Add hipMemcpy3DAsync (#1320)

* Add hipMemcpy3DAsync

* Fix CI build error

* Move back stream resolution to internal function

* Remove stream redefinition and check

hipamd/include/hip/hcc_detail/hip_runtime_api.h

@@ -2022,6 +2022,19 @@ hipError_t hipMemcpyHtoA(hipArray* dstArray, size_t dstOffset, const void* srcHo
  */
 hipError_t hipMemcpy3D(const struct hipMemcpy3DParms* p);
 
+/**
+ *  @brief Copies data between host and device asynchronously.
+ *
+ *  @param[in]   p        3D memory copy parameters
+ *  @param[in]   stream   Stream to use
+ *  @return      #hipSuccess, #hipErrorInvalidValue, #hipErrorInvalidPitchValue,
+ * #hipErrorInvalidDevicePointer, #hipErrorInvalidMemcpyDirection
+ *
+ *  @see hipMemcpy, hipMemcpy2DToArray, hipMemcpy2D, hipMemcpyFromArray, hipMemcpyToSymbol,
+ * hipMemcpyAsync
+ */
+hipError_t hipMemcpy3DAsync(const struct hipMemcpy3DParms* p, hipStream_t stream __dparm(0));
+
 // doxygen end Memory
 /**
  * @}

hipamd/include/hip/nvcc_detail/hip_runtime_api.h

@@ -607,6 +607,11 @@ inline static hipError_t hipMemcpy3D(const struct hipMemcpy3DParms *p)
     return hipCUDAErrorTohipError(cudaMemcpy3D(p));
 }
 
+inline static hipError_t hipMemcpy3DAsync(const struct hipMemcpy3DParms *p, hipStream_t stream)
+{
+    return hipCUDAErrorTohipError(cudaMemcpy3DAsync(p, stream));
+}
+
 inline static hipError_t hipMemcpy2DAsync(void* dst, size_t dpitch, const void* src, size_t spitch,
                                           size_t width, size_t height, hipMemcpyKind kind,
                                           hipStream_t stream) {