15 Commits

Author SHA1 Message Date
Christopher Tetreault
792f8e1114 [SVE] Take constant fold fast path for splatted vscale vectors
This should be a perfectly reasonable operation for scalable vectors.
Currently, it only works for zeroinitializer values of
ScalableVectorType, but the fundamental operation is sound and it should
be possible to make it work for other splats

Reviewed By: david-arm

Differential Revision: https://reviews.llvm.org/D77442
2020-11-17 12:45:31 -08:00
Simon Pilgrim
6fa7030a76 [ConstProp] Remove unused check-prefixes
Just use default CHECK and remove duplicate RUN
2020-11-09 13:12:40 +00:00
Jay Foad
1417abe54c [AMDGPU] Add new llvm.amdgcn.fma.legacy intrinsic
Differential Revision: https://reviews.llvm.org/D89558
2020-10-16 17:10:21 +01:00
Amara Emerson
322d0afd87 [llvm][mlir] Promote the experimental reduction intrinsics to be first class intrinsics.
This change renames the intrinsics to not have "experimental" in the name.

The autoupgrader will handle legacy intrinsics.

Relevant ML thread: http://lists.llvm.org/pipermail/llvm-dev/2020-April/140729.html

Differential Revision: https://reviews.llvm.org/D88787
2020-10-07 10:36:44 -07:00
Sanjay Patel
149f5b573c [APFloat] convert SNaN to QNaN in convert() and raise Invalid signal
This is an alternate fix (see D87835) for a bug where a NaN constant
gets wrongly transformed into Infinity via truncation.
In this patch, we uniformly convert any SNaN to QNaN while raising
'invalid op'.
But we don't have a way to directly specify a 32-bit SNaN value in LLVM IR,
so those are always encoded/decoded by calling convert from/to 64-bit hex.

See D88664 for a clang fix needed to allow this change.

Differential Revision: https://reviews.llvm.org/D88238
2020-10-01 14:37:38 -04:00
Sanjay Patel
e34bd1e0b0 [APFloat] prevent NaN morphing into Inf on conversion (PR43907)
We shift the significand right on a truncation, but that needs to be made NaN-safe:
always set at least 1 bit in the significand.
https://llvm.org/PR43907

See D88238 for the likely follow-up (but needs some plumbing fixes before it can proceed).

Differential Revision: https://reviews.llvm.org/D87835
2020-09-24 14:02:19 -04:00
Arthur Eubanks
f4f7df037e [DIE] Remove DeadInstEliminationPass
This pass is like DeadCodeEliminationPass, but only does one pass
through a function instead of iterating on users of eliminated
instructions.

DeadCodeEliminationPass should be used in all cases.

Reviewed By: asbirlea

Differential Revision: https://reviews.llvm.org/D87933
2020-09-21 12:12:25 -07:00
Sanjay Patel
f74a334fe3 [ConstantFolding] add undef handling for fmin/fmax intrinsics
The output here may not be optimal (yet), but it should be
consistent for commuted operands (it was not before) and
correct. We can do better by checking FMF and NaN if needed.

Code in InstSimplify generally assumes that we have already
folded code like this, so it was not handling 2 constant
inputs by commuting consistently.
2020-09-19 10:31:01 -04:00
Sanjay Patel
d3b0644e22 [InstSimplify] add tests for constant folding fmin/fmax with undef op; NFC 2020-09-18 16:09:44 -04:00
Sanjay Patel
6690de098e [InstSimplify] add another test for NaN propagation; NFC 2020-09-18 09:20:26 -04:00
Sanjay Patel
c6ebe3fd00 [InstSimplify] add tests for FP constant miscompile; NFC (PR43907) 2020-09-17 12:04:39 -04:00
Eli Friedman
37f2776d1a [ConstantFold] Fold binary arithmetic on scalable vector splats.
It's a nice simplification, and it confuses instcombine if we don't do
it.

Differential Revision: https://reviews.llvm.org/D87422
2020-09-11 16:41:58 -07:00
Bryan Chan
3404add468 [EarlyCSE] Verify hash code in regression tests
As discussed in D86843, -earlycse-debug-hash should be used in more regression
tests to catch inconsistency between the hashing and the equivalence check.

Differential Revision: https://reviews.llvm.org/D86863
2020-09-04 10:40:35 -04:00
Bryan Chan
a09eef113f Replace CRLF with LF; NFC 2020-09-03 15:30:08 -04:00
Arthur Eubanks
486ed88533 [ConstProp] Remove ConstantPropagation
As discussed in
http://lists.llvm.org/pipermail/llvm-dev/2020-July/143801.html.

Currently no users outside of unit tests.

Replace all instances in tests of -constprop with -instsimplify.
Notable changes in tests:
* vscale.ll - @llvm.sadd.sat.nxv16i8 is evaluated by instsimplify, use a fake intrinsic instead
* InsertElement.ll - insertelement undef is removed by instsimplify in @insertelement_undef
llvm/test/Transforms/ConstProp moved to llvm/test/Transforms/InstSimplify/ConstProp

Reviewed By: lattner, nikic

Differential Revision: https://reviews.llvm.org/D85159
2020-08-26 15:51:30 -07:00