OSchip
/
llvm-project
mirror of https://github.com/THU-DSP-LAB/llvm-project.git
13
0
Fork 6
Code Issues Packages Projects Releases Wiki Activity
260,307 Commits 31 Branches 0 Tags 2.2 GiB
Commit Graph

174 Commits

Author SHA1 Message Date
Sanjay Patel 104e36a0e9 [InstCombine] add/move tests for or-of-icmps; NFC 
If we had these tests, the bug caused by https://reviews.llvm.org/rL299851 would have been caught sooner.
There's also an assert in the code that should have caught that bug, but the assert line itself has a bug.

llvm-svn: 300201
 2017-04-13 15:46:39 +00:00
Sanjay Patel 50c82c4395 [InstCombine] add fold for icmp with or mask of low bits (PR32542) 
We already have these 'and' folds:

// X & -C == -C -> X >  u ~C
// X & -C != -C -> X <= u ~C
//   iff C is a power of 2

...but we were missing the 'or' siblings.

http://rise4fun.com/Alive/n6

This should improve:
https://bugs.llvm.org/show_bug.cgi?id=32524
...but there are 2 or more other pieces to fix still.

Differential Revision: https://reviews.llvm.org/D31712

llvm-svn: 299570
 2017-04-05 17:57:05 +00:00
Sanjay Patel e7e4cc5f98 [InstCombine] add tests for missing icmp fold (PR32524) 
llvm-svn: 299557
 2017-04-05 16:21:38 +00:00
Sanjay Patel d11a03b263 [InstCombine] add test for missed vector icmp fold; NFC 
Also, move the related existing scalar test to a renamed file 
where I'm planning to add more icmp-add tests.

llvm-svn: 294487
 2017-02-08 17:37:17 +00:00
Sanjay Patel 73fc8ddb06 [InstCombine] fix operand-complexity-based canonicalization (PR28296) 
The code comments didn't match the code logic, and we didn't actually distinguish the fake unary (not/neg/fneg) 
operators from arguments. Adding another level to the weighting scheme provides more structure and can help 
simplify the pattern matching in InstCombine and other places.

I fixed regressions that would have shown up from this change in:
rL290067
rL290127

But that doesn't mean there are no pattern-matching logic holes left; some combines may just be missing regression tests.

Should fix:
https://llvm.org/bugs/show_bug.cgi?id=28296

Differential Revision: https://reviews.llvm.org/D27933

llvm-svn: 294049
 2017-02-03 21:43:34 +00:00
Sanjay Patel 5f8451afad [InstCombine] use m_APInt to allow ashr folds for vectors with splat constants 
llvm-svn: 292064
 2017-01-15 16:38:19 +00:00
Sanjay Patel 40f401776b [InstCombine] optimize unsigned icmp of increment 
Allows LLVM to optimize sequences like the following:

%add = add nuw i32 %x, 1
%cmp = icmp ugt i32 %add, %y

Into:

%cmp = icmp uge i32 %x, %y

Previously, only signed comparisons were being handled.

Decrements could also be handled, but 'sub nuw %x, 1' is currently canonicalized to
'add %x, -1' in InstCombineAddSub, losing the nuw flag. Removing that canonicalization
seems like it might have far-reaching ramifications so I kept this simple for now.

Patch by Matti Niemenmaa!

Differential Revision: https://reviews.llvm.org/D24700

llvm-svn: 291975
 2017-01-13 23:25:46 +00:00
Sanjay Patel 8d4aa10960 [InstCombine] move and add tests for icmp + shl nsw; NFC 
As discussed here:
http://lists.llvm.org/pipermail/llvm-dev/2017-January/108749.html
...we should be able to better optimize this pattern.

llvm-svn: 291262
 2017-01-06 18:57:54 +00:00
Sanjay Patel f3dda13bd2 [InstCombine] Ensure that truncated int types are legal. 
Fixes the FIXMEs in D25952 and rL285075.

Patch by bryant!

Differential Revision: https://reviews.llvm.org/D25955

llvm-svn: 285108
 2016-10-25 20:11:47 +00:00
Sanjay Patel d59f7f9047 [InstCombine] add test and code comment to show potentially misguided icmp trunc transform 
llvm-svn: 285075
 2016-10-25 15:16:39 +00:00
Sanjay Patel af91d1f81e [InstCombine] allow icmp (shr/shl) folds for vectors 
These 2 helper functions were already using APInt internally, so just
change the API and caller to allow folds for splats. The scalar
regression tests look quite thorough, so I just added a couple of
tests to prove that vectors are handled too.

These folds should be grouped with the other cmp+shift folds though.
That can be an NFC follow-up.

llvm-svn: 281663
 2016-09-15 21:35:30 +00:00
Sanjay Patel 886a542e23 [InstCombine] allow icmp (sub nsw) folds for vectors 
Also, clean up the code and comments for the existing folds in foldICmpSubConstant().

llvm-svn: 281631
 2016-09-15 18:05:17 +00:00
Sanjay Patel 514068397e [InstCombine] add vector tests for icmp (sub nsw) 
llvm-svn: 281630
 2016-09-15 17:54:47 +00:00
Sanjay Patel 7577a3d799 [InstCombine] use m_APInt to allow icmp folds using known bits for splat constant vectors 
llvm-svn: 281613
 2016-09-15 14:15:47 +00:00
Sanjay Patel 9f036b5a97 [InstCombine] add vector tests for foldICmpUsingKnownBits() 
llvm-svn: 281559
 2016-09-14 23:15:11 +00:00
Matt Arsenault e2e6cfee61 Reapply "InstCombine: Reduce trunc (shl x, K) width." 
This reapplies r272987 with a fix for infinitely looping
when the truncated value is another shift of a constant.

llvm-svn: 281379
 2016-09-13 19:43:57 +00:00
Sanjay Patel 58109abe91 [InstCombine] use m_APInt to allow icmp ult X, C folds for splat constant vectors 
llvm-svn: 281107
 2016-09-09 21:59:37 +00:00
Sanjay Patel 9b40f98357 [InstCombine] use m_APInt to allow icmp (and (sh X, Y), C2), C1 folds for splat constant vectors 
llvm-svn: 280873
 2016-09-07 22:33:03 +00:00
Sanjay Patel def931e76a [InstCombine] allow icmp (and X, C2), C1 folds for splat constant vectors 
This is a revert of r280676 which was a revert of r280637;
ie, this is r280637 again. It was speculatively reverted to
help debug buildbot failures.

llvm-svn: 280861
 2016-09-07 20:50:44 +00:00
Sanjay Patel 46f9df5b71 [InstCombine] revert r280637 because it causes test failures on an ARM bot 
http://lab.llvm.org:8011/builders/clang-cmake-armv7-a15/builds/14952/steps/ninja%20check%201/logs/FAIL%3A%20LLVM%3A%3Aicmp.ll

llvm-svn: 280676
 2016-09-05 22:36:32 +00:00
Sanjay Patel c641e9d6ff [InstCombine] allow icmp (and X, C2), C1 folds for splat constant vectors 
The code to calculate 'UsesRemoved' could be simplified.
As-is, that code is a victim of PR30273:
https://llvm.org/bugs/show_bug.cgi?id=30273

llvm-svn: 280637
 2016-09-04 20:58:27 +00:00
Sanjay Patel 541aef4661 [InstCombine] allow icmp (div X, Y), C folds for splat constant vectors 
Converting all of the overflow ops to APInt looked risky, so I've left that as a TODO.

llvm-svn: 280299
 2016-08-31 21:57:21 +00:00
Sanjay Patel 5c5311f4e5 [InstCombine] use m_APInt to allow icmp (and X, Y), C folds for splat constant vectors 
llvm-svn: 279937
 2016-08-28 18:18:00 +00:00
Sanjay Patel d398d4a39e [InstCombine] use m_APInt to allow icmp eq/ne (shr X, C2), C folds for splat constant vectors 
llvm-svn: 279677
 2016-08-24 22:22:06 +00:00
Sanjay Patel a392049419 [InstCombine] use m_APInt to allow icmp (shr exact X, Y), 0 folds for splat constant vectors 
llvm-svn: 279472
 2016-08-22 20:45:06 +00:00
Sanjay Patel 643d21a62c [InstCombine] use m_APInt to allow icmp (shl X, Y), C folds for splat constant vectors, part 4 
This concludes the fixes for icmp+shl in this series:
https://reviews.llvm.org/rL279339
https://reviews.llvm.org/rL279398
https://reviews.llvm.org/rL279399

llvm-svn: 279401
 2016-08-21 17:10:07 +00:00
Sanjay Patel 7e09f13fed [InstCombine] use m_APInt to allow icmp (shl X, Y), C folds for splat constant vectors, part 2 
This is a partial enablement (move the ConstantInt guard down).

llvm-svn: 279398
 2016-08-21 16:28:22 +00:00
Sanjay Patel fa7de606c4 [InstCombine] use m_APInt to allow icmp (shl X, Y), C folds for splat constant vectors, part 1 
This is a partial enablement (move the ConstantInt guard down) because there are many
different folds here and one of the later ones will require reworking 'isSignBitCheck'.

llvm-svn: 279339
 2016-08-19 22:33:26 +00:00
Reid Kleckner a871d3872a Fix regression in InstCombine introduced by r278944 
The intended transform is:
  // Simplify icmp eq (or (ptrtoint P), (ptrtoint Q)), 0
  // -> and (icmp eq P, null), (icmp eq Q, null).

P and Q are both pointer types, but may have different types. We need
two calls to getNullValue() to make the icmps.

llvm-svn: 279271
 2016-08-19 16:53:18 +00:00
Sanjay Patel a867afe094 [InstCombine] use m_APInt to allow icmp (shl 1, Y), C folds for splat constant vectors 
llvm-svn: 279266
 2016-08-19 16:12:16 +00:00
Sanjay Patel 57b12d3876 [InstCombine] use m_APInt to allow icmp X, C folds for splat constant vectors 
Of course, we really need to refactor and fix all of the cmp predicates, 
but this one is interesting because without it, we later perform an 
information-losing transform of icmp (shl 1, Y), C, and we can't recover
the better fold.

llvm-svn: 279263
 2016-08-19 15:40:44 +00:00
Sanjay Patel 78111a7617 [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 279259
 2016-08-19 15:27:28 +00:00
Sanjay Patel 14cdf1968f [InstCombine] add missing tests for basic icmp folds 
These are implicitly included as part of larger test cases, but they don't 
exist stand-alone (and don't happen for vectors...).

llvm-svn: 279257
 2016-08-19 15:21:45 +00:00
Sanjay Patel 6347807f87 [InstCombine] use m_APInt to allow icmp (mul X, Y), C folds for splat constant vectors 
This is a sibling of:
https://reviews.llvm.org/rL278859
https://reviews.llvm.org/rL278935
https://reviews.llvm.org/rL278945
https://reviews.llvm.org/rL279066

llvm-svn: 279077
 2016-08-18 15:44:44 +00:00
Sanjay Patel 4c5e60d95c [InstCombine] use m_APInt to allow icmp (xor X, Y), C folds for splat constant vectors 
This is a sibling of:
https://reviews.llvm.org/rL278859
https://reviews.llvm.org/rL278935
https://reviews.llvm.org/rL278945

llvm-svn: 279066
 2016-08-18 14:10:48 +00:00
Sanjay Patel 3c92db7560 [InstCombine] add test for missing vector icmp fold 
Also, add a scalar test to demonstrate one of the intermediate folds that
is necessary to accomplish the existing, multi-step test. And simplify
the vector tests to only check the final piece of that multi-step transform.

llvm-svn: 278995
 2016-08-17 22:18:57 +00:00
Sanjay Patel 4f7eb2aa95 [InstCombine] use m_APInt to allow icmp (add X, Y), C folds for splat constant vectors 
This is a sibling of:
https://reviews.llvm.org/rL278859

llvm-svn: 278935
 2016-08-17 15:24:30 +00:00
Sanjay Patel e47df1ac62 [InstCombine] use m_APInt to allow icmp (sub X, Y), C folds for splat constant vectors 
llvm-svn: 278859
 2016-08-16 21:53:19 +00:00
Sanjay Patel 46a68ba618 [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278768
 2016-08-16 00:48:38 +00:00
Sanjay Patel f1bf21c56b [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278765
 2016-08-16 00:27:12 +00:00
Sanjay Patel df77a4dbb0 [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278757
 2016-08-15 22:43:52 +00:00
Sanjay Patel 41520e1712 [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278751
 2016-08-15 21:47:50 +00:00
Sanjay Patel 3e9acec2fa [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278737
 2016-08-15 20:56:11 +00:00
Sanjay Patel b37bd6d7b7 [InstCombine] add test for missing vector icmp fold 
llvm-svn: 278727
 2016-08-15 20:02:40 +00:00
Sanjay Patel 2044a8eba9 [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278709
 2016-08-15 18:45:10 +00:00
Sanjay Patel d391b0d69e [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278704
 2016-08-15 18:26:56 +00:00
Sanjay Patel a7b9bb3785 [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278683
 2016-08-15 17:10:35 +00:00
Sanjay Patel ba1f9fbddc [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278634
 2016-08-14 22:28:50 +00:00
Sanjay Patel f6559404d5 [InstCombine] remove unnecessary function attributes from tests 
llvm-svn: 278633
 2016-08-14 21:48:21 +00:00
Sanjay Patel b44ca3bfa9 [InstCombine] add tests for missing vector icmp folds 
llvm-svn: 278632
 2016-08-14 21:36:22 +00:00