Details
- Reviewers
jlebar - Commits
- rGbab95c70874e: [NVPTX] added match.{any,all}.sync instructions, intrinsics & builtins.
rG9941ee9529da: [NVPTX] added match.{any,all}.sync instructions, intrinsics & builtins.
rC314223: [NVPTX] added match.{any,all}.sync instructions, intrinsics & builtins.
rC314135: [NVPTX] added match.{any,all}.sync instructions, intrinsics & builtins.
rL314223: [NVPTX] added match.{any,all}.sync instructions, intrinsics & builtins.
rL314135: [NVPTX] added match.{any,all}.sync instructions, intrinsics & builtins.
Diff Detail
- Build Status
Buildable 10600 Build 10600: arc lint + arc unit
Event Timeline
clang/include/clang/Basic/BuiltinsNVPTX.def | ||
---|---|---|
419 | Nit, non-restrictive "which" should get a comma. :) | |
clang/lib/CodeGen/CGBuiltin.cpp | ||
9603 | Doing sext i1 -> i32 is going to cause us to store 0 or -1 in the pred (right?). The CUDA docs say
I'd guess that "true" probably means 1 (i.e. uext i1 -> i32) rather than -1, although, I guess we have to check. |
Addressed Justin's comments.
clang/lib/CodeGen/CGBuiltin.cpp | ||
---|---|---|
9603 | Right. It should've been ZExt. In similar places CUDA headers use "selp %r1, 1, 0, %p". |
llvm/trunk/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp | ||
---|---|---|
716 ↗ | (On Diff #116584) | I've unintentionally killed return true; here and that's what broke the texture tests. I'm not sure yet why my local tests worked. |
Nit, non-restrictive "which" should get a comma. :)