[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on copysign value operand (PR #97180)

llvmbot wrote: @llvm/pr-subscribers-backend-amdgpu Author: Matt Arsenault (arsenm) Changes So far the only cases that seem to benefit are the weird copysign with different typed inputs. --- Full diff: https://github.com/llvm/llvm-project/pull/97180.diff 2 Files Affected: - (modified) l

[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on copysign value operand (PR #97180)

https://github.com/arsenm ready_for_review https://github.com/llvm/llvm-project/pull/97180 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on copysign value operand (PR #97180)

arsenm wrote: > [!WARNING] > This pull request is not mergeable via GitHub because a downstack PR is > open. Once all requirements are satisfied, merge this PR as a stack href="https://app.graphite.dev/github/pr/llvm/llvm-project/97180?utm_source=stack-comment-downstack-mergeability-warning"; >

[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on copysign value operand (PR #97180)

https://github.com/arsenm created https://github.com/llvm/llvm-project/pull/97180 So far the only cases that seem to benefit are the weird copysign with different typed inputs. >From a064f83bf3579016ec5f0af50a11f2b2d3a31cf4 Mon Sep 17 00:00:00 2001 From: Matt Arsenault Date: Sat, 29 Jun 2024 0

[llvm-branch-commits] [llvm] AMDGPU: Use real copysign in fast pow (PR #97152)

https://github.com/arsenm ready_for_review https://github.com/llvm/llvm-project/pull/97152 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

[llvm-branch-commits] [llvm] AMDGPU: Use real copysign in fast pow (PR #97152)

llvmbot wrote: @llvm/pr-subscribers-backend-amdgpu Author: Matt Arsenault (arsenm) Changes Previously this would introduce some codegen regressions, but those have been avoided by simplifying demanded bits on copysign operations. --- Full diff: https://github.com/llvm/llvm-project/pull/97

[llvm-branch-commits] [llvm] AMDGPU: Use real copysign in fast pow (PR #97152)

arsenm wrote: > [!WARNING] > This pull request is not mergeable via GitHub because a downstack PR is > open. Once all requirements are satisfied, merge this PR as a stack href="https://app.graphite.dev/github/pr/llvm/llvm-project/97152?utm_source=stack-comment-downstack-mergeability-warning"; >

[llvm-branch-commits] [llvm] AMDGPU: Use real copysign in fast pow (PR #97152)

https://github.com/arsenm created https://github.com/llvm/llvm-project/pull/97152 Previously this would introduce some codegen regressions, but those have been avoided by simplifying demanded bits on copysign operations. >From 339a76086df9e50218561f568d70683f14ef1631 Mon Sep 17 00:00:00 2001 Fr

[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on fcopysign sign value (PR #97151)

llvmbot wrote: @llvm/pr-subscribers-llvm-selectiondag Author: Matt Arsenault (arsenm) Changes Math library code has quite a few places with complex bit logic that are ultimately fed into a copysign. This helps avoid some regressions in a future patch. This assumes the position in the floa

[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on fcopysign sign value (PR #97151)

https://github.com/arsenm ready_for_review https://github.com/llvm/llvm-project/pull/97151 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on fcopysign sign value (PR #97151)

arsenm wrote: > [!WARNING] > This pull request is not mergeable via GitHub because a downstack PR is > open. Once all requirements are satisfied, merge this PR as a stack href="https://app.graphite.dev/github/pr/llvm/llvm-project/97151?utm_source=stack-comment-downstack-mergeability-warning"; >

[llvm-branch-commits] [llvm] DAG: Call SimplifyDemandedBits on fcopysign sign value (PR #97151)

https://github.com/arsenm created https://github.com/llvm/llvm-project/pull/97151 Math library code has quite a few places with complex bit logic that are ultimately fed into a copysign. This helps avoid some regressions in a future patch. This assumes the position in the float type, which shou