Merge pull request #1198 from psychocrypt/fix-inputBufferSizeCuda

Fix input buffer size cuda

.gitignore

@@ -17,3 +17,8 @@ config-debug.txt
 # KDevelop files
 .kdev4/
 xmr-stak.kdev4
+
+# Idea/Clion project files
+cmake-build-release/
+cmake-build-debug/
+\.idea/

xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu

@@ -127,7 +127,7 @@ __global__ void cryptonight_extra_gpu_prepare( int threads, uint32_t * __restric
 	uint32_t ctx_b[4];
 	uint32_t ctx_key1[40];
 	uint32_t ctx_key2[40];
-	uint32_t input[21];
+	uint32_t input[28];
 
 	memcpy( input, d_input, len );
 	//*((uint32_t *)(((char *)input) + 39)) = startNonce + thread;
@@ -299,7 +299,7 @@ extern "C" int cryptonight_extra_cpu_init(nvid_ctx* ctx)
 	CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_ctx_text, 32 * sizeof(uint32_t) * wsize));
 	CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_ctx_a, 4 * sizeof(uint32_t) * wsize));
 	CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_ctx_b, ctx_b_size));
-	CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_input, 21 * sizeof (uint32_t ) ));
+	CUDA_CHECK(ctx->device_id, cudaMalloc(&ctx->d_input, 28 * 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(