Commit Graph

4201 Commits

Author SHA1 Message Date
Carl Ritson 790c75c163 [AMDGPU] Add SI_EARLY_TERMINATE_SCC0 for early terminating shader
Add pseudo instruction to allow early termination of pixel shader
anywhere based on the value of SCC.  The intention is to use this
when a mask of live lanes is updated, e.g. live lanes in WQM pass.
This facilitates early termination of shaders even when EXEC is
incomplete, e.g. in non-uniform control flow.

Reviewed By: foad

Differential Revision: https://reviews.llvm.org/D88777
2021-01-13 13:29:05 +09:00
Joe Nash 314e29ed2b [AMDGPU] Add _e64 suffix to VOP3 Insts
Previously, instructions which could be
expressed as VOP3 in addition to another
encoding had a _e64 suffix on the tablegen
record name, while those
only available as VOP3 did not. With this
patch, all VOP3s will have the _e64 suffix.
The assembly does not change, only  the mir.

Reviewed By: foad

Differential Revision: https://reviews.llvm.org/D94341

Change-Id: Ia8ec8890d47f8f94bbbdac43745b4e9dd2b03423
2021-01-12 18:33:18 -05:00
Matt Arsenault 3d39709159 AMDGPU: Remove wrapper only call limitation
This seems to only have overridden cold handling, which we probably
shouldn't do. As far as I can tell the wrapper library functions are
still inlined as appropriate.
2021-01-12 17:12:49 -05:00
Craig Topper 03c8d6a0c4 [LegalizeDAG][RISCV][PowerPC][AMDGPU][WebAssembly] Improve expansion of SETONE/SETUEQ on targets without SETO/SETUO.
If SETO/SETUO aren't legal, they'll be expanded and we'll end up
with 3 comparisons.

SETONE is equivalent to (SETOGT || SETOLT)
so if one of those operations is supported use that expansion. We
don't need both since we can commute the operands to make the other.

SETUEQ can be implemented with !(SETOGT || SETOLT) or (SETULE && SETUGE).
I've only implemented the first because it didn't look like most of the
affected targets had legal SETULE/SETUGE.

Reviewed By: frasercrmck, tlively, nemanjai

Differential Revision: https://reviews.llvm.org/D94450
2021-01-12 10:45:03 -08:00
Simon Pilgrim a4931d4fe3 [AMDGPU] Regenerate umax crash test 2021-01-12 18:02:15 +00:00
Jay Foad 794e3d94d5 [AMDGPU][GlobalISel] Remove some duplicate RUN lines
Differential Revision: https://reviews.llvm.org/D86618
2021-01-12 11:02:16 +00:00
Sebastian Neubauer 6a195491b6 [AMDGPU] Fix failing assert with scratch ST mode
In ST mode, flat scratch instructions have neither an sgpr nor a vgpr
for the address. This lead to an assertion when inserting hard clauses.

Differential Revision: https://reviews.llvm.org/D94406
2021-01-12 09:54:02 +01:00
Craig Topper b1c304c494 [CodeGen] Try to make the print of memory operand alignment a little more user friendly.
Memory operands store a base alignment that does not factor in
the effect of the offset on the alignment.

Previously the printing code only printed the base alignment if
it was different than the size. If there is an offset, the reader
would need to figure out the effective alignment themselves. This
has confused me before and someone else was recently confused on
IRC.

This patch prints the possibly offset adjusted alignment if it is
different than the size. And prints the base alignment if it is
different than the alignment. The MIR parser has been updated to
read basealign in addition to align.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D94344
2021-01-11 19:58:47 -08:00
Mircea Trofin 05e90cefeb [NFC] Disallow unused prefixes under llvm/test/CodeGen
This patch finishes addressing unused prefixes under CodeGen: 2
remaining tests fixed, and then undo-ing the lit.local.cfg changes under
various subdirs and moving the policy under CodeGen.

Differential Revision: https://reviews.llvm.org/D94430
2021-01-11 12:32:18 -08:00
Jay Foad 6dcf9207df [AMDGPU] Fix a urem combine test to test what it was supposed to 2021-01-11 13:32:34 +00:00
QingShan Zhang 7539c75bb4 [DAGCombine] Remove the check for unsafe-fp-math when we are checking the AFN
We are checking the unsafe-fp-math for sqrt but not for fpow, which behaves inconsistent.
As the direction is to remove this global option, we need to remove the unsafe-fp-math
check for sqrt and update the test with afn fast-math flags.

Reviewed By: Spatel

Differential Revision: https://reviews.llvm.org/D93891
2021-01-11 02:25:53 +00:00
Tony 2f499b9aff [AMDGPU] Add volatile support to SIMemoryLegalizer
Treat a non-atomic volatile load and store as a relaxed atomic at
system scope for the address spaces accessed. This will ensure all
relevant caches will be bypassed.

A volatile atomic is not changed and still only bypasses caches upto
the level specified by the SyncScope operand.

Differential Revision: https://reviews.llvm.org/D94214
2021-01-09 00:52:33 +00:00
Mircea Trofin a8bda3df42 [NFC] Disallow unused prefixes in CodeGen/AMDGPU
This adds the lit config, and cleans up remaining tests.

Differential Revision: https://reviews.llvm.org/D94245
2021-01-08 11:49:23 -08:00
Christudasan Devadasan ae25a397e9 AMDGPU/GlobalISel: Enable sret demotion 2021-01-08 10:56:35 +05:30
Matt Arsenault 2cbbc6e87c GlobalISel: Fail legalization on narrowing extload below memory size 2021-01-07 17:40:34 -05:00
Matt Arsenault 1f9b6ef91f GlobalISel: Add combine for G_UREM by power of 2
Really I want this in the legalizer, but this is a start.
2021-01-07 16:36:35 -05:00
Mircea Trofin ee57d30f44 [NFC] Removed unused prefixes from CodeGen/AMDGPU
Last bulk batch.

Differential Revision: https://reviews.llvm.org/D94236
2021-01-07 09:48:14 -08:00
Mircea Trofin e881a25f1e [NFC] Removed unused prefixes in CodeGen/AMDGPU
This covers tests starting with s.

Differential Revision: https://reviews.llvm.org/D94184
2021-01-07 08:00:11 -08:00
Matt Arsenault 6b7d5a928f AMDGPU/GlobalISel: Start cleaning up calling convention lowering
There are various hacks working around limitations in
handleAssignments, and the logical split between different parts isn't
correct. Start separating the type legalization to satisfy going
through the DAG infrastructure from the code required to split into
register types. The type splitting should be moved to generic code.
2021-01-07 10:36:45 -05:00
Arthur Eubanks a515342de9 [test] Pin AMDGPU/opt-pipeline.ll to legacy PM
The pipeline being tested is specifically the legacy PM pipeline.
2021-01-06 11:44:16 -08:00
Mircea Trofin 90347ab96f [NFC] Removed unused prefixes in CodeGen/AMDGPU
This covers tests starting with m-r.

Differential Revision: https://reviews.llvm.org/D94181
2021-01-06 10:32:44 -08:00
Mircea Trofin b470630913 [NFC] Removed unused prefixes from CodeGen/AMDGPU
All the 'l'-starting tests.

Differential Revision: https://reviews.llvm.org/D94151
2021-01-06 09:34:11 -08:00
Matt Arsenault ab3a3f543b AMDGPU/GlobalISel: Update fdiv lowering for denormal/ulp interaction
Change the GlobalISel fast fdiv handling to match the changes in
2531535984 and
884acbb9e1
2021-01-06 12:32:01 -05:00
Matt Arsenault 0a3cf7f476 AMDGPU/GlobalISel: Add baseline IR tests for fdiv
The fdiv lowering is currently split between an IR pass and codegen,
so make sure this works end to end. We also currently differ from the
DAG on some edge cases, which this will show in a future change.
2021-01-06 11:37:00 -05:00
Matt Arsenault 136f498919 AMDGPU: Explicitly use SelectionDAG in legacy intrinsic tests
GlobalISel will probably not support the legacy buffer intrinsics, so
don't fail when the default is switched.
2021-01-06 11:37:00 -05:00
Mircea Trofin c1cd42d698 [NFC] Removed unused prefixes in CodeGen/AMDGPU
This covers the tests starting with h-k.

Differential Revision: https://reviews.llvm.org/D94147
2021-01-05 20:22:40 -08:00
Mircea Trofin cdfd4c5c1a [NFC] Removed unused prefixes in test/CodeGen/AMDGPU
More patches to follow. This covers the pertinent tests starting with e,
f, and g.

Differential Revision: https://reviews.llvm.org/D94124
2021-01-05 19:18:30 -08:00
Changpeng Fang cb5b52a06e AMDGPU: Annotate amdgpu.noclobber for global loads only
Summary:
  This is to avoid unnecessary analysis since amdgpu.noclobber is only used for globals.

Reviewers:
  arsenm

Fixes:
   SWDEV-239161

Differential Revision:
  https://reviews.llvm.org/D94107
2021-01-05 14:47:19 -08:00
Mircea Trofin 1ebe86adf5 [NFC] Removed unused prefixes in test/CodeGen/AMDGPU
More patches to follow.

Differential Revision: https://reviews.llvm.org/D94121
2021-01-05 14:16:52 -08:00
Mircea Trofin bec987ea67 [NFC] Removed unused prefixes in CodeGen/AMDGPU
This is part of the pertinent tests, more to follow in subsequent
patches.

Differential Revision: https://reviews.llvm.org/D94114
2021-01-05 14:10:03 -08:00
Mircea Trofin a9543469d5 [NFC] Removed unused prefixes in CodeGen/AMDGPU/GlobalISel
Differential Revision: https://reviews.llvm.org/D94099
2021-01-05 12:57:17 -08:00
Jay Foad 3914bebe91 [AMDGPU] Handle v_fmac_legacy_f32 in SIFoldOperands
Convert it to v_fma_legacy_f32 if it is profitable to do so, just like
other mac instructions that are converted to their mad equivalents.

Differential Revision: https://reviews.llvm.org/D94010
2021-01-05 11:55:33 +00:00
Jay Foad 639a50e2f1 [AMDGPU] Precommit test case for D94010 2021-01-05 11:55:14 +00:00
Arthur Eubanks 8e293fe6ad [NewPM][AMDGPU] Pass TargetMachine to AMDGPUSimplifyLibCallsPass
Missed in https://reviews.llvm.org/D93863.
2021-01-04 13:48:09 -08:00
Cameron McInally 92be640bd7 [FPEnv][AMDGPU] Disable FSUB(-0,X)->FNEG(X) DAGCombine when subnormals are flushed
This patch disables the FSUB(-0,X)->FNEG(X) DAG combine when we're flushing subnormals. It requires updating the existing AMDGPU tests to use the fneg IR instruction, in place of the old fsub(-0,X) canonical form, since AMDGPU is the only backend currently checking the DenormalMode flags.

Note that this will require follow-up optimizations to make sure the FSUB(-0,X) form is handled appropriately

Differential Revision: https://reviews.llvm.org/D93243
2021-01-04 14:44:10 -06:00
Arthur Eubanks 191552344b [NewPM][AMDGPU] Make amdgpu-aa work with NewPM
An AMDGPUAA class already existed that was supposed to work with the new
PM, but it wasn't tested and was a bit broken.

Fix up the existing classes to have the right keys/parameters.
Wire up AMDGPUAA inside AMDGPUTargetMachine.

Add it to the list of alias analyses for the "default" AAManager since
in adjustPassManager() amdgpu-aa is added into the pipeline at the
beginning.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D93914
2021-01-04 12:36:27 -08:00
Arthur Eubanks 4e838ba9ea [NewPM][AMDGPU] Port amdgpu-always-inline
And add to AMDGPU opt pipeline.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D94025
2021-01-04 12:27:01 -08:00
Arthur Eubanks fd323a897c [NewPM][AMDGPU] Port amdgpu-printf-runtime-binding
And add to AMDGPU opt pipeline.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D94026
2021-01-04 12:25:50 -08:00
Arthur Eubanks e1833e7493 [NewPM][AMDGPU] Port amdgpu-unify-metadata
And add to AMDGPU opt pipeline.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D94023
2021-01-04 11:57:46 -08:00
Arthur Eubanks a5f863e076 [NewPM][AMDGPU] Port amdgpu-propagate-attributes-early/late
And add to AMDGPU opt pipeline.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D94022
2021-01-04 11:53:37 -08:00
Arthur Eubanks b8f22f9d30 [NewPM][AMDGPU] Run InternalizePass when -amdgpu-internalize-symbols
The legacy PM doesn't run EP_ModuleOptimizerEarly on -O0, so skip
running it here when given O0.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D93886
2021-01-04 11:34:40 -08:00
Roman Lebedev 4b80647367
[AMDGPU][SimplifyCFG] Teach AMDGPUUnifyDivergentExitNodes to preserve {,Post}DomTree
This is a (last big?) part of the patch series to make SimplifyCFG
preserve DomTree. Currently, it still does not actually preserve it,
even thought it is pretty much fully updated to preserve it.

Once the default is flipped, a valid DomTree must be passed into
simplifyCFG, which means that whatever pass calls simplifyCFG,
should also be smart about DomTree's.

As far as i can see from `check-llvm` with default flipped,
this is the last LLVM test batch (other than bugpoint tests)
that needed fixes to not break with default flipped.

The changes here are boringly identical to the ones i did
over 42+ times/commits recently already,
so while AMDGPU is outside of my normal ecosystem,
i'm going to go for post-commit review here,
like in all the other 42+ changes.

Note that while the pass is taught to preserve {,Post}DomTree,
it still doesn't do that by default, because simplifycfg
still doesn't do that by default, and flipping default
in this pass will implicitly flip the default for simplifycfg.
That will happen, but not right now.
2021-01-02 01:01:20 +03:00
Roman Lebedev b23b1bcc26
[NFC][CodeGen][Tests] Mark all tests that fail to preserve DomTree for SimplifyCFG as such
These tests start to fail when the SimplifyCFG's default regarding DomTree
updating is switched on, so mark them as needing changes.
2021-01-02 01:01:19 +03:00
Juneyoung Lee 9b29610228 Use unary CreateShuffleVector if possible
As mentioned in D93793, there are quite a few places where unary `IRBuilder::CreateShuffleVector(X, Mask)` can be used
instead of `IRBuilder::CreateShuffleVector(X, Undef, Mask)`.
Let's update them.

Actually, it would have been more natural if the patches were made in this order:
(1) let them use unary CreateShuffleVector first
(2) update IRBuilder::CreateShuffleVector to use poison as a placeholder value (D93793)

The order is swapped, but in terms of correctness it is still fine.

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D93923
2020-12-30 22:36:08 +09:00
Arthur Eubanks 7ecbe0c7a0 [NewPM][AMDGPU] Port amdgpu-lower-kernel-attributes
And add it to the AMDGPU opt pipeline.

This is a function pass instead of a module pass (like the legacy pass)
because it's getting added to a CGSCCPassManager, and you can't put a
module pass in a CGSCCPassManager.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D93885
2020-12-29 10:26:06 -08:00
Arthur Eubanks c2ef06d3dd [NewPM] Port infer-address-spaces
And add it to the AMDGPU opt pipeline.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D93880
2020-12-28 19:58:12 -08:00
Arthur Eubanks 0e9abcfc19 [AMDGPU][NewPM] Port amdgpu-promote-alloca(-to-vector)
And add to AMDGPU opt pipeline.

Don't pin an opt run to the legacy PM when -enable-new-pm=1 if these
passes (or passes introduced in https://reviews.llvm.org/D93863) are in
the list of passes.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D93875
2020-12-28 17:52:31 -08:00
Arthur Eubanks 9abc457724 [NewPM][AMDGPU] Port amdgpu-simplifylib/amdgpu-usenative
And add them to the pipeline via
AMDGPUTargetMachine::registerPassBuilderCallbacks(), which mirrors
AMDGPUTargetMachine::adjustPassManager().

These passes can't be unconditionally added to PassRegistry.def since
they are only present when the AMDGPU backend is enabled. And there are
no target-specific headers in llvm/include, so parsing these pass names
must occur somewhere in the AMDGPU directory. I decided the best place
was inside the TargetMachine, since the PassBuilder invokes
TargetMachine::registerPassBuilderCallbacks() anyway. If we come up with
a cleaner solution for target-specific passes in the future that's fine,
but there aren't too many target-specific IR passes living in
target-specific directories so it shouldn't be too bad to change in the
future.

Reviewed By: ychen, arsenm

Differential Revision: https://reviews.llvm.org/D93863
2020-12-28 10:38:51 -08:00
alex-t 644da789e3 [AMDGPU] Split edge to make si_if dominate end_cf
Basic block containing "if" not necessarily dominates block that is the "false" target for the if.

That "false" target block may have another predecessor besides the "if" block. IR value corresponding to the Exec mask is generated by the

si_if intrinsic and then used by the end_cf intrinsic. In this case IR verifier complains that 'Def does not dominate all uses'.

This change split the edge between the "if" block and "false" target block to make it dominated by the "if" block.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D91435
2020-12-28 17:14:02 +03:00
Juneyoung Lee 9d70dbdc2b [InstCombine] use poison as placeholder for undemanded elems
Currently undef is used as a don’t-care vector when constructing a vector using a series of insertelement.
However, this is problematic because undef isn’t undefined enough.
Especially, a sequence of insertelement can be optimized to shufflevector, but using undef as its placeholder makes shufflevector a poison-blocking instruction because undef cannot be optimized to poison.
This makes a few straightforward optimizations incorrect, such as:

```
;  https://bugs.llvm.org/show_bug.cgi?id=44185

define <4 x float> @insert_not_undef_shuffle_translate_commute(float %x, <4 x float> %y, <4 x float> %q) {
  %xv = insertelement <4 x float> %q, float %x, i32 2
  %r = shufflevector <4 x float> %y, <4 x float> %xv, <4 x i32> { 0, 6, 2, undef }
  ret <4 x float> %r ; %r[3] is undef
}
=>
define <4 x float> @insert_not_undef_shuffle_translate_commute(float %x, <4 x float> %y, <4 x float> %q) {
  %r = insertelement <4 x float> %y, float %x, i32 1
  ret <4 x float> %r ; %r[3] = %y[3], incorrect if %y[3] = poison
}

Transformation doesn't verify!
ERROR: Target is more poisonous than source
```

I’d like to suggest
1. Using poison as insertelement’s placeholder value (IRBuilder::CreateVectorSplat should be patched too)
2. Updating shufflevector’s semantics to return poison element if mask is undef

Note that poison is currently lowered into UNDEF in SelDag, so codegen part is okay.
m_Undef() matches PoisonValue as well, so existing optimizations will still fire.

The only concern is hidden miscompilations that will go incorrect when poison constant is given.
A conservative way is copying all tests having `insertelement undef` & replacing it with `insertelement poison` & run Alive2 on it, but it will create many tests and people won’t like it. :(

Instead, I’ll simply locally maintain the tests and run Alive2.
If there is any bug found, I’ll report it.

Relevant links: https://bugs.llvm.org/show_bug.cgi?id=43958 , http://lists.llvm.org/pipermail/llvm-dev/2019-November/137242.html

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D93586
2020-12-28 08:58:15 +09:00