en Copyright 2015 Java http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll#da067ed5 [AMDGPU] Set most sched model resource's BufferSize to oneUsing a BufferSize of one for memory ProcResources will result in betterILP since it more accurately models the dependencies between memory opsand their consumers on an in-order processor. After this change, thescheduler will treat the data edges from loads as blocking so thatstalls are guaranteed when waiting for data to be retreaved from memory.Since we don't actually track waitcnt here, this should do a better jobat modeling their behavior.Practically, this means that the scheduler will trigger the 'STALL'heuristic more often.This type of change needs to be evaluated experimentally. Preliminaryresults are positive.Fixes: SWDEV-282962Reviewed By: rampitecDifferential Revision: https://reviews.llvm.org/D114777 List of files: /llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll Wed, 10 Nov 2021 17:59:31 +0000 Austin Kerbow <Austin.Kerbow@amd.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll#cdfd4c5c [NFC] Removed unused prefixes in test/CodeGen/AMDGPUMore patches to follow. This covers the pertinent tests starting with e,f, and g.Differential Revision: https://reviews.llvm.org/D94124 List of files: /llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll Tue, 05 Jan 2021 22:37:29 +0000 Mircea Trofin <mtrofin@google.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll#89c8c80b AMDGPU: Change pre-gfx9 implementation of fcanonicalize to mulIf f32 denormals were enabled pre-gfx9, we would still try toimplement this with v_max_f32. Pre-gfx9, these instructions ignoredthe denormal mode and did not flush. Switch to the multiply form forf32 as a workaround which should always work in any case.This fixes conformance failures when the library implementation offmin/fmax were accidentally not inlined, forcing the assumption of noflushing on targets where denormals are not enabled by default. Thisis a workaround, since really we should not be mixing code withdifferent FP mode expectations, but prefer the lowering that will workin any mode.Now this will always use max to implement canonicalize on gfx9+. Thisis only really beneficial for f64. For f32/f16 it's a neutral choice(and worse in terms of code size in 1 case), but possibly worse forthe compiler since it does add an extra register use operand. Leavethis change for later. List of files: /llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll Wed, 22 Apr 2020 21:34:10 +0000 Matt Arsenault <Matthew.Arsenault@amd.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll#687ec75d DAG: Change behavior of fminnum/fmaxnum nodesIntroduce new versions that follow the IEEE semanticsto help with legalization that may need quieted inputs.There are some regressions from inserting unnecessarycanonicalizes when these are matched from fast mathfcmp + select which should be fixed in a future commit.llvm-svn: 344914 List of files: /llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll Mon, 22 Oct 2018 16:27:27 +0000 Matt Arsenault <Matthew.Arsenault@amd.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll#36b4b0be AMDGPU: Remove -mcpu=SILeftover from before amdgcn/r600 split.llvm-svn: 310277 List of files: /llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll Mon, 07 Aug 2017 18:30:35 +0000 Matt Arsenault <Matthew.Arsenault@amd.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll#3dbeefa9 AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernelCurrently the default C calling convention functions are treatedthe same as compute kernels. Make this explicit so the defaultcalling 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 actuallywanted a non-kernel).llvm-svn: 298444 List of files: /llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll Tue, 21 Mar 2017 21:39:51 +0000 Matt Arsenault <Matthew.Arsenault@amd.com> http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll#45bb48ea R600 -> AMDGPU renamellvm-svn: 239657 List of files: /llvm-project-15.0.7/llvm/test/CodeGen/AMDGPU/fminnum.f64.ll Sat, 13 Jun 2015 03:28:10 +0000 Tom Stellard <thomas.stellard@amd.com>