|
Revision tags: llvmorg-21-init, llvmorg-19.1.7 |
|
| #
76a4c459 |
| 02-Jan-2025 |
Rajat Bajpai <rbajpai@nvidia.com> |
[InstCombine] Fix constant swap case of fcmp + fadd + sel xfrm (#119419)
The fcmp + fadd + sel => fcmp + sel + fadd xfrm performs incorrect
transformation when select branch values are swapped. Thi
[InstCombine] Fix constant swap case of fcmp + fadd + sel xfrm (#119419)
The fcmp + fadd + sel => fcmp + sel + fadd xfrm performs incorrect
transformation when select branch values are swapped. This change fixes
this.
show more ...
|
|
Revision tags: llvmorg-19.1.6, llvmorg-19.1.5 |
|
| #
de415fbb |
| 03-Dec-2024 |
Rajat Bajpai <rbajpai@nvidia.com> |
[InstCombine][FP] Fix nnan preservation for transform fcmp + sel => fmax/fmin (#117977)
Preserve `nnan` constraint only if present on both `fcmp` and `select`.
Alive2: https://alive2.llvm.org/ce/
[InstCombine][FP] Fix nnan preservation for transform fcmp + sel => fmax/fmin (#117977)
Preserve `nnan` constraint only if present on both `fcmp` and `select`.
Alive2: https://alive2.llvm.org/ce/z/ZNDjzt
show more ...
|
|
Revision tags: llvmorg-19.1.4 |
|
| #
bf1c86ce |
| 11-Nov-2024 |
Matt Arsenault <Matthew.Arsenault@amd.com> |
InstCombine: Regenerate test after "Transform (fcmp + fadd + sel)"
Update for the splat vector syntax, which landed after the last
precheck run.
|
| #
ef2d6daf |
| 11-Nov-2024 |
Rajat Bajpai <rbajpai@nvidia.com> |
[InstCombine] Transform (fcmp + fadd + sel) into (fcmp + sel + fadd) (#106492)
Transform `fcmp + fadd + sel` into `fcmp + sel + fadd` which enables the
possibility of transforming `fcmp + sel` into
[InstCombine] Transform (fcmp + fadd + sel) into (fcmp + sel + fadd) (#106492)
Transform `fcmp + fadd + sel` into `fcmp + sel + fadd` which enables the
possibility of transforming `fcmp + sel` into `maxnum/minnum`
intrinsics.
Alive2 results:
https://alive2.llvm.org/ce/z/2cmimW
https://alive2.llvm.org/ce/z/Qh9ZJt
https://alive2.llvm.org/ce/z/vtLj3R
show more ...
|