Made it possible to run some of the GEMM kernels using CUDA (those wi…

…thout shared memory)
CNugteren · Jun 29, 2016 · 0526f9d · 0526f9d
1 parent 6177c14
commit 0526f9d
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/samples/cl_to_cuda.h b/samples/cl_to_cuda.h
@@ -17,6 +17,7 @@
 #define __local __shared__
 #define restrict __restrict__
 #define __constant const
+#define inline __device__ inline // assumes all device functions are annotated with inline in OpenCL
 
 // Replaces OpenCL synchronisation with CUDA synchronisation
 #define barrier(x) __syncthreads()