gohlml

package module
v1.15.0 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Mar 27, 2024 License: Apache-2.0 Imports: 10 Imported by: 0

README

gohlml

This repository provides a translation layer from HLML to Golang. The library is meant to be integrated into other modules requiring HLML access and to be set of wrappers around the C API provided by libhlml.so

Testing

To test the code, transfer the repository to a server where the Habana driver is installed and run the following:

go test

Code Cover

To validate metrics code coverage, run:

go test -cover

Documentation

Overview

Package gohlml is a Go package serves as a bridge to work with hlml C library. It allows access to native Habana device commands and information.

Index

Constants

View Source 
const (

	// HlmlCriticalError indicates a critical error in the device
	HlmlCriticalError = C.HLML_EVENT_CRITICAL_ERR
	// HLDriverPath indicates on habana device dir
	HLDriverPath = "/sys/class/accel"
	// HLModulePath indicates on habana module dir
	HLModulePath = "/sys/module/habanalabs"
	// BITSPerLong repsenets 64 bits in logs
	BITSPerLong = 64
)

Variables

View Source 
var (
	ErrNotIntialized      = errors.New("hlml not initialized")
	ErrInvalidArgument    = errors.New("invalid argument")
	ErrNotSupported       = errors.New("not supported")
	ErrAlreadyInitialized = errors.New("hlml already initialized")
	ErrNotFound           = errors.New("not found")
	ErrInsufficientSize   = errors.New("insufficient size")
	ErrDriverNotLoaded    = errors.New("driver not loaded")
	ErrAipIsLost          = errors.New("aip is lost")
	ErrMemoryError        = errors.New("memory error")
	ErrNoData             = errors.New("no data")
	ErrUnknownError       = errors.New("unknown error")
)

Functions

func DeleteEventSet 

func DeleteEventSet(es EventSet)

func DeviceCount 

func DeviceCount() (uint, error)

DeviceCount gets number of Habana devices in the system

func FWVersion 

func FWVersion(idx uint) (kernel string, uboot string, err error)

FWVersion returns the firmware version for a given device

func GetDeviceTypeName added in v1.14.0 

func GetDeviceTypeName() (string, error)

func InitWithLogs 

func InitWithLogs() error

InitWithLogs initializes the HLML library with logging on

func Initialize 

func Initialize() error

Initialize initializes the HLML library

func RegisterEventForDevice 

func RegisterEventForDevice(es EventSet, event int, uuid string) error

func Shutdown 

func Shutdown() error

Shutdown shutdowns the HLML library

func SystemDriverVersion 

func SystemDriverVersion() (string, error)

SystemDriverVersion returns the driver version on the system

Types

type Device 

type Device struct {
	// contains filtered or unexported fields
}

Device struct maps to C HLML structure

func DeviceHandleByIndex 

func DeviceHandleByIndex(idx uint) (Device, error)

DeviceHandleByIndex gets a handle to a particular device by index

func DeviceHandleBySerial 

func DeviceHandleBySerial(serial string) (*Device, error)

DeviceHandleBySerial gets a handle to a particular device by serial number

func DeviceHandleByUUID 

func DeviceHandleByUUID(uuid string) (Device, error)

DeviceHandleByUUID gets a handle to a particular device by UUIC

func (Device) BoardID 

func (d Device) BoardID() (uint, error)

BoardID returns an ID for the PCB board

func (Device) ClockThrottleReasons 

func (d Device) ClockThrottleReasons() (uint64, error)

ClockThrottleReasons returns current clock throttle reasons

func (Device) ECCMode added in v1.14.0 

func (d Device) ECCMode() (uint, uint, error)

ECCMode retrieves the current and pending ECC modes for the device 

1 - ECCMode enabled
0 - ECCMode disabled

func (Device) EnergyConsumptionCounter 

func (d Device) EnergyConsumptionCounter() (uint64, error)

EnergyConsumptionCounter returns energy consumption

func (Device) HLRevision 

func (d Device) HLRevision() (int, error)

HLRevision returns the revision of the HL library

func (Device) ICClockMax added in v1.14.0 

func (d Device) ICClockMax() (uint, error)

ICClockMax returns the maximum IC clock frequency for a given device

func (Device) IsReplacedRowsPendingStatus added in v1.14.0 

func (d Device) IsReplacedRowsPendingStatus() (int, error)

IsReplacedRowsPendingStatus return 0 (false) or 1 (true) if there are any rows need of replacement in a power cycle

func (Device) MMEClockMax added in v1.14.0 

func (d Device) MMEClockMax() (uint, error)

MMEClockMax returns the maximum MME clock frequency for a given device

func (Device) MacAddressInfo added in v1.14.0 

func (d Device) MacAddressInfo() (map[int]string, error)

MacAddressInfo retrieves the masks for supported ports and external ports.

func (Device) MemoryInfo 

func (d Device) MemoryInfo() (uint64, uint64, uint64, error)

MemoryInfo returns the current memory usage in bytes for total, used, free

func (Device) MinorNumber 

func (d Device) MinorNumber() (uint, error)

MinorNumber returns Minor number.

func (Device) ModuleID added in v1.6.1 

func (d Device) ModuleID() (uint, error)

ModuleID returns the device moduleID

func (Device) Name 

func (d Device) Name() (string, error)

Name returns Device Name

func (Device) NicLinkStatus added in v1.14.0 

func (d Device) NicLinkStatus(port uint) (uint, error)

NicLinkStatus gets a port and checks its status. return 1 (up) or 0 (down)

func (Device) NumaNode 

func (d Device) NumaNode() (*uint, error)

NumaNode returns the Numa affinity of the device or nil is no affinity.

func (Device) PCBAssemblyVersion 

func (d Device) PCBAssemblyVersion() (string, error)

PCBAssemblyVersion returns the PCB Assembly info

func (Device) PCBVersion 

func (d Device) PCBVersion() (string, error)

PCBVersion returns the PCB version

func (Device) PCIBus 

func (d Device) PCIBus() (uint, error)

PCIBus returns the PCI bus info for a given device

func (Device) PCIBusID 

func (d Device) PCIBusID() (string, error)

PCIBusID returns the PCI bus id for a given device

func (Device) PCIDomain 

func (d Device) PCIDomain() (uint, error)

PCIDomain returns the PCI domain for a given device

func (Device) PCIID 

func (d Device) PCIID() (uint, error)

PCIID returns the PCI id for a given device

func (Device) PCILinkSpeed 

func (d Device) PCILinkSpeed() (uint, error)

PCILinkSpeed returns the current PCI link speed for a given device

func (Device) PCILinkWidth 

func (d Device) PCILinkWidth() (uint, error)

PCILinkWidth returns the current PCI link width for a given device

func (Device) PCIReplayCounter 

func (d Device) PCIReplayCounter() (uint, error)

PCIReplayCounter returns PCIe replay count

func (Device) PCIeLinkGeneration 

func (d Device) PCIeLinkGeneration() (uint, error)

PCIeLinkGeneration returns PCIe replay count MUST run with SUDO/priviledged

func (Device) PCIeLinkWidth 

func (d Device) PCIeLinkWidth() (uint, error)

PCIeLinkWidth returns PCIe link width

func (Device) PCIeRX 

func (d Device) PCIeRX() (uint, error)

PCIeRX returns PCIe receive throughput

func (Device) PCIeTX 

func (d Device) PCIeTX() (uint, error)

PCIeTX returns PCIe transmit throughput

func (Device) PowerManagementDefaultLimit added in v1.14.0 

func (d Device) PowerManagementDefaultLimit() (uint, error)

PowerManagementDefaultLimit Retrieves default power management limit on this device, in milliwatts. Default power management limit is a power management limit that the device boots with.

func (Device) PowerUsage 

func (d Device) PowerUsage() (uint, error)

PowerUsage returns the power usage in milliwatts for a given device

func (Device) ReplacedRowDoubleBitECC added in v1.14.0 

func (d Device) ReplacedRowDoubleBitECC() (uint, error)

ReplacedRowDoubleBitECC returns the number of rows with double-bit ecc errors

func (Device) ReplacedRowSingleBitECC added in v1.14.0 

func (d Device) ReplacedRowSingleBitECC() (uint, error)

ReplacedRowSingleBitECC returns the number of rows with single-bit ecc errors

func (Device) SOCClockInfo 

func (d Device) SOCClockInfo() (uint, error)

SOCClockInfo returns the SoC clock frequency for a given device

func (Device) SOCClockMax added in v1.14.0 

func (d Device) SOCClockMax() (uint, error)

SOCClockMax returns the maximum SoC clock frequency for a given device

func (Device) SerialNumber 

func (d Device) SerialNumber() (string, error)

SerialNumber returns the device serial number

func (Device) TPCClockMax added in v1.14.0 

func (d Device) TPCClockMax() (uint, error)

TPCClockMax returns the maximum TPC clock frequency for a given device

func (Device) TemperatureOnBoard added in v1.14.0 

func (d Device) TemperatureOnBoard() (uint, error)

TemperatureOnBoard returns the temperature in celsius for a device board

func (Device) TemperatureOnChip added in v1.14.0 

func (d Device) TemperatureOnChip() (uint, error)

TemperatureOnChip returns the temperature in celsius for a the device chip

func (Device) TemperatureThresholdGPU added in v1.14.0 

func (d Device) TemperatureThresholdGPU() (uint, error)

TemperatureThresholdGPU Retrieves the known temperature threshold for the AIP with the specified threshold type in degrees

func (Device) TemperatureThresholdMemory added in v1.14.0 

func (d Device) TemperatureThresholdMemory() (uint, error)

TemperatureThresholdMemory Retrieves the known temperature threshold for the AIP with the specified threshold type in degrees

func (Device) TemperatureThresholdShutdown added in v1.14.0 

func (d Device) TemperatureThresholdShutdown() (uint, error)

TemperatureThresholdShutdown Retrieves the known temperature threshold for the AIP with the specified threshold type in degrees

func (Device) TemperatureThresholdSlowdown added in v1.14.0 

func (d Device) TemperatureThresholdSlowdown() (uint, error)

TemperatureThresholdSlowdown Retrieves the known temperature threshold for the AIP with the specified threshold type in degrees

func (Device) UUID 

func (d Device) UUID() (string, error)

UUID returns the unique id for a given device

func (Device) UtilizationInfo 

func (d Device) UtilizationInfo() (uint, error)

UtilizationInfo returns the utilization aip rate for a given device

type Event 

type Event struct {
	Serial string
	Etype  uint64
}

Event contains uuid and event type

func WaitForEvent 

func WaitForEvent(es EventSet, timeout uint) (Event, error)

type EventSet 

type EventSet struct {
	// contains filtered or unexported fields
}

EventSet is a cast of the C type of the hlml event set

func NewEventSet 

func NewEventSet() EventSet

type PCIInfo 

type PCIInfo struct {
	BusID    string
	DeviceID uint
}

PCIInfo contains the PCI properties of the device

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.