1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
Go to file
Simon Pilgrim 64ad084e97 [X86][SSE] Change BUILD_VECTOR interleaving ordering to improve coalescing/combine opportunities
We currently generate BUILD_VECTOR as a tree of UNPCKL shuffles of the same type:

e.g. for v4f32:

Step 1: unpcklps 0, 2 ==> X: <?, ?, 2, 0>
      : unpcklps 1, 3 ==> Y: <?, ?, 3, 1>
Step 2: unpcklps X, Y ==>    <3, 2, 1, 0>

The issue is because we are not placing sequential vector elements together early enough, we fail to recognise many combinable patterns - consecutive scalar loads, extractions etc.

Instead, this patch unpacks progressively larger sequential vector elements together:

e.g. for v4f32:

Step 1: unpcklps 0, 2 ==> X: <?, ?, 1, 0>
      : unpcklps 1, 3 ==> Y: <?, ?, 3, 2>
Step 2: unpcklpd X, Y ==>    <3, 2, 1, 0>

This does mean that we are creating UNPCKL shuffle of different value types, but the relevant combines that benefit from this are quite capable of handling the additional BITCASTs that are now included in the shuffle tree.

Differential Revision: https://reviews.llvm.org/D33864

llvm-svn: 304688
2017-06-04 20:12:04 +00:00
bindings [go bindings] Rmove duplicated conversion function definitions after r300843. 2017-04-20 19:06:11 +00:00
cmake CMake: Fix docs-llvm-man target when clang+llvm is in the same source tree 2017-05-15 09:34:23 +00:00
docs [docs] Update name of vectorization interleave flag. 2017-05-31 23:02:55 +00:00
examples [Examples] Fix some Clang-tidy modernize-use-using and Include What You Use warnings; other minor fixes (NFC). 2017-05-26 00:00:14 +00:00
include Initialized BackedgeTakenInfo.MaxOrZero. 2017-06-03 05:21:08 +00:00
lib [X86][SSE] Change BUILD_VECTOR interleaving ordering to improve coalescing/combine opportunities 2017-06-04 20:12:04 +00:00
projects Add temporary workaround to allow in-tree libc++ builds on Windows 2017-05-11 01:44:30 +00:00
resources In MSVC builds embed a VERSIONINFO resource in our exe and DLL files. 2015-06-12 15:58:29 +00:00
runtimes [CMake][runtimes] Add install target for runtimes builtins 2017-06-02 19:38:11 +00:00
test [X86][SSE] Change BUILD_VECTOR interleaving ordering to improve coalescing/combine opportunities 2017-06-04 20:12:04 +00:00
tools [PDB] Fix use after free. 2017-06-03 00:33:35 +00:00
unittests Fixed warning: must specify at least one argument for '...' parameter. 2017-06-04 05:31:03 +00:00
utils Make the Twine pretty-printer work with GDB 7.11 2017-06-04 03:27:12 +00:00
.arcconfig project_id is from another era in phabricator land and does not provide any value. 2016-09-27 15:47:29 +00:00
.clang-format
.clang-tidy .clang-tidy: correct style name is 'camelBack' not 'lowerCase'. 2016-09-13 19:04:26 +00:00
.gitignore gitignore: Ignore .vs folder (VS2017 config files) 2017-04-08 00:16:58 +00:00
CMakeLists.txt Update the documentation and CMake file for Visual Studio generators. 2017-05-25 21:01:30 +00:00
CODE_OWNERS.TXT CODE_OWNERS: Take code ownership of instruction scheduling. 2017-03-10 18:34:37 +00:00
configure Remove autoconf support 2016-01-26 21:29:08 +00:00
CREDITS.TXT update of the url 2017-05-14 07:55:01 +00:00
LICENSE.TXT Bump year to 2017 in LICENSE.txt 2017-01-12 18:02:42 +00:00
llvm.spec.in [Sparc] Implement i64 load/store support for 32-bit sparc. 2015-08-10 19:11:39 +00:00
LLVMBuild.txt
README.txt Test commit. Added a blank line. 2017-03-06 20:45:33 +00:00
RELEASE_TESTERS.TXT Update the release tester for MIPS. NFC. 2017-01-17 11:00:28 +00:00

Low Level Virtual Machine (LLVM)
================================

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.