diff options
Diffstat (limited to 'xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu')
| -rw-r--r-- | xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu | 11 |
1 files changed, 9 insertions, 2 deletions
diff --git a/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu b/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu index 0fc99a4..492201d 100644 --- a/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu +++ b/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu @@ -218,7 +218,6 @@ extern "C" int cryptonight_extra_cpu_init(nvid_ctx* ctx) } size_t wsize = ctx->device_blocks * ctx->device_threads; - CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_long_state, hashMemSize * wsize)); CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_ctx_state, 50 * sizeof(uint32_t) * wsize)); CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_ctx_key1, 40 * sizeof(uint32_t) * wsize)); CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_ctx_key2, 40 * sizeof(uint32_t) * wsize)); @@ -228,6 +227,10 @@ extern "C" int cryptonight_extra_cpu_init(nvid_ctx* ctx) CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_input, 21 * sizeof (uint32_t ) )); CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_result_count, sizeof (uint32_t ) )); CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_result_nonce, 10 * sizeof (uint32_t ) )); + CUDA_CHECK_MSG( + ctx->device_id, + "\n**suggestion: Try to reduce the value of the attribute 'threads' in the NVIDIA config file.**", + cudaMalloc(&ctx->d_long_state, hashMemSize * wsize)); return 1; } @@ -254,7 +257,11 @@ extern "C" void cryptonight_extra_cpu_final(nvid_ctx* ctx, uint32_t startNonce, CUDA_CHECK(ctx->device_id, cudaMemset( ctx->d_result_nonce, 0xFF, 10 * sizeof (uint32_t ) )); CUDA_CHECK(ctx->device_id, cudaMemset( ctx->d_result_count, 0, sizeof (uint32_t ) )); - CUDA_CHECK_KERNEL(ctx->device_id, cryptonight_extra_gpu_final<<<grid, block >>>( wsize, target, ctx->d_result_count, ctx->d_result_nonce, ctx->d_ctx_state )); + CUDA_CHECK_MSG_KERNEL( + ctx->device_id, + "\n**suggestion: Try to increase the value of the attribute 'bfactor' in the NVIDIA config file.**", + cryptonight_extra_gpu_final<<<grid, block >>>( wsize, target, ctx->d_result_count, ctx->d_result_nonce, ctx->d_ctx_state ) + ); CUDA_CHECK(ctx->device_id, cudaMemcpy( rescount, ctx->d_result_count, sizeof (uint32_t ), cudaMemcpyDeviceToHost )); CUDA_CHECK(ctx->device_id, cudaMemcpy( resnonce, ctx->d_result_nonce, 10 * sizeof (uint32_t ), cudaMemcpyDeviceToHost )); |
