Override the default cudnn heuristics. (#957)

This commit is contained in:
Laurent Mazare
2023-09-25 10:31:53 +01:00
committed by GitHub
parent 1ce7fe2543
commit dc47224ab9
2 changed files with 33 additions and 1 deletions

View File

@ -34,6 +34,9 @@ pub(crate) fn launch_conv2d<
params: &crate::conv::ParamsConv2D,
dev: &crate::cuda_backend::CudaDevice,
) -> crate::Result<()> {
use crate::conv::CudnnFwdAlgo as CandleAlgo;
use cudarc::cudnn::sys::cudnnConvolutionFwdAlgo_t as A;
let device_id = dev.id();
let cudnn = CUDNN.with(|cudnn| {
if let Some(cudnn) = cudnn.borrow().get(&device_id) {
@ -90,7 +93,20 @@ pub(crate) fn launch_conv2d<
w: &w,
y: &y,
};
let alg = conv2d.pick_algorithm()?;
let alg = match params.cudnn_fwd_algo {
None => conv2d.pick_algorithm()?,
Some(CandleAlgo::ImplicitGemm) => A::CUDNN_CONVOLUTION_FWD_ALGO_IMPLICIT_GEMM,
Some(CandleAlgo::ImplicitPrecompGemm) => {
A::CUDNN_CONVOLUTION_FWD_ALGO_IMPLICIT_PRECOMP_GEMM
}
Some(CandleAlgo::Gemm) => A::CUDNN_CONVOLUTION_FWD_ALGO_GEMM,
Some(CandleAlgo::Direct) => A::CUDNN_CONVOLUTION_FWD_ALGO_DIRECT,
Some(CandleAlgo::Fft) => A::CUDNN_CONVOLUTION_FWD_ALGO_FFT,
Some(CandleAlgo::FftTiling) => A::CUDNN_CONVOLUTION_FWD_ALGO_FFT_TILING,
Some(CandleAlgo::Winograd) => A::CUDNN_CONVOLUTION_FWD_ALGO_WINOGRAD,
Some(CandleAlgo::WinogradNonFused) => A::CUDNN_CONVOLUTION_FWD_ALGO_WINOGRAD_NONFUSED,
Some(CandleAlgo::Count) => A::CUDNN_CONVOLUTION_FWD_ALGO_COUNT,
};
let workspace_size = conv2d.get_workspace_size(alg)?;
let mut workspace = dev.cuda_device().alloc_zeros::<u8>(workspace_size)?;
unsafe {