gpuallocator

package

v1.28.0 Latest Latest Go to latest Published: May 25, 2025 License: Apache-2.0 Imports: 13 Imported by: 0

Details

Package gpuallocator handles GPU allocation

This section is empty.

This section is empty.

This section is empty.

type CompactFirst struct{}

CompactFirst selects GPU with minimum available resources (most utilized) to efficiently pack workloads and maximize GPU utilization

func (c CompactFirst) SelectGPUs(gpus []tfv1.GPU, count uint) ([]*tfv1.GPU, error)

SelectGPUs selects multiple GPUs from the same node with the least available resources (most packed)

type GpuAllocator struct {
	client.Client
	// contains filtered or unexported fields
}

func NewGpuAllocator(ctx context.Context, client client.Client, syncInterval time.Duration) *GpuAllocator

func (s *GpuAllocator) Alloc(
	ctx context.Context,
	poolName string,
	request tfv1.Resource,
	count uint,
) ([]*tfv1.GPU, error)

Alloc allocates a request to a gpu or multiple gpus from the same node.

func (s *GpuAllocator) Dealloc(ctx context.Context, request tfv1.Resource, gpu *tfv1.GPU) error

Dealloc deallocates a request from a gpu.

func (s *GpuAllocator) SetupWithManager(ctx context.Context, mgr manager.Manager) (<-chan struct{}, error)

SetupWithManager sets up the GpuAllocator with the Manager.

func (s *GpuAllocator) Stop()

Stop stops all background goroutines

type LowLoadFirst struct{}

LowLoadFirst selects GPU with maximum available resources (least utilized) to distribute workloads more evenly across GPUs

func (l LowLoadFirst) SelectGPUs(gpus []tfv1.GPU, count uint) ([]*tfv1.GPU, error)

SelectGPUs selects multiple GPUs from the same node with the most available resources (least loaded)

type Strategy interface {
	SelectGPUs(gpus []tfv1.GPU, count uint) ([]*tfv1.GPU, error)
}

func NewStrategy(placementMode tfv1.PlacementMode) Strategy

NewStrategy creates a strategy based on the placement mode

Path	Synopsis
filter