t hDevice, uint32_t *pCount, zes_process_state_t *pProcesses)

Get information about host processes using the device.

  • The number of processes connected to the device is dynamic. This means that between a call to determine the value of pCount and the subsequent call, the number of processes may have increased or decreased. It is recommended that a large array be passed in so as to avoid receiving the error ZE_RESULT_ERROR_INVALID_SIZE. Also, always check the returned value in pCount since it may be less than the earlier call to get the required array size.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle for the device

  • pCount – [in,out] pointer to the number of processes. if count is zero, then the driver shall update the value with the total number of processes currently attached to the device. if count is greater than the number of processes currently attached to the device, then the driver shall update the value with the correct number of processes.

  • pProcesses – [in,out][optional][range(0, *pCount)] array of process information. if count is less than the number of processes currently attached to the device, then the driver shall only retrieve information about that number of processes. In this case, the return code will ZE_RESULT_ERROR_INVALID_SIZE.

Returns

zesDevicePciGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesDevicePciGetProperties(zes_device_handle_t hDevice, zes_pci_properties_t *pProperties)

Get PCI properties - address, max speed.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pProperties – [in,out] Will contain the PCI properties.

Returns

zesDevicePciGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesDevicePciGetState(zes_device_handle_t hDevice, zes_pci_state_t *pState)

Get current PCI state - current speed.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pState – [in,out] Will contain the PCI properties.

Returns

zesDevicePciGetBars

ZE_APIEXPORT ze_result_t ZE_APICALL zesDevicePciGetBars(zes_device_handle_t hDevice, uint32_t *pCount, zes_pci_bar_properties_t *pProperties)

Get information about each configured bar.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of PCI bars. if count is zero, then the driver shall update the value with the total number of PCI bars that are setup. if count is greater than the number of PCI bars that are setup, then the driver shall update the value with the correct number of PCI bars.

  • pProperties – [in,out][optional][range(0, *pCount)] array of information about setup PCI bars. if count is less than the number of PCI bars that are setup, then the driver shall only retrieve information about that number of PCI bars.

Returns

zesDevicePciGetStats

ZE_APIEXPORT ze_result_t ZE_APICALL zesDevicePciGetStats(zes_device_handle_t hDevice, zes_pci_stats_t *pStats)

Get PCI stats - bandwidth, number of packets, number of replays.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pStats – [in,out] Will contain a snapshot of the latest stats.

Returns

Device Enums

zes_engine_type_flags_t

enum zes_engine_type_flag_t

Values:

enumerator ZES_ENGINE_TYPE_FLAG_OTHER

Undefined types of accelerators.

enumerator ZES_ENGINE_TYPE_FLAG_COMPUTE

Engines that process compute kernels only (no 3D content).

enumerator ZES_ENGINE_TYPE_FLAG_3D

Engines that process 3D content only (no compute kernels).

enumerator ZES_ENGINE_TYPE_FLAG_MEDIA

Engines that process media workloads.

enumerator ZES_ENGINE_TYPE_FLAG_DMA

Engines that copy blocks of data.

enumerator ZES_ENGINE_TYPE_FLAG_RENDER

Engines that can process both 3D content and compute kernels.

enumerator ZES_ENGINE_TYPE_FLAG_FORCE_UINT32

zes_repair_status_t

enum zes_repair_status_t

Device repair status.

Values:

enumerator ZES_REPAIR_STATUS_UNSUPPORTED

The device does not support in-field repairs.

enumerator ZES_REPAIR_STATUS_NOT_PERFORMED

The device has never been repaired.

enumerator ZES_REPAIR_STATUS_PERFORMED

The device has been repaired.

enumerator ZES_REPAIR_STATUS_FORCE_UINT32

zes_reset_reason_flags_t

enum zes_reset_reason_flag_t

Values:

enumerator ZES_RESET_REASON_FLAG_WEDGED

The device needs to be reset because one or more parts of the hardware is wedged

enumerator ZES_RESET_REASON_FLAG_REPAIR

The device needs to be reset in order to complete in-field repairs.

enumerator ZES_RESET_REASON_FLAG_FORCE_UINT32

zes_reset_type_t

enum zes_reset_type_t

Device reset type.

Values:

enumerator ZES_RESET_TYPE_WARM

Apply warm reset.

enumerator ZES_RESET_TYPE_COLD

Apply cold reset.

enumerator ZES_RESET_TYPE_FLR

Apply FLR reset.

enumerator ZES_RESET_TYPE_FORCE_UINT32

zes_device_type_t

enum zes_device_type_t

Supported device types.

Values:

enumerator ZES_DEVICE_TYPE_GPU

Graphics Processing Unit.

enumerator ZES_DEVICE_TYPE_CPU

Central Processing Unit.

enumerator ZES_DEVICE_TYPE_FPGA

Field Programmable Gate Array.

enumerator ZES_DEVICE_TYPE_MCA

Memory Copy Accelerator.

enumerator ZES_DEVICE_TYPE_VPU

Vision Processing Unit.

enumerator ZES_DEVICE_TYPE_FORCE_UINT32

zes_device_property_flags_t

enum zes_device_property_flag_t

Values:

enumerator ZES_DEVICE_PROPERTY_FLAG_INTEGRATED

Device is integrated with the Host.

enumerator ZES_DEVICE_PROPERTY_FLAG_SUBDEVICE

Device handle used for query represents a sub-device.

enumerator ZES_DEVICE_PROPERTY_FLAG_ECC

Device supports error correction memory access.

enumerator ZES_DEVICE_PROPERTY_FLAG_ONDEMANDPAGING

Device supports on-demand page-faulting.

enumerator ZES_DEVICE_PROPERTY_FLAG_FORCE_UINT32

zes_pci_bar_type_t

enum zes_pci_bar_type_t

PCI bar types.

Values:

enumerator ZES_PCI_BAR_TYPE_MMIO

MMIO registers.

enumerator ZES_PCI_BAR_TYPE_ROM

ROM aperture.

enumerator ZES_PCI_BAR_TYPE_MEM

Device memory.

enumerator ZES_PCI_BAR_TYPE_FORCE_UINT32

Device Structures

zes_device_state_t

struct zes_device_state_t

Device state.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_reset_reason_flags_t reset

[out] Indicates if the device needs to be reset and for what reasons. returns 0 (none) or combination of zes_reset_reason_flag_t

zes_repair_status_t repaired

[out] Indicates if the device has been repaired

zes_reset_properties_t

struct zes_reset_properties_t

Device reset properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t force

[in] If set to true, all applications that are currently using the device will be forcibly killed.

zes_reset_type_t resetType

[in] Type of reset needs to be performed

zes_uuid_t

struct zes_uuid_t

Device universal unique id (UUID)

Public Members

uint8_t id[ZES_MAX_UUID_SIZE]

[out] opaque data representing a device UUID

zes_device_properties_t

struct zes_device_properties_t

Device properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_device_properties_t core

[out] (Deprecated, use zes_uuid_t in the extended structure) Core device properties

uint32_t numSubdevices

[out] Number of sub-devices. A value of 0 indicates that this device doesn’t have sub-devices.

char serialNumber[ZES_STRING_PROPERTY_SIZE]

[out] Manufacturing serial number (NULL terminated string value). This value is intended to reflect the Part ID/SoC ID assigned by manufacturer that is unique for a SoC. Will be set to the string “unknown” if this cannot be determined for the device.

char boardNumber[ZES_STRING_PROPERTY_SIZE]

[out] Manufacturing board number (NULL terminated string value). Alternatively “boardSerialNumber”, this value is intended to reflect the string printed on board label by manufacturer. Will be set to the string “unknown” if this cannot be determined for the device.

char brandName[ZES_STRING_PROPERTY_SIZE]

[out] Brand name of the device (NULL terminated string value). Will be set to the string “unknown” if this cannot be determined for the device.

char modelName[ZES_STRING_PROPERTY_SIZE]

[out] Model name of the device (NULL terminated string value). Will be set to the string “unknown” if this cannot be determined for the device.

char vendorName[ZES_STRING_PROPERTY_SIZE]

[out] Vendor name of the device (NULL terminated string value). Will be set to the string “unknown” if this cannot be determined for the device.

char driverVersion[ZES_STRING_PROPERTY_SIZE]

[out] Installed driver version (NULL terminated string value). Will be set to the string “unknown” if this cannot be determined for the device.

zes_device_ext_properties_t

struct zes_device_ext_properties_t

Device properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_uuid_t uuid

[out] universal unique identifier. Note: uuid obtained from Sysman API is the same as from core API. Subdevices will have their own uuid.

zes_device_type_t type

[out] generic device type

zes_device_property_flags_t flags

[out] 0 (none) or a valid combination of zes_device_property_flag_t

zes_process_state_t

struct zes_process_state_t

Contains information about a process that has an open connection with this device.

  • The application can use the process ID to query the OS for the owner and the path to the executable.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint32_t processId

[out] Host OS process ID.

uint64_t memSize

[out] Device memory size in bytes allocated by this process (may not necessarily be resident on the device at the time of reading).

uint64_t sharedSize

[out] The size of shared device memory mapped into this process (may not necessarily be resident on the device at the time of reading).

zes_engine_type_flags_t engines

[out] Bitfield of accelerator engine types being used by this process.

zes_pci_address_t

struct zes_pci_address_t

PCI address.

Public Members

uint32_t domain

[out] BDF domain

uint32_t bus

[out] BDF bus

uint32_t device

[out] BDF device

uint32_t function

[out] BDF function

zes_pci_speed_t

struct zes_pci_speed_t

PCI speed.

Public Members

int32_t gen

[out] The link generation. A value of -1 means that this property is unknown.

int32_t width

[out] The number of lanes. A value of -1 means that this property is unknown.

int64_t maxBandwidth

[out] The maximum bandwidth in bytes/sec (sum of all lanes). A value of -1 means that this property is unknown.

zes_pci_properties_t

struct zes_pci_properties_t

Static PCI properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_pci_address_t address

[out] The BDF address

zes_pci_speed_t maxSpeed

[out] Fastest port configuration supported by the device (sum of all lanes)

ze_bool_t haveBandwidthCounters

[out] Indicates whether the rxCounter and txCounter members of zes_pci_stats_t will have valid values

ze_bool_t havePacketCounters

[out] Indicates whether the packetCounter member of zes_pci_stats_t will have a valid value

ze_bool_t haveReplayCounters

[out] Indicates whether the replayCounter member of zes_pci_stats_t will have a valid value

zes_pci_state_t

struct zes_pci_state_t

Dynamic PCI state.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_pci_link_status_t status

[out] The current status of the port

zes_pci_link_qual_issue_flags_t qualityIssues

[out] If status is ZES_PCI_LINK_STATUS_QUALITY_ISSUES, then this gives a combination of zes_pci_link_qual_issue_flag_t for quality issues that have been detected; otherwise, 0 indicates there are no quality issues with the link at this time.”

zes_pci_link_stab_issue_flags_t stabilityIssues

[out] If status is ZES_PCI_LINK_STATUS_STABILITY_ISSUES, then this gives a combination of zes_pci_link_stab_issue_flag_t for reasons for the connection instability; otherwise, 0 indicates there are no connection stability issues at this time.”

zes_pci_speed_t speed

[out] The current port configure speed

zes_pci_bar_properties_t

struct zes_pci_bar_properties_t

Properties of a pci bar.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_pci_bar_type_t type

[out] The type of bar

uint32_t index

[out] The index of the bar

uint64_t base

[out] Base address of the bar.

uint64_t size

[out] Size of the bar.

zes_pci_bar_properties_1_2_t

struct zes_pci_bar_properties_1_2_t

Properties of a pci bar, including the resizable bar.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_pci_bar_type_t type

[out] The type of bar

uint32_t index

[out] The index of the bar

uint64_t base

[out] Base address of the bar.

uint64_t size

[out] Size of the bar.

ze_bool_t resizableBarSupported

[out] Support for Resizable Bar on this device.

ze_bool_t resizableBarEnabled

[out] Resizable Bar enabled on this device

zes_pci_stats_t

struct zes_pci_stats_t

PCI stats counters.

  • Percent replays is calculated by taking two snapshots (s1, s2) and using the equation: replay = 10^6 * (s2.replayCounter - s1.replayCounter) / (s2.maxBandwidth * (s2.timestamp - s1.timestamp))

  • Percent throughput is calculated by taking two snapshots (s1, s2) and using the equation: bw = 10^6 * ((s2.rxCounter - s1.rxCounter) + (s2.txCounter - s1.txCounter)) / (s2.maxBandwidth * (s2.timestamp - s1.timestamp))

Public Members

uint64_t timestamp

[out] Monotonic timestamp counter in microseconds when the measurement was made. This timestamp should only be used to calculate delta time between snapshots of this structure. Never take the delta of this timestamp with the timestamp from a different structure since they are not guaranteed to have the same base. The absolute value of the timestamp is only valid during within the application and may be different on the next execution.

uint64_t replayCounter

[out] Monotonic counter for the number of replay packets (sum of all lanes). Will always be 0 when the haveReplayCounters member of zes_pci_properties_t is FALSE.

uint64_t packetCounter

[out] Monotonic counter for the number of packets (sum of all lanes). Will always be 0 when the havePacketCounters member of zes_pci_properties_t is FALSE.

uint64_t rxCounter

[out] Monotonic counter for the number of bytes received (sum of all lanes). Will always be 0 when the haveBandwidthCounters member of zes_pci_properties_t is FALSE.

uint64_t txCounter

[out] Monotonic counter for the number of bytes transmitted (including replays) (sum of all lanes). Will always be 0 when the haveBandwidthCounters member of zes_pci_properties_t is FALSE.

zes_pci_speed_t speed

[out] The current speed of the link (sum of all lanes)

Overclock

Overclock Functions

zesDeviceSetOverclockWaiver

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceSetOverclockWaiver(zes_device_handle_t hDevice)

Set the overclock waiver.The overclock waiver setting is persistent until the next pcode boot.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters

hDevice – [in] Sysman handle of the device.

Returns

zesDeviceGetOverclockDomains

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceGetOverclockDomains(zes_device_handle_t hDevice, uint32_t *pOverclockDomains)

Get the list of supported overclock domains for this device.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pOverclockDomains – [in,out] Returns the overclock domains that are supported (a bit for each of enum zes_overclock_domain_t). If no bits are set, the device doesn’t support overclocking.

Returns

zesDeviceGetOverclockControls

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceGetOverclockControls(zes_device_handle_t hDevice, zes_overclock_domain_t domainType, uint32_t *pAvailableControls)

Get the list of supported overclock controls available for one of the supported overclock domains on the device.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • domainType – [in] Domain type.

  • pAvailableControls – [in,out] Returns the overclock controls that are supported for the specified overclock domain (a bit for each of enum zes_overclock_control_t).

Returns

zesDeviceResetOverclockSettings

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceResetOverclockSettings(zes_device_handle_t hDevice, ze_bool_t onShippedState)

Reset all overclock settings to default values (shipped = 1 or manufacturing =0)

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • onShippedState – [in] True will reset to shipped state; false will reset to manufacturing state

Returns

zesDeviceReadOverclockState

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceReadOverclockState(zes_device_handle_t hDevice, zes_overclock_mode_t *pOverclockMode, ze_bool_t *pWaiverSetting, ze_bool_t *pOverclockState, zes_pending_action_t *pPendingAction, ze_bool_t *pPendingReset)

Determine the state of overclocking.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pOverclockMode – [out] One of overclock mode.

  • pWaiverSetting – [out] Waiver setting: 0 = Waiver not set, 1 = waiver has been set.

  • pOverclockState – [out] Current settings 0 =manufacturing state, 1= shipped state)..

  • pPendingAction – [out] This enum is returned when the driver attempts to set an overclock control or reset overclock settings.

  • pPendingReset – [out] Pending reset 0 =manufacturing state, 1= shipped state)..

Returns

zesDeviceEnumOverclockDomains

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumOverclockDomains(zes_device_handle_t hDevice, uint32_t *pCount, zes_overclock_handle_t *phDomainHandle)

Get handle of overclock domains.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phDomainHandle – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesOverclockGetDomainProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockGetDomainProperties(zes_overclock_handle_t hDomainHandle, zes_overclock_properties_t *pDomainProperties)

Get overclock domain control properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • pDomainProperties – [in,out] The overclock properties for the specified domain.

Returns

zesOverclockGetDomainVFProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockGetDomainVFProperties(zes_overclock_handle_t hDomainHandle, zes_vf_property_t *pVFProperties)

Read overclock VF min,max and step values.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • pVFProperties – [in,out] The VF min,max,step for a specified domain.

Returns

zesOverclockGetDomainControlProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockGetDomainControlProperties(zes_overclock_handle_t hDomainHandle, zes_overclock_control_t DomainControl, zes_control_property_t *pControlProperties)

Read overclock control values - min/max/step/default/ref.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • DomainControl – [in] Handle for the component.

  • pControlProperties – [in,out] overclock control values.

Returns

zesOverclockGetControlCurrentValue

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockGetControlCurrentValue(zes_overclock_handle_t hDomainHandle, zes_overclock_control_t DomainControl, double *pValue)

Read the current value for a given overclock control.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component.

  • DomainControl – [in] Overclock Control.

  • pValue – [in,out] Getting overclock control value for the specified control.

Returns

zesOverclockGetControlPendingValue

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockGetControlPendingValue(zes_overclock_handle_t hDomainHandle, zes_overclock_control_t DomainControl, double *pValue)

Read the the reset pending value for a given overclock control.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • DomainControl – [in] Overclock Control.

  • pValue – [out] Returns the pending value for a given control. The units and format of the value depend on the control type.

Returns

zesOverclockSetControlUserValue

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockSetControlUserValue(zes_overclock_handle_t hDomainHandle, zes_overclock_control_t DomainControl, double pValue, zes_pending_action_t *pPendingAction)

Set the value for a given overclock control.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • DomainControl – [in] Domain Control.

  • pValue – [in] The new value of the control. The units and format of the value depend on the control type.

  • pPendingAction – [out] Pending overclock setting.

Returns

zesOverclockGetControlState

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockGetControlState(zes_overclock_handle_t hDomainHandle, zes_overclock_control_t DomainControl, zes_control_state_t *pControlState, zes_pending_action_t *pPendingAction)

Determine the state of an overclock control.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • DomainControl – [in] Domain Control.

  • pControlState – [out] Current overclock control state.

  • pPendingAction – [out] Pending overclock setting.

Returns

zesOverclockGetVFPointValues

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockGetVFPointValues(zes_overclock_handle_t hDomainHandle, zes_vf_type_t VFType, zes_vf_array_type_t VFArrayType, uint32_t PointIndex, uint32_t *PointValue)

Read the frequency or voltage of a V-F point from the default or custom V-F curve.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • VFType – [in] Voltage or Freqency point to read.

  • VFArrayType – [in] User,Default or Live VF array to read from

  • PointIndex – [in] Point index - number between (0, max_num_points - 1).

  • PointValue – [out] Returns the frequency in 1kHz units or voltage in millivolt units from the custom V-F curve at the specified zero-based index

Returns

zesOverclockSetVFPointValues

ZE_APIEXPORT ze_result_t ZE_APICALL zesOverclockSetVFPointValues(zes_overclock_handle_t hDomainHandle, zes_vf_type_t VFType, uint32_t PointIndex, uint32_t PointValue)

Write the frequency or voltage of a V-F point to custom V-F curve.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDomainHandle – [in] Handle for the component domain.

  • VFType – [in] Voltage or Freqency point to read.

  • PointIndex – [in] Point index - number between (0, max_num_points - 1).

  • PointValue – [in] Writes frequency in 1kHz units or voltage in millivolt units to custom V-F curve at the specified zero-based index

Returns

Overclock Enums

zes_overclock_domain_t

enum zes_overclock_domain_t

Overclock domains.

Values:

enumerator ZES_OVERCLOCK_DOMAIN_CARD

Overclocking card level properties such as temperature limits.

enumerator ZES_OVERCLOCK_DOMAIN_PACKAGE

Overclocking package level properties such as power limits.

enumerator ZES_OVERCLOCK_DOMAIN_GPU_ALL

Overclocking a GPU that has all accelerator assets on the same PLL/VR.

enumerator ZES_OVERCLOCK_DOMAIN_GPU_RENDER_COMPUTE

Overclocking a GPU with render and compute assets on the same PLL/VR.

enumerator ZES_OVERCLOCK_DOMAIN_GPU_RENDER

Overclocking a GPU with render assets on its own PLL/VR.

enumerator ZES_OVERCLOCK_DOMAIN_GPU_COMPUTE

Overclocking a GPU with compute assets on its own PLL/VR.

enumerator ZES_OVERCLOCK_DOMAIN_GPU_MEDIA

Overclocking a GPU with media assets on its own PLL/VR.

enumerator ZES_OVERCLOCK_DOMAIN_VRAM

Overclocking device local memory.

enumerator ZES_OVERCLOCK_DOMAIN_ADM

Overclocking LLC/L4 cache.

enumerator ZES_OVERCLOCK_DOMAIN_FORCE_UINT32

zes_overclock_control_t

enum zes_overclock_control_t

Overclock controls.

Values:

enumerator ZES_OVERCLOCK_CONTROL_VF

This control permits setting a custom V-F curve.

enumerator ZES_OVERCLOCK_CONTROL_FREQ_OFFSET

The V-F curve of the overclock domain can be shifted up or down using this control.

enumerator ZES_OVERCLOCK_CONTROL_VMAX_OFFSET

This control is used to increase the permitted voltage above the shipped voltage maximum.

enumerator ZES_OVERCLOCK_CONTROL_FREQ

This control permits direct changes to the operating frequency.

enumerator ZES_OVERCLOCK_CONTROL_VOLT_LIMIT

This control prevents frequencies that would push the voltage above this value, typically used by V-F scanners.

enumerator ZES_OVERCLOCK_CONTROL_POWER_SUSTAINED_LIMIT

This control changes the sustained power limit (PL1).

enumerator ZES_OVERCLOCK_CONTROL_POWER_BURST_LIMIT

This control changes the burst power limit (PL2).

enumerator ZES_OVERCLOCK_CONTROL_POWER_PEAK_LIMIT

his control changes the peak power limit (PL4).

enumerator ZES_OVERCLOCK_CONTROL_ICCMAX_LIMIT

This control changes the value of IccMax..

enumerator ZES_OVERCLOCK_CONTROL_TEMP_LIMIT

This control changes the value of TjMax.

enumerator ZES_OVERCLOCK_CONTROL_ITD_DISABLE

This control permits disabling the adaptive voltage feature ITD.

enumerator ZES_OVERCLOCK_CONTROL_ACM_DISABLE

This control permits disabling the adaptive voltage feature ACM.

enumerator ZES_OVERCLOCK_CONTROL_FORCE_UINT32

zes_overclock_mode_t

enum zes_overclock_mode_t

Overclock modes.

Values:

enumerator ZES_OVERCLOCK_MODE_MODE_OFF

Overclock mode is off.

enumerator ZES_OVERCLOCK_MODE_MODE_STOCK

Stock (manufacturing settings) are being used.

enumerator ZES_OVERCLOCK_MODE_MODE_ON

Overclock mode is on.

enumerator ZES_OVERCLOCK_MODE_MODE_UNAVAILABLE

Overclocking is unavailable at this time since the system is running on battery.

enumerator ZES_OVERCLOCK_MODE_MODE_DISABLED

Overclock mode is disabled.

enumerator ZES_OVERCLOCK_MODE_FORCE_UINT32

zes_control_state_t

enum zes_control_state_t

Overclock control states.

Values:

enumerator ZES_CONTROL_STATE_STATE_UNSET

No overclock control has not been changed by the driver since the last boot/reset.

enumerator ZES_CONTROL_STATE_STATE_ACTIVE

The overclock control has been set and it is active.

enumerator ZES_CONTROL_STATE_STATE_DISABLED

The overclock control value has been disabled due to the current power configuration (typically when running on DC).

enumerator ZES_CONTROL_STATE_FORCE_UINT32

zes_pending_action_t

enum zes_pending_action_t

Overclock pending actions.

Values:

enumerator ZES_PENDING_ACTION_PENDING_NONE

There no pending actions. .

enumerator ZES_PENDING_ACTION_PENDING_IMMINENT

The requested change is in progress and should complete soon.

enumerator ZES_PENDING_ACTION_PENDING_COLD_RESET

The requested change requires a device cold reset (hotplug, system boot).

enumerator ZES_PENDING_ACTION_PENDING_WARM_RESET

The requested change requires a device warm reset (PCIe FLR).

enumerator ZES_PENDING_ACTION_FORCE_UINT32

zes_vf_program_type_t

enum zes_vf_program_type_t

Overclock V-F curve programing.

Values:

enumerator ZES_VF_PROGRAM_TYPE_VF_ARBITRARY

Can program an arbitrary number of V-F points up to the maximum number and each point can have arbitrary voltage and frequency values within the min/max/step limits

enumerator ZES_VF_PROGRAM_TYPE_VF_FREQ_FIXED

Can only program the voltage for the V-F points that it reads back - the frequency of those points cannot be changed

enumerator ZES_VF_PROGRAM_TYPE_VF_VOLT_FIXED

Can only program the frequency for the V-F points that is reads back - the voltage of each point cannot be changed.

enumerator ZES_VF_PROGRAM_TYPE_FORCE_UINT32

zes_vf_type_t

enum zes_vf_type_t

VF type.

Values:

enumerator ZES_VF_TYPE_VOLT

VF Voltage point.

enumerator ZES_VF_TYPE_FREQ

VF Frequency point.

enumerator ZES_VF_TYPE_FORCE_UINT32

zes_vf_array_type_t

enum zes_vf_array_type_t

VF type.

Values:

enumerator ZES_VF_ARRAY_TYPE_USER_VF_ARRAY

User V-F array.

enumerator ZES_VF_ARRAY_TYPE_DEFAULT_VF_ARRAY

Default V-F array.

enumerator ZES_VF_ARRAY_TYPE_LIVE_VF_ARRAY

Live V-F array.

enumerator ZES_VF_ARRAY_TYPE_FORCE_UINT32

Overclock Structures

zes_overclock_properties_t

struct zes_overclock_properties_t

Overclock properties.

  • Information on the overclock domain type and all the contols that are part of the domain.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_overclock_domain_t domainType

[out] The hardware block that this overclock domain controls (GPU, VRAM, …)

uint32_t AvailableControls

[out] Returns the overclock controls that are supported (a bit for each of enum zes_overclock_control_t). If no bits are set, the domain doesn’t support overclocking.

zes_vf_program_type_t VFProgramType

[out] Type of V-F curve programming that is permitted:.

uint32_t NumberOfVFPoints

[out] Number of VF points that can be programmed - max_num_points

zes_control_property_t

struct zes_control_property_t

Overclock Control properties.

  • Provides all the control capabilities supported by the device for the overclock domain.

Public Members

double MinValue

[out] This provides information about the limits of the control value so that the driver can calculate the set of valid values.

double MaxValue

[out] This provides information about the limits of the control value so that the driver can calculate the set of valid values.

double StepValue

[out] This provides information about the limits of the control value so that the driver can calculate the set of valid values.

double RefValue

[out] The reference value provides the anchor point, UIs can combine this with the user offset request to show the anticipated improvement.

double DefaultValue

[out] The shipped out-of-box position of this control. Driver can request this value at any time to return to the out-of-box behavior.

zes_vf_property_t

struct zes_vf_property_t

Overclock VF properties.

  • Provides all the VF capabilities supported by the device for the overclock domain.

Public Members

double MinFreq

[out] Read the minimum frequency that can be be programmed in the custom V-F point..

double MaxFreq

[out] Read the maximum frequency that can be be programmed in the custom V-F point..

double StepFreq

[out] Read the frequency step that can be be programmed in the custom V-F point..

double MinVolt

[out] Read the minimum voltage that can be be programmed in the custom V-F point..

double MaxVolt

[out] Read the maximum voltage that can be be programmed in the custom V-F point..

double StepVolt

[out] Read the voltage step that can be be programmed in the custom V-F point.

Diagnostics

Diagnostics Functions

zesDeviceEnumDiagnosticTestSuites

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumDiagnosticTestSuites(zes_device_handle_t hDevice, uint32_t *pCount, zes_diag_handle_t *phDiagnostics)

Get handle of diagnostics test suites.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phDiagnostics – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesDiagnosticsGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesDiagnosticsGetProperties(zes_diag_handle_t hDiagnostics, zes_diag_properties_t *pProperties)

Get properties of a diagnostics test suite.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDiagnostics – [in] Handle for the component.

  • pProperties – [in,out] Structure describing the properties of a diagnostics test suite

Returns

zesDiagnosticsGetTests

ZE_APIEXPORT ze_result_t ZE_APICALL zesDiagnosticsGetTests(zes_diag_handle_t hDiagnostics, uint32_t *pCount, zes_diag_test_t *pTests)

Get individual tests that can be run separately. Not all test suites permit running individual tests, check the haveTests member of zes_diag_properties_t.

  • The list of available tests is returned in order of increasing test index (see the index member of zes_diag_test_t).

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDiagnostics – [in] Handle for the component.

  • pCount – [in,out] pointer to the number of tests. if count is zero, then the driver shall update the value with the total number of tests that are available. if count is greater than the number of tests that are available, then the driver shall update the value with the correct number of tests.

  • pTests – [in,out][optional][range(0, *pCount)] array of information about individual tests sorted by increasing value of the index member of zes_diag_test_t. if count is less than the number of tests that are available, then the driver shall only retrieve that number of tests.

Returns

zesDiagnosticsRunTests

ZE_APIEXPORT ze_result_t ZE_APICALL zesDiagnosticsRunTests(zes_diag_handle_t hDiagnostics, uint32_t startIndex, uint32_t endIndex, zes_diag_result_t *pResult)

Run a diagnostics test suite, either all tests or a subset of tests.

  • WARNING: Running diagnostics may destroy current device state information. Gracefully close any running workloads before initiating.

  • To run all tests in a test suite, set start = ZES_DIAG_FIRST_TEST_INDEX and end = ZES_DIAG_LAST_TEST_INDEX.

  • If the test suite permits running individual tests, the haveTests member of zes_diag_properties_t will be true. In this case, the function zesDiagnosticsGetTests() can be called to get the list of tests and corresponding indices that can be supplied to the arguments start and end in this function.

  • This function will block until the diagnostics have completed and force reset based on result

Parameters
  • hDiagnostics – [in] Handle for the component.

  • startIndex – [in] The index of the first test to run. Set to ZES_DIAG_FIRST_TEST_INDEX to start from the beginning.

  • endIndex – [in] The index of the last test to run. Set to ZES_DIAG_LAST_TEST_INDEX to complete all tests after the start test.

  • pResult – [in,out] The result of the diagnostics

Returns

Diagnostics Enums

zes_diag_result_t

enum zes_diag_result_t

Diagnostic results.

Values:

enumerator ZES_DIAG_RESULT_NO_ERRORS

Diagnostic completed without finding errors to repair.

enumerator ZES_DIAG_RESULT_ABORT

Diagnostic had problems running tests.

enumerator ZES_DIAG_RESULT_FAIL_CANT_REPAIR

Diagnostic had problems setting up repairs.

enumerator ZES_DIAG_RESULT_REBOOT_FOR_REPAIR

Diagnostics found errors, setup for repair and reboot is required to complete the process

enumerator ZES_DIAG_RESULT_FORCE_UINT32

Diagnostics Structures

zes_diag_test_t

struct zes_diag_test_t

Diagnostic test.

Public Members

uint32_t index

[out] Index of the test

char name[ZES_STRING_PROPERTY_SIZE]

[out] Name of the test

zes_diag_properties_t

struct zes_diag_properties_t

Diagnostics test suite properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

char name[ZES_STRING_PROPERTY_SIZE]

[out] Name of the diagnostics test suite

ze_bool_t haveTests

[out] Indicates if this test suite has individual tests which can be run separately (use the function zesDiagnosticsGetTests() to get the list of these tests)

Ecc

Ecc Functions

zesDeviceEccAvailable

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEccAvailable(zes_device_handle_t hDevice, ze_bool_t *pAvailable)

Is ECC functionality available - true or false?

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Handle for the component.

  • pAvailable – [out] ECC functionality is available (true)/unavailable (false).

Returns

zesDeviceEccConfigurable

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEccConfigurable(zes_device_handle_t hDevice, ze_bool_t *pConfigurable)

Is ECC support configurable - true or false?

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Handle for the component.

  • pConfigurable – [out] ECC can be enabled/disabled (true)/enabled/disabled (false).

Returns

zesDeviceGetEccState

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceGetEccState(zes_device_handle_t hDevice, zes_device_ecc_properties_t *pState)

Get current ECC state, pending state, and pending action.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Handle for the component.

  • pState – [out] ECC state, pending state, and pending action for state change.

Returns

zesDeviceSetEccState

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceSetEccState(zes_device_handle_t hDevice, const zes_device_ecc_desc_t *newState, zes_device_ecc_properties_t *pState)

Set new ECC state.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • zesDeviceGetState should be called to determine pending action required to implement state change.

Parameters
  • hDevice – [in] Handle for the component.

  • newState – [in] Pointer to desired ECC state.

  • pState – [out] ECC state, pending state, and pending action for state change.

Returns

Ecc Enums

zes_device_ecc_state_t

enum zes_device_ecc_state_t

ECC State.

Values:

enumerator ZES_DEVICE_ECC_STATE_UNAVAILABLE

None.

enumerator ZES_DEVICE_ECC_STATE_ENABLED

ECC enabled.

enumerator ZES_DEVICE_ECC_STATE_DISABLED

ECC disabled.

enumerator ZES_DEVICE_ECC_STATE_FORCE_UINT32

zes_device_action_t

enum zes_device_action_t

State Change Requirements.

Values:

enumerator ZES_DEVICE_ACTION_NONE

No action.

enumerator ZES_DEVICE_ACTION_WARM_CARD_RESET

Warm reset of the card.

enumerator ZES_DEVICE_ACTION_COLD_CARD_RESET

Cold reset of the card.

enumerator ZES_DEVICE_ACTION_COLD_SYSTEM_REBOOT

Cold reboot of the system.

enumerator ZES_DEVICE_ACTION_FORCE_UINT32

Ecc Structures

zes_device_ecc_desc_t

struct zes_device_ecc_desc_t

ECC State Descriptor.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_device_ecc_state_t state

[out] ECC state

zes_device_ecc_properties_t

struct zes_device_ecc_properties_t

ECC State.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_device_ecc_state_t currentState

[out] Current ECC state

zes_device_ecc_state_t pendingState

[out] Pending ECC state

zes_device_action_t pendingAction

[out] Pending action

Engine

Engine Functions

zesDeviceEnumEngineGroups

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumEngineGroups(zes_device_handle_t hDevice, uint32_t *pCount, zes_engine_handle_t *phEngine)

Get handle of engine groups.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phEngine – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesEngineGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesEngineGetProperties(zes_engine_handle_t hEngine, zes_engine_properties_t *pProperties)

Get engine group properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hEngine – [in] Handle for the component.

  • pProperties – [in,out] The properties for the specified engine group.

Returns

zesEngineGetActivity

ZE_APIEXPORT ze_result_t ZE_APICALL zesEngineGetActivity(zes_engine_handle_t hEngine, zes_engine_stats_t *pStats)

Get the activity stats for an engine group.

  • This function also returns the engine activity inside a Virtual Machine (VM), in the presence of hardware virtualization (SRIOV)

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hEngine – [in] Handle for the component.

  • pStats – [in,out] Will contain a snapshot of the engine group activity counters.

Returns

Engine Enums

zes_engine_group_t

enum zes_engine_group_t

Accelerator engine groups.

Values:

enumerator ZES_ENGINE_GROUP_ALL

Access information about all engines combined.

enumerator ZES_ENGINE_GROUP_COMPUTE_ALL

Access information about all compute engines combined. Compute engines can only process compute kernels (no 3D content).

enumerator ZES_ENGINE_GROUP_MEDIA_ALL

Access information about all media engines combined.

enumerator ZES_ENGINE_GROUP_COPY_ALL

Access information about all copy (blitter) engines combined.

enumerator ZES_ENGINE_GROUP_COMPUTE_SINGLE

Access information about a single compute engine - this is an engine that can process compute kernels. Note that single engines may share the same underlying accelerator resources as other engines so activity of such an engine may not be indicative of the underlying resource utilization - use ZES_ENGINE_GROUP_3D_RENDER_COMPUTE_ALL for that.

enumerator ZES_ENGINE_GROUP_RENDER_SINGLE

Access information about a single render engine - this is an engine that can process both 3D content and compute kernels. Note that single engines may share the same underlying accelerator resources as other engines so activity of such an engine may not be indicative of the underlying resource utilization - use ZES_ENGINE_GROUP_3D_RENDER_COMPUTE_ALL for that.

enumerator ZES_ENGINE_GROUP_MEDIA_DECODE_SINGLE

[DEPRECATED] No longer supported.

enumerator ZES_ENGINE_GROUP_MEDIA_ENCODE_SINGLE

[DEPRECATED] No longer supported.

enumerator ZES_ENGINE_GROUP_COPY_SINGLE

Access information about a single media encode engine. Note that single engines may share the same underlying accelerator resources as other engines so activity of such an engine may not be indicative of the underlying resource utilization - use ZES_ENGINE_GROUP_COPY_ALL for that.

enumerator ZES_ENGINE_GROUP_MEDIA_ENHANCEMENT_SINGLE

Access information about a single media enhancement engine. Note that single engines may share the same underlying accelerator resources as other engines so activity of such an engine may not be indicative of the underlying resource utilization - use ZES_ENGINE_GROUP_MEDIA_ALL for that.

enumerator ZES_ENGINE_GROUP_3D_SINGLE

[DEPRECATED] No longer supported.

enumerator ZES_ENGINE_GROUP_3D_RENDER_COMPUTE_ALL

[DEPRECATED] No longer supported.

enumerator ZES_ENGINE_GROUP_RENDER_ALL

Access information about all render engines combined. Render engines are those than process both 3D content and compute kernels.

enumerator ZES_ENGINE_GROUP_3D_ALL

[DEPRECATED] No longer supported.

enumerator ZES_ENGINE_GROUP_MEDIA_CODEC_SINGLE

Access information about a single media engine. Note that single engines may share the same underlying accelerator resources as other engines so activity of such an engine may not be indicative of the underlying resource utilization - use ZES_ENGINE_GROUP_MEDIA_ALL for that.

enumerator ZES_ENGINE_GROUP_FORCE_UINT32

Engine Structures

zes_engine_properties_t

struct zes_engine_properties_t

Engine group properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_engine_group_t type

[out] The engine group

ze_bool_t onSubdevice

[out] True if this resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

zes_engine_stats_t

struct zes_engine_stats_t

Engine activity counters.

  • Percent utilization is calculated by taking two snapshots (s1, s2) and using the equation: util = (s2.activeTime - s1.activeTime) / (s2.timestamp - s1.timestamp)

Public Members

uint64_t activeTime

[out] Monotonic counter where the resource is actively running workloads. Time units are implementation specific since the activeTime value is only intended for calculating utilization percentage as noted above.

uint64_t timestamp

[out] Monotonic counter when activeTime counter was sampled. This timestamp should only be used to calculate delta between snapshots of this structure. Never take the delta of this timestamp with the timestamp from a different structure since they are not guaranteed to have the same base. The absolute value of the timestamp is only valid during within the application and may be different on the next execution.

Events

Events Functions

zesDeviceEventRegister

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEventRegister(zes_device_handle_t hDevice, zes_event_type_flags_t events)

Specify the list of events to listen to for a given device.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] The device handle.

  • events – [in] List of events to listen to.

Returns

zesDriverEventListen

ZE_APIEXPORT ze_result_t ZE_APICALL zesDriverEventListen(ze_driver_handle_t hDriver, uint32_t timeout, uint32_t count, zes_device_handle_t *phDevices, uint32_t *pNumDeviceEvents, zes_event_type_flags_t *pEvents)

Wait for events to be received from a one or more devices.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDriver – [in] handle of the driver instance

  • timeout – [in] if non-zero, then indicates the maximum time (in milliseconds) to yield before returning ZE_RESULT_SUCCESS or ZE_RESULT_NOT_READY; if zero, then will check status and return immediately; if UINT32_MAX, then function will not return until events arrive.

  • count – [in] Number of device handles in phDevices.

  • phDevices – [in][range(0, count)] Device handles to listen to for events. Only devices from the provided driver handle can be specified in this list.

  • pNumDeviceEvents – [in,out] Will contain the actual number of devices in phDevices that generated events. If non-zero, check pEvents to determine the devices and events that were received.

  • pEvents – [in,out] An array that will continue the list of events for each device listened in phDevices. This array must be at least as big as count. For every device handle in phDevices, this will provide the events that occurred for that device at the same position in this array. If no event was received for a given device, the corresponding array entry will be zero.

Returns

zesDriverEventListenEx

ZE_APIEXPORT ze_result_t ZE_APICALL zesDriverEventListenEx(ze_driver_handle_t hDriver, uint64_t timeout, uint32_t count, zes_device_handle_t *phDevices, uint32_t *pNumDeviceEvents, zes_event_type_flags_t *pEvents)

Wait for events to be received from a one or more devices.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDriver – [in] handle of the driver instance

  • timeout – [in] if non-zero, then indicates the maximum time (in milliseconds) to yield before returning ZE_RESULT_SUCCESS or ZE_RESULT_NOT_READY; if zero, then will check status and return immediately; if UINT64_MAX, then function will not return until events arrive.

  • count – [in] Number of device handles in phDevices.

  • phDevices – [in][range(0, count)] Device handles to listen to for events. Only devices from the provided driver handle can be specified in this list.

  • pNumDeviceEvents – [in,out] Will contain the actual number of devices in phDevices that generated events. If non-zero, check pEvents to determine the devices and events that were received.

  • pEvents – [in,out] An array that will continue the list of events for each device listened in phDevices. This array must be at least as big as count. For every device handle in phDevices, this will provide the events that occurred for that device at the same position in this array. If no event was received for a given device, the corresponding array entry will be zero.

Returns

Events Enums

zes_event_type_flags_t

enum zes_event_type_flag_t

Values:

enumerator ZES_EVENT_TYPE_FLAG_DEVICE_DETACH

Event is triggered when the device is no longer available (due to a reset or being disabled).

enumerator ZES_EVENT_TYPE_FLAG_DEVICE_ATTACH

Event is triggered after the device is available again.

enumerator ZES_EVENT_TYPE_FLAG_DEVICE_SLEEP_STATE_ENTER

Event is triggered when the driver is about to put the device into a deep sleep state

enumerator ZES_EVENT_TYPE_FLAG_DEVICE_SLEEP_STATE_EXIT

Event is triggered when the driver is waking the device up from a deep sleep state

enumerator ZES_EVENT_TYPE_FLAG_FREQ_THROTTLED

Event is triggered when the frequency starts being throttled.

enumerator ZES_EVENT_TYPE_FLAG_ENERGY_THRESHOLD_CROSSED

Event is triggered when the energy consumption threshold is reached (use zesPowerSetEnergyThreshold() to configure).

enumerator ZES_EVENT_TYPE_FLAG_TEMP_CRITICAL

Event is triggered when the critical temperature is reached (use zesTemperatureSetConfig() to configure - disabled by default).

enumerator ZES_EVENT_TYPE_FLAG_TEMP_THRESHOLD1

Event is triggered when the temperature crosses threshold 1 (use zesTemperatureSetConfig() to configure - disabled by default).

enumerator ZES_EVENT_TYPE_FLAG_TEMP_THRESHOLD2

Event is triggered when the temperature crosses threshold 2 (use zesTemperatureSetConfig() to configure - disabled by default).

enumerator ZES_EVENT_TYPE_FLAG_MEM_HEALTH

Event is triggered when the health of device memory changes.

enumerator ZES_EVENT_TYPE_FLAG_FABRIC_PORT_HEALTH

Event is triggered when the health of fabric ports change.

Event is triggered when the health of the PCI link changes.

enumerator ZES_EVENT_TYPE_FLAG_RAS_CORRECTABLE_ERRORS

Event is triggered when accelerator RAS correctable errors cross thresholds (use zesRasSetConfig() to configure - disabled by default).

enumerator ZES_EVENT_TYPE_FLAG_RAS_UNCORRECTABLE_ERRORS

Event is triggered when accelerator RAS uncorrectable errors cross thresholds (use zesRasSetConfig() to configure - disabled by default).

enumerator ZES_EVENT_TYPE_FLAG_DEVICE_RESET_REQUIRED

Event is triggered when the device needs to be reset (use zesDeviceGetState() to determine the reasons for the reset).

enumerator ZES_EVENT_TYPE_FLAG_FORCE_UINT32

Fabric

Fabric Functions

zesDeviceEnumFabricPorts

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumFabricPorts(zes_device_handle_t hDevice, uint32_t *pCount, zes_fabric_port_handle_t *phPort)

Get handle of Fabric ports in a device.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phPort – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesFabricPortGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesFabricPortGetProperties(zes_fabric_port_handle_t hPort, zes_fabric_port_properties_t *pProperties)

Get Fabric port properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPort – [in] Handle for the component.

  • pProperties – [in,out] Will contain properties of the Fabric Port.

Returns

zesFabricPortGetLinkType

Get Fabric port link type.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPort – [in] Handle for the component.

  • pLinkType – [in,out] Will contain details about the link attached to the Fabric port.

Returns

zesFabricPortGetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesFabricPortGetConfig(zes_fabric_port_handle_t hPort, zes_fabric_port_config_t *pConfig)

Get Fabric port configuration.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPort – [in] Handle for the component.

  • pConfig – [in,out] Will contain configuration of the Fabric Port.

Returns

zesFabricPortSetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesFabricPortSetConfig(zes_fabric_port_handle_t hPort, const zes_fabric_port_config_t *pConfig)

Set Fabric port configuration.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPort – [in] Handle for the component.

  • pConfig – [in] Contains new configuration of the Fabric Port.

Returns

zesFabricPortGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesFabricPortGetState(zes_fabric_port_handle_t hPort, zes_fabric_port_state_t *pState)

Get Fabric port state - status (health/degraded/failed/disabled), reasons for link degradation or instability, current rx/tx speed.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPort – [in] Handle for the component.

  • pState – [in,out] Will contain the current state of the Fabric Port

Returns

zesFabricPortGetThroughput

ZE_APIEXPORT ze_result_t ZE_APICALL zesFabricPortGetThroughput(zes_fabric_port_handle_t hPort, zes_fabric_port_throughput_t *pThroughput)

Get Fabric port throughput.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPort – [in] Handle for the component.

  • pThroughput – [in,out] Will contain the Fabric port throughput counters.

Returns

zesFabricPortGetFabricErrorCounters

ZE_APIEXPORT ze_result_t ZE_APICALL zesFabricPortGetFabricErrorCounters(zes_fabric_port_handle_t hPort, zes_fabric_port_error_counters_t *pErrors)

Get Fabric Port Error Counters.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • The memory backing the arrays for phPorts and ppThroughputs must be allocated in system memory by the user who is also responsible for releasing them when they are no longer needed.

Parameters
  • hPort – [in] Handle for the component.

  • pErrors – [in,out] Will contain the Fabric port Error counters.

Returns

zesFabricPortGetMultiPortThroughput

ZE_APIEXPORT ze_result_t ZE_APICALL zesFabricPortGetMultiPortThroughput(zes_device_handle_t hDevice, uint32_t numPorts, zes_fabric_port_handle_t *phPort, zes_fabric_port_throughput_t **pThroughput)

Get Fabric port throughput from multiple ports in a single call.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • numPorts – [in] Number of ports enumerated in function zesDeviceEnumFabricPorts

  • phPort – [in][range(0, numPorts)] array of fabric port handles provided by user to gather throughput values.

  • pThroughput – [out][range(0, numPorts)] array of fabric port throughput counters from multiple ports of type zes_fabric_port_throughput_t.

Returns

Fabric Enums

zes_fabric_port_status_t

enum zes_fabric_port_status_t

Fabric port status.

Values:

enumerator ZES_FABRIC_PORT_STATUS_UNKNOWN

The port status cannot be determined.

enumerator ZES_FABRIC_PORT_STATUS_HEALTHY

The port is up and operating as expected.

enumerator ZES_FABRIC_PORT_STATUS_DEGRADED

The port is up but has quality and/or speed degradation.

enumerator ZES_FABRIC_PORT_STATUS_FAILED

Port connection instabilities are preventing workloads making forward progress

enumerator ZES_FABRIC_PORT_STATUS_DISABLED

The port is configured down.

enumerator ZES_FABRIC_PORT_STATUS_FORCE_UINT32

zes_fabric_port_qual_issue_flags_t

enum zes_fabric_port_qual_issue_flag_t

Values:

Excessive link errors are occurring.

enumerator ZES_FABRIC_PORT_QUAL_ISSUE_FLAG_SPEED

There is a degradation in the bitrate and/or width of the link.

enumerator ZES_FABRIC_PORT_QUAL_ISSUE_FLAG_FORCE_UINT32

zes_fabric_port_failure_flags_t

enum zes_fabric_port_failure_flag_t

Values:

enumerator ZES_FABRIC_PORT_FAILURE_FLAG_FAILED

A previously operating link has failed. Hardware will automatically retrain this port. This state will persist until either the physical connection is removed or the link trains successfully.

enumerator ZES_FABRIC_PORT_FAILURE_FLAG_TRAINING_TIMEOUT

A connection has not been established within an expected time. Hardware will continue to attempt port training. This status will persist until either the physical connection is removed or the link successfully trains.

enumerator ZES_FABRIC_PORT_FAILURE_FLAG_FLAPPING

Port has excessively trained and then transitioned down for some period of time. Driver will allow port to continue to train, but will not enable the port for use until the port has been disabled and subsequently re-enabled using zesFabricPortSetConfig().

enumerator ZES_FABRIC_PORT_FAILURE_FLAG_FORCE_UINT32

Fabric Structures

zes_fabric_port_id_t

struct zes_fabric_port_id_t

Unique identifier for a fabric port.

  • This not a universal identifier. The identified is garanteed to be unique for the current hardware configuration of the system. Changes in the hardware may result in a different identifier for a given port.

  • The main purpose of this identifier to build up an instantaneous topology map of system connectivity. An application should enumerate all fabric ports and match the remotePortId member of zes_fabric_port_state_t to the portId member of zes_fabric_port_properties_t.

Public Members

uint32_t fabricId

[out] Unique identifier for the fabric end-point

uint32_t attachId

[out] Unique identifier for the device attachment point

uint8_t portNumber

[out] The logical port number (this is typically marked somewhere on the physical device)

zes_fabric_port_speed_t

struct zes_fabric_port_speed_t

Fabric port speed in one direction.

Public Members

int64_t bitRate

[out] Bits/sec that the link is operating at. A value of -1 means that this property is unknown.

int32_t width

[out] The number of lanes. A value of -1 means that this property is unknown.

zes_fabric_port_properties_t

struct zes_fabric_port_properties_t

Fabric port properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

char model[ZES_MAX_FABRIC_PORT_MODEL_SIZE]

[out] Description of port technology. Will be set to the string “unkown” if this cannot be determined for this port.

ze_bool_t onSubdevice

[out] True if the port is located on a sub-device; false means that the port is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

zes_fabric_port_id_t portId

[out] The unique port identifier

zes_fabric_port_speed_t maxRxSpeed

[out] Maximum speed supported by the receive side of the port (sum of all lanes)

zes_fabric_port_speed_t maxTxSpeed

[out] Maximum speed supported by the transmit side of the port (sum of all lanes)

zes_fabric_port_config_t

struct zes_fabric_port_config_t

Fabric port configuration.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t enabled

[in,out] Port is configured up/down

ze_bool_t beaconing

[in,out] Beaconing is configured on/off

zes_fabric_port_state_t

struct zes_fabric_port_state_t

Fabric port state.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_fabric_port_status_t status

[out] The current status of the port

zes_fabric_port_qual_issue_flags_t qualityIssues

[out] If status is ZES_FABRIC_PORT_STATUS_DEGRADED, then this gives a combination of zes_fabric_port_qual_issue_flag_t for quality issues that have been detected; otherwise, 0 indicates there are no quality issues with the link at this time.

zes_fabric_port_failure_flags_t failureReasons

[out] If status is ZES_FABRIC_PORT_STATUS_FAILED, then this gives a combination of zes_fabric_port_failure_flag_t for reasons for the connection instability; otherwise, 0 indicates there are no connection stability issues at this time.

zes_fabric_port_id_t remotePortId

[out] The unique port identifier for the remote connection point if status is ZES_FABRIC_PORT_STATUS_HEALTHY, ZES_FABRIC_PORT_STATUS_DEGRADED or ZES_FABRIC_PORT_STATUS_FAILED

zes_fabric_port_speed_t rxSpeed

[out] Current maximum receive speed (sum of all lanes)

zes_fabric_port_speed_t txSpeed

[out] Current maximum transmit speed (sum of all lanes)

zes_fabric_port_throughput_t

struct zes_fabric_port_throughput_t

Fabric port throughput.

Public Members

uint64_t timestamp

[out] Monotonic timestamp counter in microseconds when the measurement was made. This timestamp should only be used to calculate delta time between snapshots of this structure. Never take the delta of this timestamp with the timestamp from a different structure since they are not guaranteed to have the same base. The absolute value of the timestamp is only valid during within the application and may be different on the next execution.

uint64_t rxCounter

[out] Monotonic counter for the number of bytes received (sum of all lanes). This includes all protocol overhead, not only the GPU traffic.

uint64_t txCounter

[out] Monotonic counter for the number of bytes transmitted (sum of all lanes). This includes all protocol overhead, not only the GPU traffic.

zes_fabric_port_error_counters_t

struct zes_fabric_port_error_counters_t

Fabric Port Error Counters.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint64_t linkFailureCount

[out] Link Failure Error Count reported per port

uint64_t fwCommErrorCount

[out] Firmware Communication Error Count reported per device

uint64_t fwErrorCount

[out] Firmware reported Error Count reported per device

uint64_t linkDegradeCount

[out] Link Degrade Error Count reported per port

Fan

Fan Functions

zesDeviceEnumFans

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumFans(zes_device_handle_t hDevice, uint32_t *pCount, zes_fan_handle_t *phFan)

Get handle of fans.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phFan – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesFanGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesFanGetProperties(zes_fan_handle_t hFan, zes_fan_properties_t *pProperties)

Get fan properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFan – [in] Handle for the component.

  • pProperties – [in,out] Will contain the properties of the fan.

Returns

zesFanGetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesFanGetConfig(zes_fan_handle_t hFan, zes_fan_config_t *pConfig)

Get fan configurations and the current fan speed mode (default, fixed, temp-speed table)

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFan – [in] Handle for the component.

  • pConfig – [in,out] Will contain the current configuration of the fan.

Returns

zesFanSetDefaultMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesFanSetDefaultMode(zes_fan_handle_t hFan)

Configure the fan to run with hardware factory settings (set mode to ZES_FAN_SPEED_MODE_DEFAULT)

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters

hFan – [in] Handle for the component.

Returns

zesFanSetFixedSpeedMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesFanSetFixedSpeedMode(zes_fan_handle_t hFan, const zes_fan_speed_t *speed)

Configure the fan to rotate at a fixed speed (set mode to ZES_FAN_SPEED_MODE_FIXED)

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFan – [in] Handle for the component.

  • speed – [in] The fixed fan speed setting

Returns

zesFanSetSpeedTableMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesFanSetSpeedTableMode(zes_fan_handle_t hFan, const zes_fan_speed_table_t *speedTable)

Configure the fan to adjust speed based on a temperature/speed table (set mode to ZES_FAN_SPEED_MODE_TABLE)

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFan – [in] Handle for the component.

  • speedTable – [in] A table containing temperature/speed pairs.

Returns

zesFanGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesFanGetState(zes_fan_handle_t hFan, zes_fan_speed_units_t units, int32_t *pSpeed)

Get current state of a fan - current mode and speed.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFan – [in] Handle for the component.

  • units – [in] The units in which the fan speed should be returned.

  • pSpeed – [in,out] Will contain the current speed of the fan in the units requested. A value of -1 indicates that the fan speed cannot be measured.

Returns

Fan Enums

zes_fan_speed_mode_t

enum zes_fan_speed_mode_t

Fan resource speed mode.

Values:

enumerator ZES_FAN_SPEED_MODE_DEFAULT

The fan speed is operating using the hardware default settings.

enumerator ZES_FAN_SPEED_MODE_FIXED

The fan speed is currently set to a fixed value.

enumerator ZES_FAN_SPEED_MODE_TABLE

The fan speed is currently controlled dynamically by hardware based on a temp/speed table

enumerator ZES_FAN_SPEED_MODE_FORCE_UINT32

zes_fan_speed_units_t

enum zes_fan_speed_units_t

Fan speed units.

Values:

enumerator ZES_FAN_SPEED_UNITS_RPM

The fan speed is in units of revolutions per minute (rpm)

enumerator ZES_FAN_SPEED_UNITS_PERCENT

The fan speed is a percentage of the maximum speed of the fan.

enumerator ZES_FAN_SPEED_UNITS_FORCE_UINT32

Fan Structures

zes_fan_speed_t

struct zes_fan_speed_t

Fan speed.

Public Members

int32_t speed

[in,out] The speed of the fan. On output, a value of -1 indicates that there is no fixed fan speed setting.

zes_fan_speed_units_t units

[in,out] The units that the fan speed is expressed in. On output, if fan speed is -1 then units should be ignored.

zes_fan_temp_speed_t

struct zes_fan_temp_speed_t

Fan temperature/speed pair.

Public Members

uint32_t temperature

[in,out] Temperature in degrees Celsius.

zes_fan_speed_t speed

[in,out] The speed of the fan

zes_fan_speed_table_t

struct zes_fan_speed_table_t

Fan speed table.

Public Members

int32_t numPoints

[in,out] The number of valid points in the fan speed table. 0 means that there is no fan speed table configured. -1 means that a fan speed table is not supported by the hardware.

zes_fan_temp_speed_t table[ZES_FAN_TEMP_SPEED_PAIR_COUNT]

[in,out] Array of temperature/fan speed pairs. The table is ordered based on temperature from lowest to highest.

zes_fan_properties_t

struct zes_fan_properties_t

Fan properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

ze_bool_t canControl

[out] Indicates if software can control the fan speed assuming the user has permissions

uint32_t supportedModes

[out] Bitfield of supported fan configuration modes (1<<zes_fan_speed_mode_t)

uint32_t supportedUnits

[out] Bitfield of supported fan speed units (1<<zes_fan_speed_units_t)

int32_t maxRPM

[out] The maximum RPM of the fan. A value of -1 means that this property is unknown.

int32_t maxPoints

[out] The maximum number of points in the fan temp/speed table. A value of -1 means that this fan doesn’t support providing a temp/speed table.

zes_fan_config_t

struct zes_fan_config_t

Fan configuration.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_fan_speed_mode_t mode

[in,out] The fan speed mode (fixed, temp-speed table)

zes_fan_speed_t speedFixed

[in,out] The current fixed fan speed setting

zes_fan_speed_table_t speedTable

[out] A table containing temperature/speed pairs

Firmware

Firmware Functions

zesDeviceEnumFirmwares

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumFirmwares(zes_device_handle_t hDevice, uint32_t *pCount, zes_firmware_handle_t *phFirmware)

Get handle of firmwares.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phFirmware – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesFirmwareGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesFirmwareGetProperties(zes_firmware_handle_t hFirmware, zes_firmware_properties_t *pProperties)

Get firmware properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFirmware – [in] Handle for the component.

  • pProperties – [in,out] Pointer to an array that will hold the properties of the firmware

Returns

zesFirmwareFlash

ZE_APIEXPORT ze_result_t ZE_APICALL zesFirmwareFlash(zes_firmware_handle_t hFirmware, void *pImage, uint32_t size)

Flash a new firmware image.

  • Any running workload must be gracefully closed before invoking this function.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • This is a non-blocking call. Application may call zesFirmwareGetFlashProgress to get completion status.

Parameters
  • hFirmware – [in] Handle for the component.

  • pImage – [in] Image of the new firmware to flash.

  • size – [in] Size of the flash image.

Returns

zesFirmwareGetFlashProgress

ZE_APIEXPORT ze_result_t ZE_APICALL zesFirmwareGetFlashProgress(zes_firmware_handle_t hFirmware, uint32_t *pCompletionPercent)

Get Firmware Flash Progress.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFirmware – [in] Handle for the component.

  • pCompletionPercent – [in,out] Pointer to the Completion Percentage of Firmware Update

Returns

Firmware Structures

zes_firmware_properties_t

struct zes_firmware_properties_t

Firmware properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

ze_bool_t canControl

[out] Indicates if software can flash the firmware assuming the user has permissions

char name[ZES_STRING_PROPERTY_SIZE]

[out] NULL terminated string value. The string “unknown” will be returned if this property cannot be determined.

char version[ZES_STRING_PROPERTY_SIZE]

[out] NULL terminated string value. The string “unknown” will be returned if this property cannot be determined.

Frequency

Frequency Functions

zesDeviceEnumFrequencyDomains

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumFrequencyDomains(zes_device_handle_t hDevice, uint32_t *pCount, zes_freq_handle_t *phFrequency)

Get handle of frequency domains.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phFrequency – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesFrequencyGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyGetProperties(zes_freq_handle_t hFrequency, zes_freq_properties_t *pProperties)

Get frequency properties - available frequencies.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFrequency – [in] Handle for the component.

  • pProperties – [in,out] The frequency properties for the specified domain.

Returns

zesFrequencyGetAvailableClocks

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyGetAvailableClocks(zes_freq_handle_t hFrequency, uint32_t *pCount, double *phFrequency)

Get available non-overclocked hardware clock frequencies for the frequency domain.

  • The list of available frequencies is returned in order of slowest to fastest.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFrequency – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of frequencies. if count is zero, then the driver shall update the value with the total number of frequencies that are available. if count is greater than the number of frequencies that are available, then the driver shall update the value with the correct number of frequencies.

  • phFrequency – [in,out][optional][range(0, *pCount)] array of frequencies in units of MHz and sorted from slowest to fastest. if count is less than the number of frequencies that are available, then the driver shall only retrieve that number of frequencies.

Returns

zesFrequencyGetRange

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyGetRange(zes_freq_handle_t hFrequency, zes_freq_range_t *pLimits)

Get current frequency limits.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFrequency – [in] Handle for the component.

  • pLimits – [in,out] The range between which the hardware can operate for the specified domain.

Returns

zesFrequencySetRange

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencySetRange(zes_freq_handle_t hFrequency, const zes_freq_range_t *pLimits)

Set frequency range between which the hardware can operate.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFrequency – [in] Handle for the component.

  • pLimits – [in] The limits between which the hardware can operate for the specified domain.

Returns

zesFrequencyGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyGetState(zes_freq_handle_t hFrequency, zes_freq_state_t *pState)

Get current frequency state - frequency request, actual frequency, TDP limits.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFrequency – [in] Handle for the component.

  • pState – [in,out] Frequency state for the specified domain.

Returns

zesFrequencyGetThrottleTime

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyGetThrottleTime(zes_freq_handle_t hFrequency, zes_freq_throttle_time_t *pThrottleTime)

Get frequency throttle time.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hFrequency – [in] Handle for the component.

  • pThrottleTime – [in,out] Will contain a snapshot of the throttle time counters for the specified domain.

Returns

zesFrequencyOcGetCapabilities

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcGetCapabilities(zes_freq_handle_t hFrequency, zes_oc_capabilities_t *pOcCapabilities)

Get the overclocking capabilities.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • pOcCapabilities – [in,out] Pointer to the capabilities structure.

Returns

zesFrequencyOcGetFrequencyTarget

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcGetFrequencyTarget(zes_freq_handle_t hFrequency, double *pCurrentOcFrequency)

Get the current overclocking frequency target, if extended moded is supported, will returned in 1 Mhz granularity.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • pCurrentOcFrequency – [out] Overclocking Frequency in MHz, if extended moded is supported, will returned in 1 Mhz granularity, else, in multiples of 50 Mhz. This cannot be greater than the maxOcFrequency member of zes_oc_capabilities_t.

Returns

zesFrequencyOcSetFrequencyTarget

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcSetFrequencyTarget(zes_freq_handle_t hFrequency, double CurrentOcFrequency)

Set the current overclocking frequency target, if extended moded is supported, can be set in 1 Mhz granularity.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • CurrentOcFrequency – [in] Overclocking Frequency in MHz, if extended moded is supported, it could be set in 1 Mhz granularity, else, in multiples of 50 Mhz. This cannot be greater than the maxOcFrequency member of zes_oc_capabilities_t.

Returns

zesFrequencyOcGetVoltageTarget

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcGetVoltageTarget(zes_freq_handle_t hFrequency, double *pCurrentVoltageTarget, double *pCurrentVoltageOffset)

Get the current overclocking voltage settings.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • pCurrentVoltageTarget – [out] Overclock voltage in Volts. This cannot be greater than the maxOcVoltage member of zes_oc_capabilities_t.

  • pCurrentVoltageOffset – [out] This voltage offset is applied to all points on the voltage/frequency curve, including the new overclock voltageTarget. Valid range is between the minOcVoltageOffset and maxOcVoltageOffset members of zes_oc_capabilities_t.

Returns

zesFrequencyOcSetVoltageTarget

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcSetVoltageTarget(zes_freq_handle_t hFrequency, double CurrentVoltageTarget, double CurrentVoltageOffset)

Set the current overclocking voltage settings.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • CurrentVoltageTarget – [in] Overclock voltage in Volts. This cannot be greater than the maxOcVoltage member of zes_oc_capabilities_t.

  • CurrentVoltageOffset – [in] This voltage offset is applied to all points on the voltage/frequency curve, include the new overclock voltageTarget. Valid range is between the minOcVoltageOffset and maxOcVoltageOffset members of zes_oc_capabilities_t.

Returns

zesFrequencyOcSetMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcSetMode(zes_freq_handle_t hFrequency, zes_oc_mode_t CurrentOcMode)

Set the current overclocking mode.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • CurrentOcMode – [in] Current Overclocking Mode zes_oc_mode_t.

Returns

zesFrequencyOcGetMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcGetMode(zes_freq_handle_t hFrequency, zes_oc_mode_t *pCurrentOcMode)

Get the current overclocking mode.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • pCurrentOcMode – [out] Current Overclocking Mode zes_oc_mode_t.

Returns

zesFrequencyOcGetIccMax

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcGetIccMax(zes_freq_handle_t hFrequency, double *pOcIccMax)

Get the maximum current limit setting.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • pOcIccMax – [in,out] Will contain the maximum current limit in Amperes on successful return.

Returns

zesFrequencyOcSetIccMax

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcSetIccMax(zes_freq_handle_t hFrequency, double ocIccMax)

Change the maximum current limit setting.

  • Setting ocIccMax to 0.0 will return the value to the factory default.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • ocIccMax – [in] The new maximum current limit in Amperes.

Returns

zesFrequencyOcGetTjMax

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcGetTjMax(zes_freq_handle_t hFrequency, double *pOcTjMax)

Get the maximum temperature limit setting.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • pOcTjMax – [in,out] Will contain the maximum temperature limit in degrees Celsius on successful return.

Returns

zesFrequencyOcSetTjMax

ZE_APIEXPORT ze_result_t ZE_APICALL zesFrequencyOcSetTjMax(zes_freq_handle_t hFrequency, double ocTjMax)

Change the maximum temperature limit setting.

  • Setting ocTjMax to 0.0 will return the value to the factory default.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hFrequency – [in] Handle for the component.

  • ocTjMax – [in] The new maximum temperature limit in degrees Celsius.

Returns

Frequency Enums

zes_freq_domain_t

enum zes_freq_domain_t

Frequency domains.

Values:

enumerator ZES_FREQ_DOMAIN_GPU

GPU Core Domain.

enumerator ZES_FREQ_DOMAIN_MEMORY

Local Memory Domain.

enumerator ZES_FREQ_DOMAIN_MEDIA

GPU Media Domain.

enumerator ZES_FREQ_DOMAIN_FORCE_UINT32

zes_freq_throttle_reason_flags_t

enum zes_freq_throttle_reason_flag_t

Values:

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_AVE_PWR_CAP

frequency throttled due to average power excursion (PL1)

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_BURST_PWR_CAP

frequency throttled due to burst power excursion (PL2)

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_CURRENT_LIMIT

frequency throttled due to current excursion (PL4)

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_THERMAL_LIMIT

frequency throttled due to thermal excursion (T > TjMax)

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_PSU_ALERT

frequency throttled due to power supply assertion

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_SW_RANGE

frequency throttled due to software supplied frequency range

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_HW_RANGE

frequency throttled due to a sub block that has a lower frequency range when it receives clocks

enumerator ZES_FREQ_THROTTLE_REASON_FLAG_FORCE_UINT32

zes_oc_mode_t

enum zes_oc_mode_t

Overclocking modes.

  • [DEPRECATED] No longer supported.

Values:

enumerator ZES_OC_MODE_OFF

Overclocking if off - hardware is running using factory default voltages/frequencies.

enumerator ZES_OC_MODE_OVERRIDE

Overclock override mode - In this mode, a fixed user-supplied voltage is applied independent of the frequency request. The maximum permitted frequency can also be increased. This mode disables INTERPOLATIVE and FIXED modes.

enumerator ZES_OC_MODE_INTERPOLATIVE

Overclock interpolative mode - In this mode, the voltage/frequency curve can be extended with a new voltage/frequency point that will be interpolated. The existing voltage/frequency points can also be offset (up or down) by a fixed voltage. This mode disables FIXED and OVERRIDE modes.

enumerator ZES_OC_MODE_FIXED

Overclocking fixed Mode - In this mode, hardware will disable most frequency throttling and lock the frequency and voltage at the specified overclock values. This mode disables OVERRIDE and INTERPOLATIVE modes. This mode can damage the part, most of the protections are disabled on this mode.

enumerator ZES_OC_MODE_FORCE_UINT32

Frequency Structures

zes_freq_properties_t

struct zes_freq_properties_t

Frequency properties.

  • Indicates if this frequency domain can be overclocked (if true, functions such as zesFrequencyOcSetFrequencyTarget() are supported).

  • The min/max hardware frequencies are specified for non-overclock configurations. For overclock configurations, use zesFrequencyOcGetFrequencyTarget() to determine the maximum frequency that can be requested.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_freq_domain_t type

[out] The hardware block that this frequency domain controls (GPU, memory, …)

ze_bool_t onSubdevice

[out] True if this resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

ze_bool_t canControl

[out] Indicates if software can control the frequency of this domain assuming the user has permissions

ze_bool_t isThrottleEventSupported

[out] Indicates if software can register to receive event ZES_EVENT_TYPE_FLAG_FREQ_THROTTLED

double min

[out] The minimum hardware clock frequency in units of MHz.

double max

[out] The maximum non-overclock hardware clock frequency in units of MHz.

zes_freq_range_t

struct zes_freq_range_t

Frequency range between which the hardware can operate.

  • When setting limits, they will be clamped to the hardware limits.

  • When setting limits, ensure that the max frequency is greater than or equal to the min frequency specified.

  • When setting limits to return to factory settings, specify -1 for both the min and max limit.

Public Members

double min

[in,out] The min frequency in MHz below which hardware frequency management will not request frequencies. On input, setting to 0 will permit the frequency to go down to the hardware minimum while setting to -1 will return the min frequency limit to the factory value (can be larger than the hardware min). On output, a negative value indicates that no external minimum frequency limit is in effect.

double max

[in,out] The max frequency in MHz above which hardware frequency management will not request frequencies. On input, setting to 0 or a very big number will permit the frequency to go all the way up to the hardware maximum while setting to -1 will return the max frequency to the factory value (which can be less than the hardware max). On output, a negative number indicates that no external maximum frequency limit is in effect.

zes_freq_state_t

struct zes_freq_state_t

Frequency state.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

double currentVoltage

[out] Current voltage in Volts. A negative value indicates that this property is not known.

double request

[out] The current frequency request in MHz. A negative value indicates that this property is not known.

double tdp

[out] The maximum frequency in MHz supported under the current TDP conditions. This fluctuates dynamically based on the power and thermal limits of the part. A negative value indicates that this property is not known.

double efficient

[out] The efficient minimum frequency in MHz. A negative value indicates that this property is not known.

double actual

[out] The resolved frequency in MHz. A negative value indicates that this property is not known.

zes_freq_throttle_reason_flags_t throttleReasons

[out] The reasons that the frequency is being limited by the hardware. Returns 0 (frequency not throttled) or a combination of zes_freq_throttle_reason_flag_t.

zes_freq_throttle_time_t

struct zes_freq_throttle_time_t

Frequency throttle time snapshot.

  • Percent time throttled is calculated by taking two snapshots (s1, s2) and using the equation: throttled = (s2.throttleTime - s1.throttleTime) / (s2.timestamp - s1.timestamp)

Public Members

uint64_t throttleTime

[out] The monotonic counter of time in microseconds that the frequency has been limited by the hardware.

uint64_t timestamp

[out] Microsecond timestamp when throttleTime was captured. This timestamp should only be used to calculate delta time between snapshots of this structure. Never take the delta of this timestamp with the timestamp from a different structure since they are not guaranteed to have the same base. The absolute value of the timestamp is only valid during within the application and may be different on the next execution.

zes_oc_capabilities_t

struct zes_oc_capabilities_t

Overclocking properties.

  • Provides all the overclocking capabilities and properties supported by the device for the frequency domain.

  • [DEPRECATED] No longer supported.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t isOcSupported

[out] Indicates if any overclocking features are supported on this frequency domain.

double maxFactoryDefaultFrequency

[out] Factory default non-overclock maximum frequency in Mhz.

double maxFactoryDefaultVoltage

[out] Factory default voltage used for the non-overclock maximum frequency in MHz.

double maxOcFrequency

[out] Maximum hardware overclocking frequency limit in Mhz.

double minOcVoltageOffset

[out] The minimum voltage offset that can be applied to the voltage/frequency curve. Note that this number can be negative.

double maxOcVoltageOffset

[out] The maximum voltage offset that can be applied to the voltage/frequency curve.

double maxOcVoltage

[out] The maximum overclock voltage that hardware supports.

ze_bool_t isTjMaxSupported

[out] Indicates if the maximum temperature limit (TjMax) can be changed for this frequency domain.

ze_bool_t isIccMaxSupported

[out] Indicates if the maximum current (IccMax) can be changed for this frequency domain.

ze_bool_t isHighVoltModeCapable

[out] Indicates if this frequency domains supports a feature to set very high voltages.

ze_bool_t isHighVoltModeEnabled

[out] Indicates if very high voltages are permitted on this frequency domain.

ze_bool_t isExtendedModeSupported

[out] Indicates if the extended overclocking features are supported. If this is supported, increments are on 1 Mhz basis.

ze_bool_t isFixedModeSupported

[out] Indicates if the fixed mode is supported. In this mode, hardware will disable most frequency throttling and lock the frequency and voltage at the specified overclock values.

Led

Led Functions

zesDeviceEnumLeds

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumLeds(zes_device_handle_t hDevice, uint32_t *pCount, zes_led_handle_t *phLed)

Get handle of LEDs.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phLed – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesLedGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesLedGetProperties(zes_led_handle_t hLed, zes_led_properties_t *pProperties)

Get LED properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hLed – [in] Handle for the component.

  • pProperties – [in,out] Will contain the properties of the LED.

Returns

zesLedGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesLedGetState(zes_led_handle_t hLed, zes_led_state_t *pState)

Get current state of a LED - on/off, color.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hLed – [in] Handle for the component.

  • pState – [in,out] Will contain the current state of the LED.

Returns

zesLedSetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesLedSetState(zes_led_handle_t hLed, ze_bool_t enable)

Turn the LED on/off.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hLed – [in] Handle for the component.

  • enable – [in] Set to TRUE to turn the LED on, FALSE to turn off.

Returns

zesLedSetColor

ZE_APIEXPORT ze_result_t ZE_APICALL zesLedSetColor(zes_led_handle_t hLed, const zes_led_color_t *pColor)

Set the color of the LED.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hLed – [in] Handle for the component.

  • pColor – [in] New color of the LED.

Returns

Led Structures

zes_led_properties_t

struct zes_led_properties_t

LED properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

ze_bool_t canControl

[out] Indicates if software can control the LED assuming the user has permissions

ze_bool_t haveRGB

[out] Indicates if the LED is RGB capable

zes_led_color_t

struct zes_led_color_t

LED color.

Public Members

double red

[in,out][range(0.0, 1.0)] The LED red value. On output, a value less than 0.0 indicates that the color is not known.

double green

[in,out][range(0.0, 1.0)] The LED green value. On output, a value less than 0.0 indicates that the color is not known.

double blue

[in,out][range(0.0, 1.0)] The LED blue value. On output, a value less than 0.0 indicates that the color is not known.

zes_led_state_t

struct zes_led_state_t

LED state.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t isOn

[out] Indicates if the LED is on or off

zes_led_color_t color

[out] Color of the LED

Memory

Memory Functions

zesDeviceEnumMemoryModules

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumMemoryModules(zes_device_handle_t hDevice, uint32_t *pCount, zes_mem_handle_t *phMemory)

Get handle of memory modules.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phMemory – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesMemoryGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesMemoryGetProperties(zes_mem_handle_t hMemory, zes_mem_properties_t *pProperties)

Get memory properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hMemory – [in] Handle for the component.

  • pProperties – [in,out] Will contain memory properties.

Returns

zesMemoryGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesMemoryGetState(zes_mem_handle_t hMemory, zes_mem_state_t *pState)

Get memory state - health, allocated.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hMemory – [in] Handle for the component.

  • pState – [in,out] Will contain the current health and allocated memory.

Returns

zesMemoryGetBandwidth

ZE_APIEXPORT ze_result_t ZE_APICALL zesMemoryGetBandwidth(zes_mem_handle_t hMemory, zes_mem_bandwidth_t *pBandwidth)

Get memory bandwidth.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hMemory – [in] Handle for the component.

  • pBandwidth – [in,out] Will contain the total number of bytes read from and written to memory, as well as the current maximum bandwidth.

Returns

Memory Enums

zes_mem_type_t

enum zes_mem_type_t

Memory module types.

Values:

enumerator ZES_MEM_TYPE_HBM

HBM memory.

enumerator ZES_MEM_TYPE_DDR

DDR memory.

enumerator ZES_MEM_TYPE_DDR3

DDR3 memory.

enumerator ZES_MEM_TYPE_DDR4

DDR4 memory.

enumerator ZES_MEM_TYPE_DDR5

DDR5 memory.

enumerator ZES_MEM_TYPE_LPDDR

LPDDR memory.

enumerator ZES_MEM_TYPE_LPDDR3

LPDDR3 memory.

enumerator ZES_MEM_TYPE_LPDDR4

LPDDR4 memory.

enumerator ZES_MEM_TYPE_LPDDR5

LPDDR5 memory.

enumerator ZES_MEM_TYPE_SRAM

SRAM memory.

enumerator ZES_MEM_TYPE_L1

L1 cache.

enumerator ZES_MEM_TYPE_L3

L3 cache.

enumerator ZES_MEM_TYPE_GRF

Execution unit register file.

enumerator ZES_MEM_TYPE_SLM

Execution unit shared local memory.

enumerator ZES_MEM_TYPE_GDDR4

GDDR4 memory.

enumerator ZES_MEM_TYPE_GDDR5

GDDR5 memory.

enumerator ZES_MEM_TYPE_GDDR5X

GDDR5X memory.

enumerator ZES_MEM_TYPE_GDDR6

GDDR6 memory.

enumerator ZES_MEM_TYPE_GDDR6X

GDDR6X memory.

enumerator ZES_MEM_TYPE_GDDR7

GDDR7 memory.

enumerator ZES_MEM_TYPE_FORCE_UINT32

zes_mem_loc_t

enum zes_mem_loc_t

Memory module location.

Values:

enumerator ZES_MEM_LOC_SYSTEM

System memory.

enumerator ZES_MEM_LOC_DEVICE

On board local device memory.

enumerator ZES_MEM_LOC_FORCE_UINT32

zes_mem_health_t

enum zes_mem_health_t

Memory health.

Values:

enumerator ZES_MEM_HEALTH_UNKNOWN

The memory health cannot be determined.

enumerator ZES_MEM_HEALTH_OK

All memory channels are healthy.

enumerator ZES_MEM_HEALTH_DEGRADED

Excessive correctable errors have been detected on one or more channels. Device should be reset.

enumerator ZES_MEM_HEALTH_CRITICAL

Operating with reduced memory to cover banks with too many uncorrectable errors.

enumerator ZES_MEM_HEALTH_REPLACE

Device should be replaced due to excessive uncorrectable errors.

enumerator ZES_MEM_HEALTH_FORCE_UINT32

Memory Structures

zes_mem_properties_t

struct zes_mem_properties_t

Memory properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_mem_type_t type

[out] The memory type

ze_bool_t onSubdevice

[out] True if this resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

zes_mem_loc_t location

[out] Location of this memory (system, device)

uint64_t physicalSize

[out] Physical memory size in bytes. A value of 0 indicates that this property is not known. However, a call to zesMemoryGetState() will correctly return the total size of usable memory.

int32_t busWidth

[out] Width of the memory bus. A value of -1 means that this property is unknown.

int32_t numChannels

[out] The number of memory channels. A value of -1 means that this property is unknown.

zes_mem_state_t

struct zes_mem_state_t

Memory state - health, allocated.

  • Percent allocation is given by 100 * (size - free / size.

  • Percent free is given by 100 * free / size.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_mem_health_t health

[out] Indicates the health of the memory

uint64_t free

[out] The free memory in bytes

uint64_t size

[out] The total allocatable memory in bytes (can be less than the physicalSize member of zes_mem_properties_t)

zes_mem_bandwidth_t

struct zes_mem_bandwidth_t

Memory bandwidth.

  • Percent bandwidth is calculated by taking two snapshots (s1, s2) and using the equation: bw = 10^6 * ((s2.readCounter - s1.readCounter) + (s2.writeCounter - s1.writeCounter)) / (s2.maxBandwidth * (s2.timestamp - s1.timestamp))

  • Counter can roll over and rollover needs to be handled by comparing the current read against the previous read

  • Counter is a 32 byte transaction count, which means the calculated delta (delta = current_value - previous_value or delta = 2^32 - previous_value + current_value in case of rollover) needs to be multiplied by 32 to get delta between samples in actual byte count

Public Members

uint64_t readCounter

[out] Total bytes read from memory

uint64_t writeCounter

[out] Total bytes written to memory

uint64_t maxBandwidth

[out] Current maximum bandwidth in units of bytes/sec

uint64_t timestamp

[out] The timestamp in microseconds when these measurements were sampled. This timestamp should only be used to calculate delta time between snapshots of this structure. Never take the delta of this timestamp with the timestamp from a different structure since they are not guaranteed to have the same base. The absolute value of the timestamp is only valid during within the application and may be different on the next execution.

zes_mem_ext_bandwidth_t

struct zes_mem_ext_bandwidth_t

Extension properties for Memory bandwidth.

  • Number of counter bits

  • [DEPRECATED] No longer supported.

Public Members

uint32_t memoryTimestampValidBits

[out] Returns the number of valid bits in the timestamp values

Performance

Performance Functions

zesDeviceEnumPerformanceFactorDomains

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumPerformanceFactorDomains(zes_device_handle_t hDevice, uint32_t *pCount, zes_perf_handle_t *phPerf)

Get handles to accelerator domains whose performance can be optimized via a Performance Factor.

  • A Performance Factor should be tuned for each workload.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phPerf – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesPerformanceFactorGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesPerformanceFactorGetProperties(zes_perf_handle_t hPerf, zes_perf_properties_t *pProperties)

Get properties about a Performance Factor domain.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPerf – [in] Handle for the Performance Factor domain.

  • pProperties – [in,out] Will contain information about the specified Performance Factor domain.

Returns

zesPerformanceFactorGetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesPerformanceFactorGetConfig(zes_perf_handle_t hPerf, double *pFactor)

Get current Performance Factor for a given domain.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPerf – [in] Handle for the Performance Factor domain.

  • pFactor – [in,out] Will contain the actual Performance Factor being used by the hardware (may not be the same as the requested Performance Factor).

Returns

zesPerformanceFactorSetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesPerformanceFactorSetConfig(zes_perf_handle_t hPerf, double factor)

Change the performance factor for a domain.

  • The Performance Factor is a number between 0 and 100.

  • A Performance Factor is a hint to the hardware. Depending on the hardware, the request may not be granted. Follow up this function with a call to zesPerformanceFactorGetConfig() to determine the actual factor being used by the hardware.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPerf – [in] Handle for the Performance Factor domain.

  • factor – [in] The new Performance Factor.

Returns

Performance Structures

zes_perf_properties_t

struct zes_perf_properties_t

Static information about a Performance Factor domain.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if this Performance Factor affects accelerators located on a sub-device

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

zes_engine_type_flags_t engines

[out] Bitfield of accelerator engine types that are affected by this Performance Factor.

Power

Power Functions

zesDeviceEnumPowerDomains

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumPowerDomains(zes_device_handle_t hDevice, uint32_t *pCount, zes_pwr_handle_t *phPower)

Get handle of power domains.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phPower – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesDeviceGetCardPowerDomain

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceGetCardPowerDomain(zes_device_handle_t hDevice, zes_pwr_handle_t *phPower)

Get handle of the PCIe card-level power.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • phPower – [in,out] power domain handle for the entire PCIe card.

Returns

zesPowerGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerGetProperties(zes_pwr_handle_t hPower, zes_power_properties_t *pProperties)

Get properties related to a power domain.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPower – [in] Handle for the component.

  • pProperties – [in,out] Structure that will contain property data.

Returns

zesPowerGetEnergyCounter

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerGetEnergyCounter(zes_pwr_handle_t hPower, zes_power_energy_counter_t *pEnergy)

Get energy counter.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPower – [in] Handle for the component.

  • pEnergy – [in,out] Will contain the latest snapshot of the energy counter and timestamp when the last counter value was measured.

Returns

zesPowerGetLimits

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerGetLimits(zes_pwr_handle_t hPower, zes_power_sustained_limit_t *pSustained, zes_power_burst_limit_t *pBurst, zes_power_peak_limit_t *pPeak)

Get power limits.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] Use zesPowerGetLimitsExt.

Parameters
  • hPower – [in] Handle for the component.

  • pSustained – [in,out][optional] The sustained power limit. If this is null, the current sustained power limits will not be returned.

  • pBurst – [in,out][optional] The burst power limit. If this is null, the current peak power limits will not be returned.

  • pPeak – [in,out][optional] The peak power limit. If this is null, the peak power limits will not be returned.

Returns

zesPowerSetLimits

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerSetLimits(zes_pwr_handle_t hPower, const zes_power_sustained_limit_t *pSustained, const zes_power_burst_limit_t *pBurst, const zes_power_peak_limit_t *pPeak)

Set power limits.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] Use zesPowerSetLimitsExt.

Parameters
  • hPower – [in] Handle for the component.

  • pSustained – [in][optional] The sustained power limit. If this is null, no changes will be made to the sustained power limits.

  • pBurst – [in][optional] The burst power limit. If this is null, no changes will be made to the burst power limits.

  • pPeak – [in][optional] The peak power limit. If this is null, no changes will be made to the peak power limits.

Returns

zesPowerGetEnergyThreshold

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerGetEnergyThreshold(zes_pwr_handle_t hPower, zes_energy_threshold_t *pThreshold)

Get energy threshold.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPower – [in] Handle for the component.

  • pThreshold – [in,out] Returns information about the energy threshold setting - enabled/energy threshold/process ID.

Returns

zesPowerSetEnergyThreshold

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerSetEnergyThreshold(zes_pwr_handle_t hPower, double threshold)

Set energy threshold.

  • An event ZES_EVENT_TYPE_FLAG_ENERGY_THRESHOLD_CROSSED will be generated when the delta energy consumed starting from this call exceeds the specified threshold. Use the function zesDeviceEventRegister() to start receiving the event.

  • Only one running process can control the energy threshold at a given time. If another process attempts to change the energy threshold, the error ZE_RESULT_ERROR_NOT_AVAILABLE will be returned. The function zesPowerGetEnergyThreshold() to determine the process ID currently controlling this setting.

  • Calling this function will remove any pending energy thresholds and start counting from the time of this call.

  • Once the energy threshold has been reached and the event generated, the threshold is automatically removed. It is up to the application to request a new threshold.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPower – [in] Handle for the component.

  • threshold – [in] The energy threshold to be set in joules.

Returns

Power Enums

zes_power_domain_t

enum zes_power_domain_t

Power Domain.

Values:

enumerator ZES_POWER_DOMAIN_UNKNOWN

The PUnit power domain level cannot be determined.

enumerator ZES_POWER_DOMAIN_CARD

The PUnit power domain is a card-level power domain.

enumerator ZES_POWER_DOMAIN_PACKAGE

The PUnit power domain is a package-level power domain.

enumerator ZES_POWER_DOMAIN_STACK

The PUnit power domain is a stack-level power domain.

enumerator ZES_POWER_DOMAIN_MEMORY

The PUnit power domain is a memory-level power domain.

enumerator ZES_POWER_DOMAIN_GPU

The PUnit power domain is a GPU-level power domain.

enumerator ZES_POWER_DOMAIN_FORCE_UINT32

zes_power_level_t

enum zes_power_level_t

Power Level Type.

Values:

enumerator ZES_POWER_LEVEL_UNKNOWN

The PUnit power monitoring duration cannot be determined.

enumerator ZES_POWER_LEVEL_SUSTAINED

The PUnit determines effective power draw by computing a moving average of the actual power draw over a time interval (longer than BURST).

enumerator ZES_POWER_LEVEL_BURST

The PUnit determines effective power draw by computing a moving average of the actual power draw over a time interval (longer than PEAK).

enumerator ZES_POWER_LEVEL_PEAK

The PUnit determines effective power draw by computing a moving average of the actual power draw over a very short time interval.

enumerator ZES_POWER_LEVEL_INSTANTANEOUS

The PUnit predicts effective power draw using the current device configuration (frequency, voltage, etc…) & throttles proactively to stay within the specified limit.

enumerator ZES_POWER_LEVEL_FORCE_UINT32

zes_power_source_t

enum zes_power_source_t

Power Source Type.

Values:

enumerator ZES_POWER_SOURCE_ANY

Limit active no matter whether the power source is mains powered or battery powered.

enumerator ZES_POWER_SOURCE_MAINS

Limit active only when the device is mains powered.

enumerator ZES_POWER_SOURCE_BATTERY

Limit active only when the device is battery powered.

enumerator ZES_POWER_SOURCE_FORCE_UINT32

zes_limit_unit_t

enum zes_limit_unit_t

Limit Unit.

Values:

enumerator ZES_LIMIT_UNIT_UNKNOWN

The PUnit power monitoring unit cannot be determined.

enumerator ZES_LIMIT_UNIT_CURRENT

The limit is specified in milliamperes of current drawn.

enumerator ZES_LIMIT_UNIT_POWER

The limit is specified in milliwatts of power generated.

enumerator ZES_LIMIT_UNIT_FORCE_UINT32

Power Structures

zes_power_properties_t

struct zes_power_properties_t

Properties related to device power settings.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if this resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

ze_bool_t canControl

[out] Software can change the power limits of this domain assuming the user has permissions.

ze_bool_t isEnergyThresholdSupported

[out] Indicates if this power domain supports the energy threshold event (ZES_EVENT_TYPE_FLAG_ENERGY_THRESHOLD_CROSSED).

int32_t defaultLimit

[out] (Deprecated) The factory default TDP power limit of the part in milliwatts. A value of -1 means that this is not known.

int32_t minLimit

[out] (Deprecated) The minimum power limit in milliwatts that can be requested. A value of -1 means that this is not known.

int32_t maxLimit

[out] (Deprecated) The maximum power limit in milliwatts that can be requested. A value of -1 means that this is not known.

zes_power_energy_counter_t

struct zes_power_energy_counter_t

Energy counter snapshot.

  • Average power is calculated by taking two snapshots (s1, s2) and using the equation: PowerWatts = (s2.energy - s1.energy) / (s2.timestamp - s1.timestamp)

Public Members

uint64_t energy

[out] The monotonic energy counter in microjoules.

uint64_t timestamp

[out] Microsecond timestamp when energy was captured. This timestamp should only be used to calculate delta time between snapshots of this structure. Never take the delta of this timestamp with the timestamp from a different structure since they are not guaranteed to have the same base. The absolute value of the timestamp is only valid during within the application and may be different on the next execution.

zes_power_sustained_limit_t

struct zes_power_sustained_limit_t

Sustained power limits.

  • The power controller (Punit) will throttle the operating frequency if the power averaged over a window (typically seconds) exceeds this limit.

  • [DEPRECATED] No longer supported.

Public Members

ze_bool_t enabled

[in,out] indicates if the limit is enabled (true) or ignored (false)

int32_t power

[in,out] power limit in milliwatts

int32_t interval

[in,out] power averaging window (Tau) in milliseconds

zes_power_burst_limit_t

struct zes_power_burst_limit_t

Burst power limit.

  • The power controller (Punit) will throttle the operating frequency of the device if the power averaged over a few milliseconds exceeds a limit known as PL2. Typically PL2 > PL1 so that it permits the frequency to burst higher for short periods than would be otherwise permitted by PL1.

  • [DEPRECATED] No longer supported.

Public Members

ze_bool_t enabled

[in,out] indicates if the limit is enabled (true) or ignored (false)

int32_t power

[in,out] power limit in milliwatts

zes_power_peak_limit_t

struct zes_power_peak_limit_t

Peak power limit.

  • The power controller (Punit) will reactively/proactively throttle the operating frequency of the device when the instantaneous/100usec power exceeds this limit. The limit is known as PL4 or Psys. It expresses the maximum power that can be drawn from the power supply.

  • If this power limit is removed or set too high, the power supply will generate an interrupt when it detects an overcurrent condition and the power controller will throttle the device frequencies down to min. It is thus better to tune the PL4 value in order to avoid such excursions.

  • [DEPRECATED] No longer supported.

Public Members

int32_t powerAC

[in,out] power limit in milliwatts for the AC power source.

int32_t powerDC

[in,out] power limit in milliwatts for the DC power source. On input, this is ignored if the product does not have a battery. On output, this will be -1 if the product does not have a battery.

zes_energy_threshold_t

struct zes_energy_threshold_t

Energy threshold.

  • .

Public Members

ze_bool_t enable

[in,out] Indicates if the energy threshold is enabled.

double threshold

[in,out] The energy threshold in Joules. Will be 0.0 if no threshold has been set.

uint32_t processId

[in,out] The host process ID that set the energy threshold. Will be 0xFFFFFFFF if no threshold has been set.

Psu

Psu Functions

zesDeviceEnumPsus

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumPsus(zes_device_handle_t hDevice, uint32_t *pCount, zes_psu_handle_t *phPsu)

Get handle of power supplies.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phPsu – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesPsuGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesPsuGetProperties(zes_psu_handle_t hPsu, zes_psu_properties_t *pProperties)

Get power supply properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPsu – [in] Handle for the component.

  • pProperties – [in,out] Will contain the properties of the power supply.

Returns

zesPsuGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesPsuGetState(zes_psu_handle_t hPsu, zes_psu_state_t *pState)

Get current power supply state.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPsu – [in] Handle for the component.

  • pState – [in,out] Will contain the current state of the power supply.

Returns

Psu Enums

zes_psu_voltage_status_t

enum zes_psu_voltage_status_t

PSU voltage status.

Values:

enumerator ZES_PSU_VOLTAGE_STATUS_UNKNOWN

The status of the power supply voltage controllers cannot be determined

enumerator ZES_PSU_VOLTAGE_STATUS_NORMAL

No unusual voltages have been detected.

enumerator ZES_PSU_VOLTAGE_STATUS_OVER

Over-voltage has occurred.

enumerator ZES_PSU_VOLTAGE_STATUS_UNDER

Under-voltage has occurred.

enumerator ZES_PSU_VOLTAGE_STATUS_FORCE_UINT32

Psu Structures

zes_psu_properties_t

struct zes_psu_properties_t

Static properties of the power supply.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

ze_bool_t haveFan

[out] True if the power supply has a fan

int32_t ampLimit

[out] The maximum electrical current in milliamperes that can be drawn. A value of -1 indicates that this property cannot be determined.

zes_psu_state_t

struct zes_psu_state_t

Dynamic state of the power supply.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_psu_voltage_status_t voltStatus

[out] The current PSU voltage status

ze_bool_t fanFailed

[out] Indicates if the fan has failed

int32_t temperature

[out] Read the current heatsink temperature in degrees Celsius. A value of -1 indicates that this property cannot be determined.

int32_t current

[out] The amps being drawn in milliamperes. A value of -1 indicates that this property cannot be determined.

Ras

Ras Functions

zesDeviceEnumRasErrorSets

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumRasErrorSets(zes_device_handle_t hDevice, uint32_t *pCount, zes_ras_handle_t *phRas)

Get handle of all RAS error sets on a device.

  • A RAS error set is a collection of RAS error counters of a given type (correctable/uncorrectable) from hardware blocks contained within a sub-device or within the device.

  • A device without sub-devices will typically return two handles, one for correctable errors sets and one for uncorrectable error sets.

  • A device with sub-devices will return RAS error sets for each sub-device and possibly RAS error sets for hardware blocks outside the sub-devices.

  • If the function completes successfully but pCount is set to 0, RAS features are not available/enabled on this device.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phRas – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesRasGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesRasGetProperties(zes_ras_handle_t hRas, zes_ras_properties_t *pProperties)

Get RAS properties of a given RAS error set - this enables discovery of the type of RAS error set (correctable/uncorrectable) and if located on a sub-device.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hRas – [in] Handle for the component.

  • pProperties – [in,out] Structure describing RAS properties

Returns

zesRasGetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesRasGetConfig(zes_ras_handle_t hRas, zes_ras_config_t *pConfig)

Get RAS error thresholds that control when RAS events are generated.

  • The driver maintains counters for all RAS error sets and error categories. Events are generated when errors occur. The configuration enables setting thresholds to limit when events are sent.

  • When a particular RAS correctable error counter exceeds the configured threshold, the event ZES_EVENT_TYPE_FLAG_RAS_CORRECTABLE_ERRORS will be triggered.

  • When a particular RAS uncorrectable error counter exceeds the configured threshold, the event ZES_EVENT_TYPE_FLAG_RAS_UNCORRECTABLE_ERRORS will be triggered.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hRas – [in] Handle for the component.

  • pConfig – [in,out] Will be populed with the current RAS configuration - thresholds used to trigger events

Returns

zesRasSetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesRasSetConfig(zes_ras_handle_t hRas, const zes_ras_config_t *pConfig)

Set RAS error thresholds that control when RAS events are generated.

  • The driver maintains counters for all RAS error sets and error categories. Events are generated when errors occur. The configuration enables setting thresholds to limit when events are sent.

  • When a particular RAS correctable error counter exceeds the specified threshold, the event ZES_EVENT_TYPE_FLAG_RAS_CORRECTABLE_ERRORS will be generated.

  • When a particular RAS uncorrectable error counter exceeds the specified threshold, the event ZES_EVENT_TYPE_FLAG_RAS_UNCORRECTABLE_ERRORS will be generated.

  • Call zesRasGetState() and set the clear flag to true to restart event generation once counters have exceeded thresholds.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hRas – [in] Handle for the component.

  • pConfig – [in] Change the RAS configuration - thresholds used to trigger events

Returns

zesRasGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesRasGetState(zes_ras_handle_t hRas, ze_bool_t clear, zes_ras_state_t *pState)

Get the current value of RAS error counters for a particular error set.

  • Clearing errors will affect other threads/applications - the counter values will start from zero.

  • Clearing errors requires write permissions.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hRas – [in] Handle for the component.

  • clear – [in] Set to 1 to clear the counters of this type

  • pState – [in,out] Breakdown of where errors have occurred

Returns

Ras Enums

zes_ras_error_type_t

enum zes_ras_error_type_t

RAS error type.

Values:

enumerator ZES_RAS_ERROR_TYPE_CORRECTABLE

Errors were corrected by hardware.

enumerator ZES_RAS_ERROR_TYPE_UNCORRECTABLE

Error were not corrected.

enumerator ZES_RAS_ERROR_TYPE_FORCE_UINT32

zes_ras_error_cat_t

enum zes_ras_error_cat_t

RAS error categories.

Values:

enumerator ZES_RAS_ERROR_CAT_RESET

The number of accelerator engine resets attempted by the driver.

enumerator ZES_RAS_ERROR_CAT_PROGRAMMING_ERRORS

The number of hardware exceptions generated by the way workloads have programmed the hardware

enumerator ZES_RAS_ERROR_CAT_DRIVER_ERRORS

The number of low level driver communication errors have occurred.

enumerator ZES_RAS_ERROR_CAT_COMPUTE_ERRORS

The number of errors that have occurred in the compute accelerator hardware

enumerator ZES_RAS_ERROR_CAT_NON_COMPUTE_ERRORS

The number of errors that have occurred in the fixed-function accelerator hardware

enumerator ZES_RAS_ERROR_CAT_CACHE_ERRORS

The number of errors that have occurred in caches (L1/L3/register file/shared local memory/sampler)

enumerator ZES_RAS_ERROR_CAT_DISPLAY_ERRORS

The number of errors that have occurred in the display.

enumerator ZES_RAS_ERROR_CAT_FORCE_UINT32

Ras Structures

zes_ras_properties_t

struct zes_ras_properties_t

RAS properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_ras_error_type_t type

[out] The type of RAS error

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

zes_ras_state_t

struct zes_ras_state_t

RAS error details.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint64_t category[ZES_MAX_RAS_ERROR_CATEGORY_COUNT]

[in][out] Breakdown of error by category

zes_ras_config_t

struct zes_ras_config_t

RAS error configuration - thresholds used for triggering RAS events (ZES_EVENT_TYPE_FLAG_RAS_CORRECTABLE_ERRORS, ZES_EVENT_TYPE_FLAG_RAS_UNCORRECTABLE_ERRORS)

  • The driver maintains a total counter which is updated every time a hardware block covered by the corresponding RAS error set notifies that an error has occurred. When this total count goes above the totalThreshold specified below, a RAS event is triggered.

  • The driver also maintains a counter for each category of RAS error (see zes_ras_state_t for a breakdown). Each time a hardware block of that category notifies that an error has occurred, that corresponding category counter is updated. When it goes above the threshold specified in detailedThresholds, a RAS event is triggered.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint64_t totalThreshold

[in,out] If the total RAS errors exceeds this threshold, the event will be triggered. A value of 0ULL disables triggering the event based on the total counter.

zes_ras_state_t detailedThresholds

[in,out] If the RAS errors for each category exceed the threshold for that category, the event will be triggered. A value of 0ULL will disable an event being triggered for that category.

Scheduler

Scheduler Functions

zesDeviceEnumSchedulers

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumSchedulers(zes_device_handle_t hDevice, uint32_t *pCount, zes_sched_handle_t *phScheduler)

Returns handles to scheduler components.

  • Each scheduler component manages the distribution of work across one or more accelerator engines.

  • If an application wishes to change the scheduler behavior for all accelerator engines of a specific type (e.g. compute), it should select all the handles where the engines member zes_sched_properties_t contains that type.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phScheduler – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesSchedulerGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerGetProperties(zes_sched_handle_t hScheduler, zes_sched_properties_t *pProperties)

Get properties related to a scheduler component.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hScheduler – [in] Handle for the component.

  • pProperties – [in,out] Structure that will contain property data.

Returns

zesSchedulerGetCurrentMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerGetCurrentMode(zes_sched_handle_t hScheduler, zes_sched_mode_t *pMode)

Get current scheduling mode in effect on a scheduler component.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hScheduler – [in] Sysman handle for the component.

  • pMode – [in,out] Will contain the current scheduler mode.

Returns

zesSchedulerGetTimeoutModeProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerGetTimeoutModeProperties(zes_sched_handle_t hScheduler, ze_bool_t getDefaults, zes_sched_timeout_properties_t *pConfig)

Get scheduler config for mode ZES_SCHED_MODE_TIMEOUT.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hScheduler – [in] Sysman handle for the component.

  • getDefaults – [in] If TRUE, the driver will return the system default properties for this mode, otherwise it will return the current properties.

  • pConfig – [in,out] Will contain the current parameters for this mode.

Returns

zesSchedulerGetTimesliceModeProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerGetTimesliceModeProperties(zes_sched_handle_t hScheduler, ze_bool_t getDefaults, zes_sched_timeslice_properties_t *pConfig)

Get scheduler config for mode ZES_SCHED_MODE_TIMESLICE.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hScheduler – [in] Sysman handle for the component.

  • getDefaults – [in] If TRUE, the driver will return the system default properties for this mode, otherwise it will return the current properties.

  • pConfig – [in,out] Will contain the current parameters for this mode.

Returns

zesSchedulerSetTimeoutMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerSetTimeoutMode(zes_sched_handle_t hScheduler, zes_sched_timeout_properties_t *pProperties, ze_bool_t *pNeedReload)

Change scheduler mode to ZES_SCHED_MODE_TIMEOUT or update scheduler mode parameters if already running in this mode.

  • This mode is optimized for multiple applications or contexts submitting work to the hardware. When higher priority work arrives, the scheduler attempts to pause the current executing work within some timeout interval, then submits the other work.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hScheduler – [in] Sysman handle for the component.

  • pProperties – [in] The properties to use when configurating this mode.

  • pNeedReload – [in,out] Will be set to TRUE if a device driver reload is needed to apply the new scheduler mode.

Returns

zesSchedulerSetTimesliceMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerSetTimesliceMode(zes_sched_handle_t hScheduler, zes_sched_timeslice_properties_t *pProperties, ze_bool_t *pNeedReload)

Change scheduler mode to ZES_SCHED_MODE_TIMESLICE or update scheduler mode parameters if already running in this mode.

  • This mode is optimized to provide fair sharing of hardware execution time between multiple contexts submitting work to the hardware concurrently.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hScheduler – [in] Sysman handle for the component.

  • pProperties – [in] The properties to use when configurating this mode.

  • pNeedReload – [in,out] Will be set to TRUE if a device driver reload is needed to apply the new scheduler mode.

Returns

zesSchedulerSetExclusiveMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerSetExclusiveMode(zes_sched_handle_t hScheduler, ze_bool_t *pNeedReload)

Change scheduler mode to ZES_SCHED_MODE_EXCLUSIVE.

  • This mode is optimized for single application/context use-cases. It permits a context to run indefinitely on the hardware without being preempted or terminated. All pending work for other contexts must wait until the running context completes with no further submitted work.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hScheduler – [in] Sysman handle for the component.

  • pNeedReload – [in,out] Will be set to TRUE if a device driver reload is needed to apply the new scheduler mode.

Returns

zesSchedulerSetComputeUnitDebugMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesSchedulerSetComputeUnitDebugMode(zes_sched_handle_t hScheduler, ze_bool_t *pNeedReload)

Change scheduler mode to ZES_SCHED_MODE_COMPUTE_UNIT_DEBUG.

  • This is a special mode that must ben enabled when debugging an application that uses this device e.g. using the Level0 Debug API.

  • It ensures that only one command queue can execute work on the hardware at a given time. Work is permitted to run as long as needed without enforcing any scheduler fairness policies.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • [DEPRECATED] No longer supported.

Parameters
  • hScheduler – [in] Sysman handle for the component.

  • pNeedReload – [in,out] Will be set to TRUE if a device driver reload is needed to apply the new scheduler mode.

Returns

Scheduler Enums

zes_sched_mode_t

enum zes_sched_mode_t

Scheduler mode.

Values:

enumerator ZES_SCHED_MODE_TIMEOUT

Multiple applications or contexts are submitting work to the hardware. When higher priority work arrives, the scheduler attempts to pause the current executing work within some timeout interval, then submits the other work.

enumerator ZES_SCHED_MODE_TIMESLICE

The scheduler attempts to fairly timeslice hardware execution time between multiple contexts submitting work to the hardware concurrently.

enumerator ZES_SCHED_MODE_EXCLUSIVE

Any application or context can run indefinitely on the hardware without being preempted or terminated. All pending work for other contexts must wait until the running context completes with no further submitted work.

enumerator ZES_SCHED_MODE_COMPUTE_UNIT_DEBUG

[DEPRECATED] No longer supported.

enumerator ZES_SCHED_MODE_FORCE_UINT32

Scheduler Structures

zes_sched_properties_t

struct zes_sched_properties_t

Properties related to scheduler component.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t onSubdevice

[out] True if this resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

ze_bool_t canControl

[out] Software can change the scheduler component configuration assuming the user has permissions.

zes_engine_type_flags_t engines

[out] Bitfield of accelerator engine types that are managed by this scheduler component. Note that there can be more than one scheduler component for the same type of accelerator engine.

uint32_t supportedModes

[out] Bitfield of scheduler modes that can be configured for this scheduler component (bitfield of 1<<zes_sched_mode_t).

zes_sched_timeout_properties_t

struct zes_sched_timeout_properties_t

Configuration for timeout scheduler mode (ZES_SCHED_MODE_TIMEOUT)

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint64_t watchdogTimeout

[in,out] The maximum time in microseconds that the scheduler will wait for a batch of work submitted to a hardware engine to complete or to be preempted so as to run another context. If this time is exceeded, the hardware engine is reset and the context terminated. If set to ZES_SCHED_WATCHDOG_DISABLE, a running workload can run as long as it wants without being terminated, but preemption attempts to run other contexts are permitted but not enforced.

zes_sched_timeslice_properties_t

struct zes_sched_timeslice_properties_t

Configuration for timeslice scheduler mode (ZES_SCHED_MODE_TIMESLICE)

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint64_t interval

[in,out] The average interval in microseconds that a submission for a context will run on a hardware engine before being preempted out to run a pending submission for another context.

uint64_t yieldTimeout

[in,out] The maximum time in microseconds that the scheduler will wait to preempt a workload running on an engine before deciding to reset the hardware engine and terminating the associated context.

Standby

Standby Functions

zesDeviceEnumStandbyDomains

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumStandbyDomains(zes_device_handle_t hDevice, uint32_t *pCount, zes_standby_handle_t *phStandby)

Get handle of standby controls.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phStandby – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesStandbyGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesStandbyGetProperties(zes_standby_handle_t hStandby, zes_standby_properties_t *pProperties)

Get standby hardware component properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hStandby – [in] Handle for the component.

  • pProperties – [in,out] Will contain the standby hardware properties.

Returns

zesStandbyGetMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesStandbyGetMode(zes_standby_handle_t hStandby, zes_standby_promo_mode_t *pMode)

Get the current standby promotion mode.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hStandby – [in] Handle for the component.

  • pMode – [in,out] Will contain the current standby mode.

Returns

zesStandbySetMode

ZE_APIEXPORT ze_result_t ZE_APICALL zesStandbySetMode(zes_standby_handle_t hStandby, zes_standby_promo_mode_t mode)

Set standby promotion mode.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hStandby – [in] Handle for the component.

  • mode – [in] New standby mode.

Returns

Standby Enums

zes_standby_type_t

enum zes_standby_type_t

Standby hardware components.

Values:

enumerator ZES_STANDBY_TYPE_GLOBAL

Control the overall standby policy of the device/sub-device.

enumerator ZES_STANDBY_TYPE_FORCE_UINT32

zes_standby_promo_mode_t

enum zes_standby_promo_mode_t

Standby promotion modes.

Values:

enumerator ZES_STANDBY_PROMO_MODE_DEFAULT

Best compromise between performance and energy savings.

enumerator ZES_STANDBY_PROMO_MODE_NEVER

The device/component will never shutdown. This can improve performance but uses more energy.

enumerator ZES_STANDBY_PROMO_MODE_FORCE_UINT32

Standby Structures

zes_standby_properties_t

struct zes_standby_properties_t

Standby hardware component properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_standby_type_t type

[out] Which standby hardware component this controls

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

Temperature

Temperature Functions

zesDeviceEnumTemperatureSensors

ZE_APIEXPORT ze_result_t ZE_APICALL zesDeviceEnumTemperatureSensors(zes_device_handle_t hDevice, uint32_t *pCount, zes_temp_handle_t *phTemperature)

Get handle of temperature sensors.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hDevice – [in] Sysman handle of the device.

  • pCount – [in,out] pointer to the number of components of this type. if count is zero, then the driver shall update the value with the total number of components of this type that are available. if count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • phTemperature – [in,out][optional][range(0, *pCount)] array of handle of components of this type. if count is less than the number of components of this type that are available, then the driver shall only retrieve that number of component handles.

Returns

zesTemperatureGetProperties

ZE_APIEXPORT ze_result_t ZE_APICALL zesTemperatureGetProperties(zes_temp_handle_t hTemperature, zes_temp_properties_t *pProperties)

Get temperature sensor properties.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hTemperature – [in] Handle for the component.

  • pProperties – [in,out] Will contain the temperature sensor properties.

Returns

zesTemperatureGetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesTemperatureGetConfig(zes_temp_handle_t hTemperature, zes_temp_config_t *pConfig)

Get temperature configuration for this sensor - which events are triggered and the trigger conditions.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hTemperature – [in] Handle for the component.

  • pConfig – [in,out] Returns current configuration.

Returns

zesTemperatureSetConfig

ZE_APIEXPORT ze_result_t ZE_APICALL zesTemperatureSetConfig(zes_temp_handle_t hTemperature, const zes_temp_config_t *pConfig)

Set temperature configuration for this sensor - indicates which events are triggered and the trigger conditions.

Parameters
  • hTemperature – [in] Handle for the component.

  • pConfig – [in] New configuration.

Returns

zesTemperatureGetState

ZE_APIEXPORT ze_result_t ZE_APICALL zesTemperatureGetState(zes_temp_handle_t hTemperature, double *pTemperature)

Get the temperature from a specified sensor.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hTemperature – [in] Handle for the component.

  • pTemperature – [in,out] Will contain the temperature read from the specified sensor in degrees Celsius.

Returns

Temperature Enums

zes_temp_sensors_t

enum zes_temp_sensors_t

Temperature sensors.

Values:

enumerator ZES_TEMP_SENSORS_GLOBAL

The maximum temperature across all device sensors.

enumerator ZES_TEMP_SENSORS_GPU

The maximum temperature across all sensors in the GPU.

enumerator ZES_TEMP_SENSORS_MEMORY

The maximum temperature across all sensors in the local memory.

enumerator ZES_TEMP_SENSORS_GLOBAL_MIN

The minimum temperature across all device sensors.

enumerator ZES_TEMP_SENSORS_GPU_MIN

The minimum temperature across all sensors in the GPU.

enumerator ZES_TEMP_SENSORS_MEMORY_MIN

The minimum temperature across all sensors in the local device memory.

enumerator ZES_TEMP_SENSORS_GPU_BOARD

The maximum temperature across all sensors in the GPU Board.

enumerator ZES_TEMP_SENSORS_GPU_BOARD_MIN

The minimum temperature across all sensors in the GPU Board.

enumerator ZES_TEMP_SENSORS_FORCE_UINT32

Temperature Structures

zes_temp_properties_t

struct zes_temp_properties_t

Temperature sensor properties.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_temp_sensors_t type

[out] Which part of the device the temperature sensor measures

ze_bool_t onSubdevice

[out] True if the resource is located on a sub-device; false means that the resource is on the device of the calling Sysman handle

uint32_t subdeviceId

[out] If onSubdevice is true, this gives the ID of the sub-device

double maxTemperature

[out] Will contain the maximum temperature for the specific device in degrees Celsius.

ze_bool_t isCriticalTempSupported

[out] Indicates if the critical temperature event ZES_EVENT_TYPE_FLAG_TEMP_CRITICAL is supported

ze_bool_t isThreshold1Supported

[out] Indicates if the temperature threshold 1 event ZES_EVENT_TYPE_FLAG_TEMP_THRESHOLD1 is supported

ze_bool_t isThreshold2Supported

[out] Indicates if the temperature threshold 2 event ZES_EVENT_TYPE_FLAG_TEMP_THRESHOLD2 is supported

zes_temp_threshold_t

struct zes_temp_threshold_t

Temperature sensor threshold.

Public Members

ze_bool_t enableLowToHigh

[in,out] Trigger an event when the temperature crosses from below the threshold to above.

ze_bool_t enableHighToLow

[in,out] Trigger an event when the temperature crosses from above the threshold to below.

double threshold

[in,out] The threshold in degrees Celsius.

zes_temp_config_t

struct zes_temp_config_t

Temperature configuration - which events should be triggered and the trigger conditions.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

ze_bool_t enableCritical

[in,out] Indicates if event ZES_EVENT_TYPE_FLAG_TEMP_CRITICAL should be triggered by the driver.

zes_temp_threshold_t threshold1

[in,out] Configuration controlling if and when event ZES_EVENT_TYPE_FLAG_TEMP_THRESHOLD1 should be triggered by the driver.

zes_temp_threshold_t threshold2

[in,out] Configuration controlling if and when event ZES_EVENT_TYPE_FLAG_TEMP_THRESHOLD2 should be triggered by the driver.

Powerlimits

Powerlimits Functions

zesPowerGetLimitsExt

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerGetLimitsExt(zes_pwr_handle_t hPower, uint32_t *pCount, zes_power_limit_ext_desc_t *pSustained)

Get power limits.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

  • This function returns all the power limits associated with the supplied power domain.

Parameters
  • hPower – [in] Power domain handle instance.

  • pCount – [in,out] Pointer to the number of power limit descriptors. If count is zero, then the driver shall update the value with the total number of components of this type that are available. If count is greater than the number of components of this type that are available, then the driver shall update the value with the correct number of components.

  • pSustained – [in,out][optional][range(0, *pCount)] Array of query results for power limit descriptors. If count is less than the number of components of this type that are available, then the driver shall only retrieve that number of components.

Returns

zesPowerSetLimitsExt

ZE_APIEXPORT ze_result_t ZE_APICALL zesPowerSetLimitsExt(zes_pwr_handle_t hPower, uint32_t *pCount, zes_power_limit_ext_desc_t *pSustained)

Set power limits.

  • The application can only modify unlocked members of the limit descriptors returned by zesPowerGetLimitsExt.

  • Not all the limits returned by zesPowerGetLimitsExt need to be supplied to this function.

  • Limits do not have to be supplied in the same order as returned by zesPowerGetLimitsExt.

  • The same limit can be supplied multiple times. Limits are applied in the order in which they are supplied.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hPower – [in] Handle for the component.

  • pCount – [in] Pointer to the number of power limit descriptors.

  • pSustained – [in][optional][range(0, *pCount)] Array of power limit descriptors.

Returns

Powerlimits Enums

zes_power_limits_ext_version_t

enum zes_power_limits_ext_version_t

Power Limits Extension Version(s)

Values:

enumerator ZES_POWER_LIMITS_EXT_VERSION_1_0

version 1.0

enumerator ZES_POWER_LIMITS_EXT_VERSION_CURRENT

latest known version

enumerator ZES_POWER_LIMITS_EXT_VERSION_FORCE_UINT32

Powerlimits Structures

zes_power_limit_ext_desc_t

struct zes_power_limit_ext_desc_t

Device power/current limit descriptor.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_power_level_t level

[in,out] duration type over which the power draw is measured, i.e. sustained, burst, peak, or critical.

zes_power_source_t source

[out] source of power used by the system, i.e. AC or DC.

zes_limit_unit_t limitUnit

[out] unit used for specifying limit, i.e. current units (milliamps) or power units (milliwatts).

ze_bool_t enabledStateLocked

[out] indicates if the power limit state (enabled/ignored) can be set (false) or is locked (true).

ze_bool_t enabled

[in,out] indicates if the limit is enabled (true) or ignored (false). If enabledStateIsLocked is True, this value is ignored.

ze_bool_t intervalValueLocked

[out] indicates if the interval can be modified (false) or is fixed (true).

int32_t interval

[in,out] power averaging window in milliseconds. If intervalValueLocked is true, this value is ignored.

ze_bool_t limitValueLocked

[out] indicates if the limit can be set (false) or if the limit is fixed (true).

int32_t limit

[in,out] limit value. If limitValueLocked is true, this value is ignored. The value should be provided in the unit specified by limitUnit.

zes_power_ext_properties_t

struct zes_power_ext_properties_t

Extension properties related to device power settings.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_power_domain_t domain

[out] domain that the power limit belongs to.

zes_power_limit_ext_desc_t *defaultLimit

[out] the factory default limit of the part.

Engineactivity

Engineactivity Functions

zesEngineGetActivityExt

ZE_APIEXPORT ze_result_t ZE_APICALL zesEngineGetActivityExt(zes_engine_handle_t hEngine, uint32_t *pCount, zes_engine_stats_t *pStats)

Get activity stats for Physical Function (PF) and each Virtual Function (VF) associated with engine group.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hEngine – [in] Handle for the component.

  • pCount – [in,out] Pointer to the number of VF engine stats descriptors.

    • if count is zero, the driver shall update the value with the total number of engine stats available.

    • if count is greater than the total number of engine stats available, the driver shall update the value with the correct number of engine stats available.

    • The count returned is the sum of number of VF instances currently available and the PF instance.

  • pStats – [in,out][optional][range(0, *pCount)] array of engine group activity counters.

    • if count is less than the total number of engine stats available, then driver shall only retrieve that number of stats.

    • the implementation shall populate the vector with engine stat for PF at index 0 of the vector followed by user provided pCount-1 number of VF engine stats.

Returns

Engineactivity Enums

zes_engine_activity_ext_version_t

enum zes_engine_activity_ext_version_t

Engine Activity Extension Version(s)

Values:

enumerator ZES_ENGINE_ACTIVITY_EXT_VERSION_1_0

version 1.0

enumerator ZES_ENGINE_ACTIVITY_EXT_VERSION_CURRENT

latest known version

enumerator ZES_ENGINE_ACTIVITY_EXT_VERSION_FORCE_UINT32

Engineactivity Structures

zes_engine_ext_properties_t

struct zes_engine_ext_properties_t

Extension properties related to Engine Groups.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint32_t countOfVirtualFunctionInstance

[out] Number of Virtual Function(VF) instances associated with engine to monitor the utilization of hardware across all Virtual Function from a Physical Function (PF) instance. These VF-by-VF views should provide engine group and individual engine level granularity. This count represents the number of VF instances that are actively using the resource represented by the engine handle.

Rasstate

Rasstate Functions

zesRasGetStateExp

ZE_APIEXPORT ze_result_t ZE_APICALL zesRasGetStateExp(zes_ras_handle_t hRas, uint32_t *pCount, zes_ras_state_exp_t *pState)

Ras Get State.

  • This function retrieves error counters for different RAS error categories.

  • The application may call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hRas – [in] Handle for the component.

  • pCount – [in,out] pointer to the number of RAS state structures that can be retrieved. if count is zero, then the driver shall update the value with the total number of error categories for which state can be retrieved. if count is greater than the number of RAS states available, then the driver shall update the value with the correct number of RAS states available.

  • pState – [in,out][optional][range(0, *pCount)] array of query results for RAS error states for different categories. if count is less than the number of RAS states available, then driver shall only retrieve that number of RAS states.

Returns

zesRasClearStateExp

ZE_APIEXPORT ze_result_t ZE_APICALL zesRasClearStateExp(zes_ras_handle_t hRas, zes_ras_error_category_exp_t category)

Ras Clear State.

  • This function clears error counters for a RAS error category.

  • Clearing errors will affect other threads/applications - the counter values will start from zero.

  • Clearing errors requires write permissions.

  • The application should not call this function from simultaneous threads.

  • The implementation of this function should be lock-free.

Parameters
  • hRas – [in] Handle for the component.

  • category – [in] category for which error counter is to be cleared.

Returns

Rasstate Enums

zes_ras_state_exp_version_t

enum zes_ras_state_exp_version_t

RAS Get State Extension Version(s)

Values:

enumerator ZES_RAS_STATE_EXP_VERSION_1_0

version 1.0

enumerator ZES_RAS_STATE_EXP_VERSION_CURRENT

latest known version

enumerator ZES_RAS_STATE_EXP_VERSION_FORCE_UINT32

zes_ras_error_category_exp_t

enum zes_ras_error_category_exp_t

RAS error categories.

Values:

enumerator ZES_RAS_ERROR_CATEGORY_EXP_RESET

The number of accelerator engine resets attempted by the driver.

enumerator ZES_RAS_ERROR_CATEGORY_EXP_PROGRAMMING_ERRORS

The number of hardware exceptions generated by the way workloads have programmed the hardware

enumerator ZES_RAS_ERROR_CATEGORY_EXP_DRIVER_ERRORS

The number of low level driver communication errors have occurred.

enumerator ZES_RAS_ERROR_CATEGORY_EXP_COMPUTE_ERRORS

The number of errors that have occurred in the compute accelerator hardware

enumerator ZES_RAS_ERROR_CATEGORY_EXP_NON_COMPUTE_ERRORS

The number of errors that have occurred in the fixed-function accelerator hardware

enumerator ZES_RAS_ERROR_CATEGORY_EXP_CACHE_ERRORS

The number of errors that have occurred in caches (L1/L3/register file/shared local memory/sampler)

enumerator ZES_RAS_ERROR_CATEGORY_EXP_DISPLAY_ERRORS

The number of errors that have occurred in the display.

enumerator ZES_RAS_ERROR_CATEGORY_EXP_MEMORY_ERRORS

The number of errors that have occurred in Memory.

enumerator ZES_RAS_ERROR_CATEGORY_EXP_SCALE_ERRORS

The number of errors that have occurred in Scale Fabric.

enumerator ZES_RAS_ERROR_CATEGORY_EXP_L3FABRIC_ERRORS

The number of errors that have occurred in L3 Fabric.

enumerator ZES_RAS_ERROR_CATEGORY_EXP_FORCE_UINT32

Rasstate Structures

zes_ras_state_exp_t

struct zes_ras_state_exp_t

Extension structure for providing RAS error counters for different error sets.

Public Members

zes_ras_error_category_exp_t category

[out] category for which error counter is provided.

uint64_t errorCounter

[out] Current value of RAS counter for specific error category.

Mempageofflinestate

Mempageofflinestate Enums

zes_mem_page_offline_state_exp_version_t

enum zes_mem_page_offline_state_exp_version_t

Memory State Extension Version(s)

Values:

enumerator ZES_MEM_PAGE_OFFLINE_STATE_EXP_VERSION_1_0

version 1.0

enumerator ZES_MEM_PAGE_OFFLINE_STATE_EXP_VERSION_CURRENT

latest known version

enumerator ZES_MEM_PAGE_OFFLINE_STATE_EXP_VERSION_FORCE_UINT32

Mempageofflinestate Structures

zes_mem_page_offline_state_exp_t

struct zes_mem_page_offline_state_exp_t

Extension properties for Memory State.

Public Members

zes_structure_type_t stype

[in] type of this structure

const void *pNext

[in][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

uint32_t memoryPageOffline

[out] Returns the number of Memory Pages Offline

uint32_t maxMemoryPageOffline

[out] Returns the Allowed Memory Pages Offline

Memorytimestampvalidbits

Memorytimestampvalidbits Enums

zes_mem_timestamp_bits_exp_version_t

enum zes_mem_timestamp_bits_exp_version_t

Memory Timestamp Valid Bits Extension Version(s)

Values:

enumerator ZES_MEM_TIMESTAMP_BITS_EXP_VERSION_1_0

version 1.0

enumerator ZES_MEM_TIMESTAMP_BITS_EXP_VERSION_CURRENT

latest known version

enumerator ZES_MEM_TIMESTAMP_BITS_EXP_VERSION_FORCE_UINT32

Memorytimestampvalidbits Structures

zes_mem_timestamp_bits_exp_t

struct zes_mem_timestamp_bits_exp_t

Extension properties for reporting valid bit count for memory timestamp value.

Public Members

uint32_t memoryTimestampValidBits

[out] Returns the number of valid bits in the timestamp values

Powerdomainproperties

Powerdomainproperties Enums

zes_power_domain_properties_exp_version_t

enum zes_power_domain_properties_exp_version_t

Power Domain Properties Extension Version(s)

Values:

enumerator ZES_POWER_DOMAIN_PROPERTIES_EXP_VERSION_1_0

version 1.0

enumerator ZES_POWER_DOMAIN_PROPERTIES_EXP_VERSION_CURRENT

latest known version

enumerator ZES_POWER_DOMAIN_PROPERTIES_EXP_VERSION_FORCE_UINT32

Powerdomainproperties Structures

zes_power_domain_exp_properties_t

struct zes_power_domain_exp_properties_t

Extension structure for providing power domain information associated with a power handle.

Public Members

zes_structure_type_t stype

[in] type of this structure

void *pNext

[in,out][optional] must be null or a pointer to an extension-specific structure (i.e. contains stype and pNext).

zes_power_domain_t powerDomain

[out] Power domain associated with the power handle.