Merge pull request #265 from phani544/nvccTests

[nvccTests]Enabled inline_asm_vadd on nvcc
2017-11-20 07:28:29 +05:30
parent aeadc3f18f 9fa3ed4f6f
commit 29c0ab8401
1 changed files with 7 additions and 1 deletions
@@ -39,7 +39,13 @@ __global__ void vadd_asm(hipLaunchParm lp,
 {
    int i = blockDim.x * blockIdx.x + threadIdx.x;

-    asm volatile ("v_add_f32_e32 %0, %1, %2" : "=v" (out[i]) : "v"(in[i]),"v" (out[i]));
+#ifdef __HIP_PLATFORM_NVCC__
+        asm volatile("add.f32 %0,%1,%2;":"=f"(out[i]):"f"(in[i]),"f"(out[i]));
+#endif
+
+#ifdef __HIP_PLATFORM_HCC__
+   asm volatile ("v_add_f32_e32 %0, %1, %2" : "=v" (out[i]) : "v"(in[i]),"v" (out[i]));
+#endif
 }

 // CPU implementation of Vector Result