| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | AMDGPU: Implement computeKnownBitsForTargetNode for mbcnt | Matt Arsenault | 2017-11-13 | 1 | -2/+2 |
| | | | | | llvm-svn: 318100 | ||||
| * | AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel | Matt Arsenault | 2017-03-21 | 1 | -1/+1 |
| | | | | | | | | | | | | | Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444 | ||||
| * | AMDGPU: Remove llvm.SI.tid intrinsic | Matt Arsenault | 2016-06-17 | 1 | -3/+7 |
| | | | | | | | Mesa doesn't emit this for llvm >= 3.8 anymore. llvm-svn: 273050 | ||||
| * | AMDGPU: Avoid using 64-bit shift for i64 (shl x, 32) | Matt Arsenault | 2015-07-14 | 1 | -2/+3 |
| | | | | | | | | | | | | | | | | | | This can be done only with moves which theoretically will optimize better later. Although this transform increases the instruction count, it should be code size / cycle count neutral in the worst VALU case. It also seems to slightly improve a couple of testcases due to other DAG combines this exposes. This is probably slightly worse for the SALU case, so it might be better to handle this during moveToVALU, although then you lose some simplifications like the load width reducing in the simple testcase. llvm-svn: 242177 | ||||
| * | R600 -> AMDGPU rename | Tom Stellard | 2015-06-13 | 1 | -0/+17 |
| llvm-svn: 239657 | |||||
 |
index : bcm5719-llvm | |
| Project Ortega BCM5719 LLVM | Raptor Computing Systems |
| summaryrefslogtreecommitdiffstats |
|