summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/X86/vec_set-3.ll
Commit message (Collapse)AuthorAgeFilesLines
* [X86][SSE] Check vec_set BUILD_VECTOR tests on both 32 and 64-bit targetsSimon Pilgrim2017-05-101-14/+31
| | | | llvm-svn: 302683
* [X86][SSE] Regenerated the vec_set tests.Simon Pilgrim2016-04-011-9/+8
| | | | | | Replaced lots of dodgy greps with actual codegen llvm-svn: 265163
* [x86] Enable the new vector shuffle lowering by default.Chandler Carruth2014-10-041-19/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Update the entire regression test suite for the new shuffles. Remove most of the old testing which was devoted to the old shuffle lowering path and is no longer relevant really. Also remove a few other random tests that only really exercised shuffles and only incidently or without any interesting aspects to them. Benchmarking that I have done shows a few small regressions with this on LNT, zero measurable regressions on real, large applications, and for several benchmarks where the loop vectorizer fires in the hot path it shows 5% to 40% improvements for SSE2 and SSE3 code running on Sandy Bridge machines. Running on AMD machines shows even more dramatic improvements. When using newer ISA vector extensions the gains are much more modest, but the code is still better on the whole. There are a few regressions being tracked (PR21137, PR21138, PR21139) but by and large this is expected to be a win for x86 generated code performance. It is also more correct than the code it replaces. I have fuzz tested this extensively with ISA extensions up through AVX2 and found no crashes or miscompiles (yet...). The old lowering had a few miscompiles and crashers after a somewhat smaller amount of fuzz testing. There is one significant area where the new code path lags behind and that is in AVX-512 support. However, there was *extremely little* support for that already and so this isn't a significant step backwards and the new framework will probably make it easier to implement lowering that uses the full power of AVX-512's table-based shuffle+blend (IMO). Many thanks to Quentin, Andrea, Robert, and others for benchmarking assistance. Thanks to Adam and others for help with AVX-512. Thanks to Hal, Eric, and *many* others for answering my incessant questions about how the backend actually works. =] I will leave the old code path in the tree until the 3 PRs above are at least resolved to folks' satisfaction. Then I will rip it (and 1000s of lines of code) out. =] I don't expect this flag to stay around for very long. It may not survive next week. llvm-svn: 219046
* [x86] Adjust the patterns for lowering X86vzmovl nodes which don'tChandler Carruth2014-10-031-1/+1
| | | | | | | | | | | | | | perform a load to use blendps rather than movss when it is available. For non-loads, blendps is *much* faster. It can execute on two ports in Sandy Bridge and Ivy Bridge, and *three* ports on Haswell. This fixes one of the "regressions" from aggressively taking the "insertion" path in the new vector shuffle lowering. This does highlight one problem with blendps -- it isn't commuted as heavily as it should be. That's future work though. llvm-svn: 219022
* [x86] Fix an embarressing bug in the INSERTPS formation code. The maskChandler Carruth2014-09-051-1/+1
| | | | | | | | | | | | | | computation was totally wrong, but somehow it didn't really show up with llc. I've added an assert that triggers on multiple existing test cases and updated one of them to show the correct value. There appear to still be more bugs lurking around insertps's mask. =/ However, note that this only really impacts the new vector shuffle lowering. llvm-svn: 217289
* [x86] Teach the new v4i32 shuffle lowering some more tricks to recognizeChandler Carruth2014-09-041-0/+14
| | | | | | | | | | | | | | | vzext patterns and insert-element patterns that for SSE4 have dedicated instructions. With this we can enable the experimental mode in a regression test that happens to cover some of the past set of issues. You can see that the new logic does significantly better here on the floating point cases. A follow-up to this change and the previous ones will hoist the logic into helpers so it can be shared across element type sizes as in this particular case it generalizes cleanly. llvm-svn: 217136
* [x86] Clean up some tests to use FileCheck and combine two into a singleChandler Carruth2014-08-281-12/+35
| | | | | | | | | file. Changing code that is covered by these tests is just too hard to debug currently, and now it will be clear the nature of the changes. llvm-svn: 216643
* Add mcpu to tests to prevent them from using AVX instructions on Sandy ↵Craig Topper2012-04-271-1/+1
| | | | | | Bridge after r155618. llvm-svn: 155696
* Eliminate more uses of llvm-as and llvm-dis.Dan Gohman2009-09-081-1/+1
| | | | llvm-svn: 81290
* Remove obsolete -f flags.Dan Gohman2009-08-251-1/+1
| | | | llvm-svn: 79992
* Favors pshufd over shufps when shuffling elements from one vector. pshufd is ↵Evan Cheng2008-04-051-4/+3
| | | | | | faster than shufps. llvm-svn: 49244
* Remove llvm-upgrade and update tests.Tanya Lattner2008-02-211-11/+12
| | | | llvm-svn: 47432
* Convert tests using "| wc -l | grep ..." to use the count script.Dan Gohman2007-08-151-2/+2
| | | | llvm-svn: 41097
* For PR1319:Reid Spencer2007-04-161-4/+3
| | | | | | | | Remove && from the end of the lines to prevent tests from throwing run lines into the background. Also, clean up places where the same command is run multiple times by using a temporary file. llvm-svn: 36142
* For PR411:Reid Spencer2007-01-301-3/+5
| | | | | | | | Update these tests to not use the same name even though the type of the value differs. After PR411 hits, type planes will be gone and it will be illegal for a name to be used twice, regardless of type. llvm-svn: 33660
* Regression is gone, don't try to find it on clean target.Reid Spencer2007-01-171-0/+16
llvm-svn: 33296
OpenPOWER on IntegriCloud