nvlink

type ExpectedLinkStates struct {
	// AtLeastGPUsWithAllLinksFeatureEnabled is the expected/minimum number of GPUs with all links feature enabled.
	// This is useful to detect the following scenarios:
	// - DeviceGetNvLinkState() returns SUCCESS (not ERROR_NOT_SUPPORTED)
	// - Each link returns state != FEATURE_ENABLED (they're FEATURE_DISABLED or inactive)
	// - Thus, safe to assume that NVLink is supported (no NOT_SUPPORTED error)
	// - But ALL links are inactive/disabled
	// - (e.g., nvidia-smi returns "Unable to retrieve NVLink information as all links are inActive")
	// e.g., if set to 8 and one GPU has some nvlinks feature disabled, it will be considered as unhealthy.
	AtLeastGPUsWithAllLinksFeatureEnabled int `json:"at_least_gpus_with_all_links_feature_enabled"`
}

type NVLink struct {
	// Represents the GPU UUID.
	UUID string `json:"uuid"`

	// BusID is the GPU bus ID from the nvml API.
	//  e.g., "0000:0f:00.0"
	BusID string `json:"bus_id"`

	// States is the list of nvlink states.
	States NVLinkStates `json:"states"`

	// Supported is true if the NVLink is supported by the device.
	Supported bool `json:"supported"`
}

type NVLinkState struct {
	// Link is the nvlink link number.
	Link int `json:"link"`

	// FeatureEnabled is true if the nvlink feature is enabled.
	FeatureEnabled bool `json:"feature_enabled"`
	// ReplayErrors is the number of replay errors.
	ReplayErrors uint64 `json:"replay_errors"`
	// RecoveryErrors is the number of recovery errors.
	RecoveryErrors uint64 `json:"recovery_errors"`
	// CRCErrors is the number of crc errors.
	CRCErrors uint64 `json:"crc_errors"`

	// ThroughputRawTxBytes is the NVLink TX Data throughput + protocol overhead in bytes.
	ThroughputRawTxBytes uint64 `json:"throughput_raw_tx_bytes"`
	// ThroughputRawRxBytes is the NVLink RX Data throughput + protocol overhead in bytes.
	ThroughputRawRxBytes uint64 `json:"throughput_raw_rx_bytes"`
}

type NVLinkStates []NVLinkState