Mathis/CUDANet
1
0
Fork 0
mirror of https://github.com/lordmathis/CUDANet.git synced 2025-11-05 17:34:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

Implement to_cuda function

Browse Source
This commit is contained in:
Matúš Námešný
2024-02-09 19:18:30 +01:00
parent 6645cd31ba
commit 3e6b7dc7e6
4 changed files with 37 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
src/layers/dense.cpp
View File

@@ -1,5 +1,10 @@
#include "dense.h"
#include "cuda_helper.h"
#include <cstdlib>
#include <cublas_v2.h>
#include <cstdio>
#include <stdexcept>
Layers::Dense::Dense(int inputSize, int outputSize, cublasHandle_t cublasHandle)
@@ -13,12 +18,10 @@ Layers::Dense::Dense(int inputSize, int outputSize, cublasHandle_t cublasHandle)
initializeBiases();
// Allocate GPU memory for weights and biases
cudaMalloc((void**)&d_weights, sizeof(float) * inputSize * outputSize);
cudaMalloc((void**)&d_biases, sizeof(float) * biases.size());
CUDA_CHECK(cudaMalloc((void**)&d_weights, sizeof(float) * inputSize * outputSize));
CUDA_CHECK(cudaMalloc((void**)&d_biases, sizeof(float) * biases.size()));
// Copy weights and biases to GPU
cudaMemcpy(d_weights, weights.data(), sizeof(float) * inputSize * outputSize, cudaMemcpyHostToDevice);
cudaMemcpy(d_biases, biases.data(), sizeof(float) * biases.size(), cudaMemcpyHostToDevice);
to_cuda();
}
Layers::Dense::~Dense() {
@@ -49,4 +52,9 @@ void Layers::Dense::forward(const float* input, float* output) {
// Add biases
cublasSaxpy(cublasHandle, outputSize, &alpha, d_biases, 1, output, 1);
}
void Layers::Dense::to_cuda() {
CUDA_CHECK(cudaMemcpy(d_weights, weights.data(), sizeof(float) * inputSize * outputSize, cudaMemcpyHostToDevice));
CUDA_CHECK(cudaMemcpy(d_biases, biases.data(), sizeof(float) * biases.size(), cudaMemcpyHostToDevice));
}