You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Now find the `<libname>.ll` under target/amdgcn-amd-amdhsa folder and copy it to a device.ll file (or adjust the file names below).
83
-
If you work on an NVIDIA or Intel gpu, please adjust the names acordingly and open an issue to share your results (either if you succeed or fail).
84
-
First we compile our .ll files (good for manual inspections) to .bc files and clean up leftover artifacts. The cleanup is important, otherwise caching might interfere on following runs.
Especially for the last command I recommend to not fix the paths, but rather just re-generate them by copying a bare-mode openmp example and compiling it with your clang. By adding `-###` to your clang invocation, you can see the invidual steps.
91
+
Especially for the last three commands I recommend to not fix the paths, but rather just re-generate them by copying a bare-mode openmp example and compiling it with your clang. By adding `-###` to your clang invocation, you can see the invidual steps.
92
+
You can ignore other steps, e.g. the invocation of a "clang-offload-packager".
103
93
```
104
94
myclang++ -fuse-ld=lld -O3 -fopenmp -fopenmp-offload-mandatory --offload-arch=gfx90a omp_bare.cpp -o main -###
0 commit comments