Skip to main content

Module gpu

Module gpu 

Source
Expand description

GPU acceleration support for SIMD operations

This module provides CUDA and OpenCL kernel interfaces for GPU-accelerated machine learning operations with fallback to CPU SIMD implementations.

Modules§

cuda
CUDA specific implementation
opencl
OpenCL specific implementation

Structs§

GpuBuffer
GPU memory buffer wrapper
GpuContext
GPU context for managing resources
GpuDevice
GPU device information
GpuManager
GPU manager for handling multiple devices and backends
GpuStream
GPU stream for asynchronous operations
KernelConfig
GPU kernel launch parameters

Enums§

GpuBackend
GPU computation backends

Statics§

GPU_MANAGER

Traits§

GpuOperations
GPU operations interface

Functions§

get_gpu_devices
Get available GPU devices
initialize_gpu
Initialize GPU support
is_gpu_available
Check if GPU acceleration is available