Module gpu_optimizer 

GPU-accelerated optimizer operations

This module provides GPU acceleration for optimization using SciRS2’s GPU abstractions. Enables 10-50x speedup for large models through GPU parallelism and tensor cores.

Features

• GPU-accelerated parameter updates
• Tensor core support for mixed-precision training
• Multi-backend support (CUDA, Metal, OpenCL, WebGPU via SciRS2)
• Automatic host-device data transfer
• GPU memory tracking and management

Performance

Achieves 10-50x speedup over CPU for models with millions of parameters.

SciRS2 Integration

This module uses SciRS2-Core GPU abstractions exclusively:

• scirs2_core::gpu::GpuContext for GPU context management
• scirs2_core::gpu::GpuBuffer for GPU memory allocation
• scirs2_core::gpu::GpuKernel for GPU kernel execution
• scirs2_core::tensor_cores for mixed-precision optimization
• scirs2_core::array_protocol::GPUArray for GPU array interface

Structs

GpuConfig

GPU optimizer configuration

GpuMemoryStats

GPU memory statistics

GpuOptimizer

GPU-accelerated optimizer wrapper

GpuUtils

GPU optimizer utilities