Env. var NVCUDASAMPLES_ROOT is changed to NVCUDASAMPLESX_Y_ROOT where X - major ver, Y - minor ver.
Reason: NVCUDASAMPLES_ROOT contains path to CUDA SDK installed last, while NVCUDASAMPLESX_Y_ROOT contains samples of the same version as of CUDA_TOOLKIT_ROOT_DIR.
[ROCm/hip commit: 5df3a9bb3b]
// hipified to #include<hip/hip_runtime.h>
#include<cuda.h> // 1st cuda main include (Driver API)
// to eliminate
#include<cuda_runtime.h> // 2nd cuda main include (Runtime API)
HIP has one header hip_runtime.h for both CUDA APIs, thus second cuda main include directive is eliminated entirely.
[ROCm/hip commit: 7e7cfa10cc]
All but the most contrived use of macros is now properly handled -
have a look at the new testcases this commit adds. You can have
macros in kernel calls, macros spanning chunks of your arguments,
the call, call parameters, or callee can all be macros or
partially macros.
[ROCm/hip commit: 094b2b9b05]
It seems the test is already broken, but look how awesome the
error message is now:
/home/chris/HIP/tests/hipify-clang/axpy.cu:31:12: error: expected string not found in input
// CHECK: hipLaunchKernel(HIP_KERNEL_NAME(axpy), dim3(1), dim3(kDataLen), 0, 0, a, device_x, device_y);
^
<stdin>:31:2: note: scanning from here
//
^
<stdin>:33:2: note: possible intended match here
hipLaunchKernelGGL(axpy, dim3(1), dim3(kDataLen), 0, 0, a, device_x, device_y);
^
[ROCm/hip commit: 66cc45afc5]
Evgeny Mankov
Evgeny Mankov
Chris Kitching
Maneesh Gupta