Nvidia uses fatbinaries to bundle all of their device code. This patch
adds the magic number "0x50ed55ba" used in their propeitary format to
the list of magic identifies. This is technically undocumented and could
unlikely be changed by Nvidia in the future.
Details
Details
- Reviewers
tra jdoerfert - Commits
- rG24ebdb6c255e: [CUDA] Add CUDA fatbinary magic
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Unit Tests
Unit Tests
Time | Test | |
---|---|---|
60,040 ms | x64 debian > libFuzzer.libFuzzer::large.test |
I'd move it up to case 0x50: