Browse Source

neoscrypt: last vstudio bool warnings fix

master
Tanguy Pruvot 10 years ago
parent
commit
954dbf2388
  1. 8
      neoscrypt/cuda_neoscrypt.cu

8
neoscrypt/cuda_neoscrypt.cu

@ -599,7 +599,7 @@ static __device__ __forceinline__ void neoscrypt_salsa(uint16 *XV) @@ -599,7 +599,7 @@ static __device__ __forceinline__ void neoscrypt_salsa(uint16 *XV)
#define SHIFT 130
__global__ __launch_bounds__(128, 1)
void neoscrypt_gpu_hash_k0(uint32_t threads, uint32_t startNonce, bool stratum)
void neoscrypt_gpu_hash_k0(uint32_t threads, uint32_t startNonce, int stratum)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
// if (thread < threads)
@ -686,7 +686,7 @@ void neoscrypt_gpu_hash_k3(uint32_t threads, uint32_t startNonce) @@ -686,7 +686,7 @@ void neoscrypt_gpu_hash_k3(uint32_t threads, uint32_t startNonce)
}
__global__ __launch_bounds__(128, 1)
void neoscrypt_gpu_hash_k4(uint32_t threads, uint32_t startNonce, uint32_t *nonceRes, bool stratum)
void neoscrypt_gpu_hash_k4(uint32_t threads, uint32_t startNonce, uint32_t *nonceRes, int stratum)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
@ -746,11 +746,11 @@ uint32_t neoscrypt_cpu_hash_k4(int thr_id, uint32_t threads, uint32_t startNounc @@ -746,11 +746,11 @@ uint32_t neoscrypt_cpu_hash_k4(int thr_id, uint32_t threads, uint32_t startNounc
dim3 grid((threads + threadsperblock - 1) / threadsperblock);
dim3 block(threadsperblock);
neoscrypt_gpu_hash_k0 <<< grid, block >>>(threads, startNounce, (bool) have_stratum);
neoscrypt_gpu_hash_k0 <<< grid, block >>>(threads, startNounce, have_stratum);
neoscrypt_gpu_hash_k01 <<< grid, block >>>(threads, startNounce);
neoscrypt_gpu_hash_k2 <<< grid, block >>>(threads, startNounce);
neoscrypt_gpu_hash_k3 <<< grid, block >>>(threads, startNounce);
neoscrypt_gpu_hash_k4 <<< grid, block >>>(threads, startNounce, d_NNonce[thr_id], (bool) have_stratum);
neoscrypt_gpu_hash_k4 <<< grid, block >>>(threads, startNounce, d_NNonce[thr_id], have_stratum);
MyStreamSynchronize(NULL, order, thr_id);
cudaMemcpy(&result[thr_id], d_NNonce[thr_id], sizeof(uint32_t), cudaMemcpyDeviceToHost);

Loading…
Cancel
Save