GpuExecutionContext

Models & Types · AiDotNet.Tensors.Engines.Gpu

No summary documentation available yet.

Implements: IDisposable

Properties

Property Summary
GpuExecutionContext Current { get; }
GpuExecutionMode EffectiveExecutionMode { get; }
GpuExecutionOptions Options { get; }
GpuStreamPool StreamPool { get; }
GpuTensorRegistry Registry { get; }
IAsyncGpuBackend AsyncBackend { get; }
IDirectGpuBackend Backend { get; }
bool IsGpuAvailable { get; }
bool SupportsAsync { get; }

Methods

Method Summary
GpuExecutionContext Begin(IDirectGpuBackend backend, GpuExecutionOptions options)
GpuMemoryStats GetMemoryStats()
Tensor<T> Empty(int[] shape, GpuTensorRole role)
Tensor<T> Upload(T[] data, int[] shape, GpuTensorRole role)
Tensor<T> Upload(Tensor<T> tensor, GpuTensorRole role)
Tensor<T> Zeros(int[] shape, GpuTensorRole role)
bool ShouldUseGpu(int elementCount)
void Dispose()
void Synchronize()