Details
Details
Diff Detail
Diff Detail
- Repository
- rL LLVM
- Build Status
Buildable 26167 Build 26166: arc lint + arc unit
Event Timeline
lib/Target/AMDGPU/AMDGPUGISel.td | ||
---|---|---|
152 | Typo leaglizing | |
153–154 | I think turning everything into i32 vectors in SelectionDAG was a mistake, but I don't think GlobalISel will have the same issues with it. Floating point loads don't need to be explicitly handled anymore? | |
lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp | ||
282–296 | Why is this target specific? |
lib/Target/AMDGPU/AMDGPUGISel.td | ||
---|---|---|
153–154 | Right, loads are select based on type width only. |
lib/Target/AMDGPU/SMInstructions.td | ||
---|---|---|
689–690 | This isn't necessary since the verifier enforces there is always one MMO for g_load |
Typo leaglizing