caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
445,768 Commits 1 Branch 0 Tags
40cc041d4eec7306a67bfdbfeaa8796b8d856b60
Commit Graph

1184 Commits

Author SHA1 Message Date
Matt Arsenault
6acf6661dd ConstantFolding: Ignore output denorm mode for canonicalize 
Alive2 opt plugin is now happy with the test. Fixes issue 59245
 2022-12-13 08:44:21 -05:00
Matt Arsenault
4fa54f8b81 ConstantFolding: Fix handling of canonicalize for ppc_fp128 0s 2022-12-13 08:44:20 -05:00
Sanjay Patel
9055661b95 [InstSimplify] try harder to propagate existing NaN values through FP folds 
Any undef element in a vector would trigger the whole constant
to be replaced with a canonical NaN. This propagates each
element when possible.

issue #59122
 2022-12-12 17:52:14 -05:00
Sanjay Patel
44c3c93eae [InstSimplify] add tests for vectors with NaN + partial undef; NFC 
issue #59122
 2022-12-12 17:52:14 -05:00
Bjorn Pettersson
3528e63d89 [test] Remove duplicate RUN lines in Transform tests 2022-12-08 11:47:16 +01:00
Nikita Popov
8a09875dd1 [InstSimplify] Do not remove insertvalue of undef into poison 
We cannot remove an insertvalue of undef if it inserts into a
potentially poison value. The new implementation matches that of
insertelement.

See https://alive2.llvm.org/ce/z/pCcFsW for the previously
incorrect transform.
 2022-12-08 11:35:59 +01:00
Nikita Popov
8992fce0e0 [InstSimplify] Fix check lines (NFC) 
Accidentally used the wrong binary in the previous commit.
 2022-12-08 11:22:49 +01:00
Nikita Popov
e161454d2e [InstSimplify] Add test for insertvalue of undef (NFC) 2022-12-08 11:16:46 +01:00
Roman Lebedev
5b4b842ffb [NFC] Port all InstSimplify tests to -passes= syntax 2022-12-08 02:38:45 +03:00
Matt Arsenault
7f4429c0e4 ValueTracking: Teach CannotBeOrderedLessThanZero about copysign 2022-12-06 09:01:39 -05:00
Matt Arsenault
dbca874faa ValueTracking: Teach CannotBeOrderedLessThanZero about trivial ops 
Handle canonicalize and arithmetic.fence
 2022-12-05 08:39:07 -05:00
Matt Arsenault
db0f258479 ValueTracking: Teach isKnownNeverNaN about arithmetic_fence 2022-12-05 08:39:07 -05:00
Matt Arsenault
dac496fb1f ValueTracking: Teach isKnownNeverInfinity about arithmetic.fence 2022-12-05 08:39:07 -05:00
Sanjay Patel
47f5da47f5 [InstSimplify] (X && Y) ? X : Y --> Y 
Similar to the recent fold that was added for 'or' in D138815:
https://alive2.llvm.org/ce/z/PBapTJ
 2022-11-30 15:44:48 -05:00
Sanjay Patel
d422f7e494 [InstSimplify] add tests for select with common 'and' ops; NFC 2022-11-30 15:44:48 -05:00
chenglin.bi
f297332749 [InstSimplify] Fold (X || Y) ? X : Y --> X 
(X || Y) ? X : Y --> X
https://alive2.llvm.org/ce/z/oRQJee

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D138815
 2022-11-30 10:14:17 +08:00
chenglin.bi
1fd4d91fa6 [InstSimplify] Fold !(X || Y) && X --> false 
!(X || Y) && X --> false
https://alive2.llvm.org/ce/z/693Jgv

Fix: [56654](https://github.com/llvm/llvm-project/issues/56654)
Fix: [56780](https://github.com/llvm/llvm-project/issues/56780)

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D138853
 2022-11-29 22:45:24 +08:00
chenglin.bi
0752fb57e4 [InstSimplify] Fold (X || Y) ? false : X --> false 
(X || Y) ? false : X --> false
https://alive2.llvm.org/ce/z/y93yUm

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D138700
 2022-11-29 22:08:50 +08:00
chenglin.bi
0869a96ca9 [InstSimplify] add precommit test for pattern !(X || Y) && X --> false; NFC 2022-11-29 04:07:43 +08:00
chenglin.bi
52dd5b6e95 [InstSimplify] add precommit test for pattern (X || Y) ? false : X -> false; NFC 2022-11-29 03:47:34 +08:00
chenglin.bi
617be4f37b [InstSimplify] Add precommit tests for select+or patterns; NFC 2022-11-28 22:58:39 +08:00
chenglin.bi
b400dde473 [InstSimplify] Use dominate condtion to simplify instructions 
Fix #56795

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D138542
 2022-11-26 21:41:41 +08:00
chenglin.bi
79fd94e9e7 [InstSimplify] Add precommit test for D138542; NFC 2022-11-24 09:04:33 +08:00
Matt Arsenault
841a0edd03 ConstantFolding: Constant fold some canonicalizes 
+/-0 is obviously foldable. Other non-special, non-subnormal
values are also probably OK. For denormal values, check
the calling function's denormal mode. For now, don't fold
denormals to the input for IEEE mode because as far as I know
the langref is still pretending LLVM's float isn't IEEE.

Also folds undef to 0, although NaN may make more sense. Skips
folding nans and infinities, although it should be OK to fold those
in a future change.
 2022-11-18 10:35:19 -08:00
Matt Arsenault
d2bbe5a5ff InstSimplify: Add baseline tests for canonicalize 2022-11-18 08:53:50 -08:00
Matt Arsenault
af7e80b7cb ValueTracking: Look through copysign in isKnownNeverInfinity 2022-11-17 08:52:09 -08:00
Matt Arsenault
29e4363dda ValueTracking: Look through fneg in isKnownNeverNaN 2022-11-17 08:35:49 -08:00
Matt Arsenault
ba1669c81f ValueTracking: Look through fabs and fneg in isKnownNeverInfinity 2022-11-17 00:06:15 -08:00
Matt Arsenault
d24fe812ec ValueTracking: Look through canonicalize in isKnownNeverInfinity 2022-11-17 00:06:15 -08:00
Matt Arsenault
fde4ef1973 InstSimplify: Fold arithmetic_fence as idempotent 2022-11-15 22:29:34 -08:00
Sanjay Patel
0a37290dc8 [InstSimplify] restrict logic fold with partial undef vector 
https://alive2.llvm.org/ce/z/4ncsnX

Fixes #58977
 2022-11-14 12:09:23 -05:00
Sanjay Patel
c2c48f0c96 [InstSimplify] add test for fsub with inf operand; NFC 
Verify that constant negation works with a partial undef vector.
Also, remove a bogus TODO comment on a related test.
 2022-11-11 09:17:20 -05:00
Sanjay Patel
21f1b2da95 [InstSimplify] fold fsub nnan with Inf operand 
Similar to fbc2c8f2fb, but if we have a non-canonical
fsub with constant operand 1, then flip the sign of the
Infinity:
https://alive2.llvm.org/ce/z/vKWfhW

If Infinity is operand 0, then the sign remains:
https://alive2.llvm.org/ce/z/73d97C
 2022-11-11 08:42:44 -05:00
Sanjay Patel
fbc2c8f2fb [InstSimplify] fold X +nnan Inf 
If we exclude NaN (and therefore the opposite Inf),
anything plus Inf is Inf:
https://alive2.llvm.org/ce/z/og3dj9
 2022-11-10 17:13:26 -05:00
Sanjay Patel
58167f67dc [InstSimplify] add tests for fadd/fsub with inf constant operand; NFC 2022-11-10 17:13:26 -05:00
Bjorn Pettersson
ac696ac453 Use opt -passes=<name> instead of opt -name 
Updated the RUN line in several test cases to use the new PM syntax
  opt -passes=<pipeline>
instead of the deprecated syntax
  opt -pass1 -pass2
 2022-11-08 12:15:42 +01:00
Matt Arsenault
7dd27a75a2 InstSimplify: Fold fdiv nnan ninf x, 0 -> poison 
https://alive2.llvm.org/ce/z/JxX5in
 2022-11-07 08:43:22 -08:00
Matt Arsenault
1ce5f93d03 InstSimplify: Add new baseline tests for fdiv 2022-11-07 08:43:22 -08:00
David Green
b46427b9a2 [InstSimplify] (~A & B) | ~(A | B) --> ~A with logical and 
According to https://alive2.llvm.org/ce/z/opsdrb, it is valid to convert
(~A & B) | ~(A | B) --> ~A even if the And is a Logical And. This came
up from the vector masking of predicated blocks.

Differential Revision: https://reviews.llvm.org/D137435
 2022-11-07 10:03:18 +00:00
David Green
9e885d9aab [InstSimplify] Add tests for (~A & B) | ~(A | B) --> ~A with logical And. NFC 2022-11-07 09:04:06 +00:00
Daniel Sanders
021e6e05d3 [instsimplify] Move (extelt (inselt Vec, Value, Index), Index) -> Value from InstCombine 
As requested in https://reviews.llvm.org/D135625#3858141

Differential Revision: https://reviews.llvm.org/D136099
 2022-10-17 15:22:06 -07:00
Sanjay Patel
0a1210e482 [InstSimplify] try harder to fold fmul with 0.0 operand 
https://alive2.llvm.org/ce/z/oShzr3

This was noted as a missing fold in D134876 (with additional
examples based on issue #58046).

I'm assuming that fmul with a zero operand is rare enough
that the use of ValueTracking will not noticeably increase
compile-time.

This adjusts a PowerPC codegen test that was added with D88388
because it would get folded away and no longer provide coverage
for the bug fix.
 2022-10-04 11:20:01 -04:00
Sanjay Patel
b794d72bf4 [InstSimplify] add tests for fmul with 0.0 operand; NFC 2022-10-04 10:29:53 -04:00
Sanjay Patel
ba7da14d83 Revert "[InstSimplify] reduce code duplication for fmul folds; NFC" 
This reverts commit 7b7940f9da.
This missed a test update.
 2022-10-03 11:21:23 -04:00
Sanjay Patel
7b7940f9da [InstSimplify] reduce code duplication for fmul folds; NFC 
The constant is already commuted for an fmul opcode,
but this code can be called more directly for fma,
so we have to swap for that caller. There are tests
in InstSimplify and InstCombine to verify that this
works as expected.
 2022-10-03 10:36:02 -04:00
Sanjay Patel
4490cfbaf4 [ValueTracking] peek through fpext in isKnownNeverInfinity() 
https://alive2.llvm.org/ce/z/BkNoRW
 2022-10-02 11:20:23 -04:00
Sanjay Patel
0243b424d7 [InstSimplify] add tests for FP infinity compare with fpext; NFC 2022-10-02 11:20:23 -04:00
Arthur Eubanks
e23aee7175 [test] Update some legacy PM tests 2022-09-30 11:31:02 -07:00
Sanjay Patel
3f906f057c [InstSimplify] look through vector select (shuffle) in min/max fold 
This is an extension of the existing min/max+select fold (which already
has a very large number of variations) to allow a vector shuffle because
that's what we have in the motivating example from issue #42100.

A couple of Alive2 checks of variants (I don't know how to generalize
these in Alive):
https://alive2.llvm.org/ce/z/jUFAqT

And verify the PR42100 test:
https://alive2.llvm.org/ce/z/3EcASf

It's possible there is some generalization of the fold or a
VectorCombine/SLP answer for the motivating test, but I haven't found a
better/smaller solution yet.

We can also add even more variants here as follow-up patches. For example,
we can have shuffle followed by min/max; we also don't have this
canonicalization or the reverse:
https://alive2.llvm.org/ce/z/StHD9f

Differential Revision: https://reviews.llvm.org/D134879
 2022-09-30 08:27:00 -04:00
Sanjay Patel
8bfba17b40 [InstSimplify][PhaseOrdering] add tests for vector select of min/max; NFC 
The phase ordering test is the almost unoptimized IR for the example
in issue #42100; it was passed through -mem2reg to reduce obvious
excessive load/store and other noise.

D134879
 2022-09-29 12:06:55 -04:00