1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-22 02:33:06 +01:00
Go to file
David Sherwood cdd50ed2ff [LoopVectorize] Don't interleave scalar ordered reductions for inner loops
Consider the following loop:

  void foo(float *dst, float *src, int N) {
    for (int i = 0; i < N; i++) {
      dst[i] = 0.0;
      for (int j = 0; j < N; j++) {
        dst[i] += src[(i * N) + j];
      }
    }
  }

When we are not building with -Ofast we may attempt to vectorise the
inner loop using ordered reductions instead. In addition we also try
to select an appropriate interleave count for the inner loop. However,
when choosing a VF=1 the inner loop will be scalar and there is existing
code in selectInterleaveCount that limits the interleave count to 2
for reductions due to concerns about increasing the critical path.
For ordered reductions this problem is even worse due to the additional
data dependency, and so I've added code to simply disable interleaving
for scalar ordered reductions for now.

Test added here:

  Transforms/LoopVectorize/AArch64/strict-fadd-vf1.ll

Differential Revision: https://reviews.llvm.org/D106646
2021-07-27 17:41:01 +01:00
benchmarks
bindings [IR] Rename comdat noduplicates to comdat nodeduplicate 2021-07-20 12:47:10 -07:00
cmake [CMake] Add LIBXML2_DEFINITIONS when testing for symbol existance 2021-07-24 09:55:14 +02:00
docs [LangRef][NFC] Fix variable name in llvm.maxnum docs 2021-07-27 12:04:28 +01:00
examples [ORC] Require ExecutorProcessControl when constructing an ExecutionSession. 2021-07-27 16:53:49 +10:00
include [WebAssembly] Codegen for extmul SIMD instructions 2021-07-27 08:41:30 -07:00
lib [LoopVectorize] Don't interleave scalar ordered reductions for inner loops 2021-07-27 17:41:01 +01:00
projects [RFC][debuginfo-test] Rename debug-info lit tests for general purposes 2021-06-28 11:31:40 +01:00
resources
runtimes [runtimes] Fix umbrella component targets 2021-06-12 19:49:44 -07:00
test [LoopVectorize] Don't interleave scalar ordered reductions for inner loops 2021-07-27 17:41:01 +01:00
tools [ORC] Require ExecutorProcessControl when constructing an ExecutionSession. 2021-07-27 16:53:49 +10:00
unittests [ORC] Require ExecutorProcessControl when constructing an ExecutionSession. 2021-07-27 16:53:49 +10:00
utils [gn build] Port 2487db1f2862 2021-07-27 06:54:07 +00:00
.clang-format Test commit. 2014-03-02 13:08:46 +00:00
.clang-tidy NFC: .clang-tidy: Inherit configs from parents to improve maintainability 2021-06-08 08:25:59 -07:00
.gitattributes Fix the "git modified" issue on the preserve-comments-crlf.s. 2019-09-10 12:17:49 +00:00
.gitignore Remove .gitignore entries not relevant in the monorepo. 2021-04-07 12:25:02 -07:00
CMakeLists.txt [MLGO] Strip TF_PIP cmake variable 2021-07-22 16:28:13 -07:00
CODE_OWNERS.TXT [NFC] Update code owners file 2021-07-20 11:29:10 -07:00
configure Remove autoconf support 2016-01-26 21:29:08 +00:00
CREDITS.TXT Update personal info in CREDITS.TXT 2021-04-11 19:25:02 +08:00
LICENSE.TXT Fix typos throughout the license files that somehow I and my reviewers 2019-01-21 09:52:34 +00:00
llvm.spec.in Update structured references to the license to the new license. 2019-01-19 11:30:51 +00:00
README.txt Test commit. 2020-03-14 18:08:26 -07:00
RELEASE_TESTERS.TXT Update the list of platforms & archs 2018-12-16 14:47:16 +00:00

The LLVM Compiler Infrastructure
================================

This directory and its subdirectories contain source code for LLVM,
a toolkit for the construction of highly optimized compilers,
optimizers, and runtime environments.

LLVM is open source software. You may freely distribute it under the terms of
the license agreement found in LICENSE.txt.

Please see the documentation provided in docs/ for further
assistance with LLVM, and in particular docs/GettingStarted.rst for getting
started with LLVM and docs/README.txt for an overview of LLVM's
documentation setup.

If you are writing a package for LLVM, see docs/Packaging.rst for our
suggestions.