github/concrete
1
1
Fork 0
mirror of https://github.com/zama-ai/concrete.git synced 2026-02-08 19:44:57 -05:00
Code Issues Packages Projects Releases Wiki Activity

chore(core): fix clippy error in cuda backend and fix formatting

Browse Source
This commit is contained in:
Agnes Leroy
2022-11-10 14:12:44 +01:00
committed by Agnès Leroy
parent 80f4ca7338
commit da654ee9cb
2 changed files with 14 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
src/bootstrap_wop.cuh
View File

@@ -302,8 +302,8 @@ void host_cmux_tree(void *v_stream, Torus *glwe_array_out, Torus *ggsw_in,
int ggsw_size = r * polynomial_size * (glwe_dimension + 1) *
(glwe_dimension + 1) * level_count;
double2 *d_ggsw_fft_in =
(double2 *)cuda_malloc_async(ggsw_size * sizeof(double), *stream, gpu_index);
double2 *d_ggsw_fft_in = (double2 *)cuda_malloc_async(
ggsw_size * sizeof(double), *stream, gpu_index);
batch_fft_ggsw_vector<Torus, STorus, params>(v_stream, d_ggsw_fft_in, ggsw_in,
r, glwe_dimension,
@@ -328,10 +328,10 @@ void host_cmux_tree(void *v_stream, Torus *glwe_array_out, Torus *ggsw_in,
// Allocate buffers
int glwe_size = (glwe_dimension + 1) * polynomial_size;
Torus *d_buffer1 =
(Torus *)cuda_malloc_async(num_lut * glwe_size * sizeof(Torus), *stream, gpu_index);
Torus *d_buffer2 =
(Torus *)cuda_malloc_async(num_lut * glwe_size * sizeof(Torus), *stream, gpu_index);
Torus *d_buffer1 = (Torus *)cuda_malloc_async(
num_lut * glwe_size * sizeof(Torus), *stream, gpu_index);
Torus *d_buffer2 = (Torus *)cuda_malloc_async(
num_lut * glwe_size * sizeof(Torus), *stream, gpu_index);
checkCudaErrors(cudaMemcpyAsync(d_buffer1, lut_vector,
num_lut * glwe_size * sizeof(Torus),

15
src/device.cu
View File

@@ -33,14 +33,15 @@ void *cuda_malloc(uint64_t size, uint32_t gpu_index) {
/// Allocates a size-byte array at the device memory. Tries to do it
/// asynchronously.
void *cuda_malloc_async(uint64_t size, cudaStream_t stream, uint32_t gpu_index) {
void *cuda_malloc_async(uint64_t size, cudaStream_t stream,
uint32_t gpu_index) {
void *ptr;
int support_async_alloc;
checkCudaErrors(cudaDeviceGetAttribute(&support_async_alloc, cudaDevAttrMemoryPoolsSupported,
gpu_index));
checkCudaErrors(cudaDeviceGetAttribute(
&support_async_alloc, cudaDevAttrMemoryPoolsSupported, gpu_index));
if(support_async_alloc)
if (support_async_alloc)
checkCudaErrors(cudaMallocAsync((void **)&ptr, size, stream));
else
checkCudaErrors(cudaMalloc((void **)&ptr, size));
@@ -160,10 +161,10 @@ int cuda_drop(void *ptr, uint32_t gpu_index) {
int cuda_drop_async(void *ptr, cudaStream_t stream, uint32_t gpu_index) {
int support_async_alloc;
checkCudaErrors(cudaDeviceGetAttribute(&support_async_alloc, cudaDevAttrMemoryPoolsSupported,
gpu_index));
checkCudaErrors(cudaDeviceGetAttribute(
&support_async_alloc, cudaDevAttrMemoryPoolsSupported, gpu_index));
if(support_async_alloc)
if (support_async_alloc)
checkCudaErrors(cudaFreeAsync(ptr, stream));
else
checkCudaErrors(cudaFree(ptr));