Arm MVE has multiple instructions such as VMLAVA.s8, which (in this
case) can take two 128bit vectors, sign extend the inputs to i32,
multiplying them together and sum the result into a 32bit general
purpose register. So taking 16 i8's as inputs, they can multiply and
accumulate the result into a single i32 without any rounding/truncating
along the way. There are also reduction instructions for plain integer
add and min/max, and operations that sum into a pair of 32bit registers
together treated as a 64bit integer (even though MVE does not have a
plain 64bit addition instruction). So giving the vectorizer the ability
to use these instructions both enables us to vectorize at higher
bitwidths, and to vectorize things we previously could not.
In order to do that we need a way to represent that the reduction
operation, specified with a llvm.experimental.vector.reduce when
vectorizing for Arm, occurs inside the loop not after it like most
reductions. This patch attempts to do that, teaching the vectorizer
about in-loop reductions. It does this through a vplan recipe
representing the reductions that the original chain of reduction
operations is replaced by. Cost modelling is currently just done through
a prefersInloopReduction TTI hook (which follows in a later patch).
Differential Revision: https://reviews.llvm.org/D75069
This reverts commit e9761688e4. It breaks the build:
```
~/src/llvm-project/llvm/lib/Analysis/IVDescriptors.cpp:868:10: error: no viable conversion from returned value of type 'SmallVector<[...], 8>' to function return type 'SmallVector<[...], 4>'
return ReductionOperations;
```
Arm MVE has multiple instructions such as VMLAVA.s8, which (in this
case) can take two 128bit vectors, sign extend the inputs to i32,
multiplying them together and sum the result into a 32bit general
purpose register. So taking 16 i8's as inputs, they can multiply and
accumulate the result into a single i32 without any rounding/truncating
along the way. There are also reduction instructions for plain integer
add and min/max, and operations that sum into a pair of 32bit registers
together treated as a 64bit integer (even though MVE does not have a
plain 64bit addition instruction). So giving the vectorizer the ability
to use these instructions both enables us to vectorize at higher
bitwidths, and to vectorize things we previously could not.
In order to do that we need a way to represent that the reduction
operation, specified with a llvm.experimental.vector.reduce when
vectorizing for Arm, occurs inside the loop not after it like most
reductions. This patch attempts to do that, teaching the vectorizer
about in-loop reductions. It does this through a vplan recipe
representing the reductions that the original chain of reduction
operations is replaced by. Cost modelling is currently just done through
a prefersInloopReduction TTI hook (which follows in a later patch).
Differential Revision: https://reviews.llvm.org/D75069
Forward declare DemandedBits in IVDescriptors, and move include
into the cpp file. Also drop the include from LoopUtils, which
does not need it at all.
This version contains 2 fixes for reported issues:
1. Make sure we do not try to sink terminator instructions.
2. Make sure we bail out, if we try to sink an instruction that needs to
stay in place for another recurrence.
Original message:
If the recurrence PHI node has a single user, we can sink any
instruction without side effects, given that all users are dominated by
the instruction computing the incoming value of the next iteration
('Previous'). We can sink instructions that may cause traps, because
that only causes the trap to occur later, but not on any new paths.
With the relaxed check, we also have to make sure that we do not have a
direct cycle (meaning PHI user == 'Previous), which indicates a
reduction relation, which potentially gets missed by
ReductionDescriptor.
As follow-ups, we can also sink stores, iff they do not alias with
other instructions we move them across and we could also support sinking
chains of instructions and multiple users of the PHI.
Fixes PR43398.
Reviewers: hsaito, dcaballe, Ayal, rengolin
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D69228
It broke Chromium, causing "Instruction does not dominate all uses!" errors.
See https://bugs.chromium.org/p/chromium/issues/detail?id=1022297#c1 for a
reproducer.
> If the recurrence PHI node has a single user, we can sink any
> instruction without side effects, given that all users are dominated by
> the instruction computing the incoming value of the next iteration
> ('Previous'). We can sink instructions that may cause traps, because
> that only causes the trap to occur later, but not on any new paths.
>
> With the relaxed check, we also have to make sure that we do not have a
> direct cycle (meaning PHI user == 'Previous), which indicates a
> reduction relation, which potentially gets missed by
> ReductionDescriptor.
>
> As follow-ups, we can also sink stores, iff they do not alias with
> other instructions we move them across and we could also support sinking
> chains of instructions and multiple users of the PHI.
>
> Fixes PR43398.
>
> Reviewers: hsaito, dcaballe, Ayal, rengolin
>
> Reviewed By: Ayal
>
> Differential Revision: https://reviews.llvm.org/D69228
If the recurrence PHI node has a single user, we can sink any
instruction without side effects, given that all users are dominated by
the instruction computing the incoming value of the next iteration
('Previous'). We can sink instructions that may cause traps, because
that only causes the trap to occur later, but not on any new paths.
With the relaxed check, we also have to make sure that we do not have a
direct cycle (meaning PHI user == 'Previous), which indicates a
reduction relation, which potentially gets missed by
ReductionDescriptor.
As follow-ups, we can also sink stores, iff they do not alias with
other instructions we move them across and we could also support sinking
chains of instructions and multiple users of the PHI.
Fixes PR43398.
Reviewers: hsaito, dcaballe, Ayal, rengolin
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D69228
The changes here are based on the corresponding diffs for allowing FMF on 'select':
D61917 <https://reviews.llvm.org/D61917>
As discussed there, we want to have fast-math-flags be a property of an FP value
because the alternative (having them on things like fcmp) leads to logical
inconsistency such as:
https://bugs.llvm.org/show_bug.cgi?id=38086
The earlier patch for select made almost no practical difference because most
unoptimized conditional code begins life as a phi (based on what I see in clang).
Similarly, I don't expect this patch to do much on its own either because
SimplifyCFG promptly drops the flags when converting to select on a minimal
example like:
https://bugs.llvm.org/show_bug.cgi?id=39535
But once we have this plumbing in place, we should be able to wire up the FMF
propagation and start solving cases like that.
The change to RecurrenceDescriptor::AddReductionVar() is required to prevent a
regression in a LoopVectorize test. We are intersecting the FMF of any
FPMathOperator there, so if a phi is not properly annotated, new math
instructions may not be either. Once we fix the propagation in SimplifyCFG, it
may be safe to remove that hack.
Differential Revision: https://reviews.llvm.org/D67564
llvm-svn: 372878
The changes here are based on the corresponding diffs for allowing FMF on 'select':
D61917
As discussed there, we want to have fast-math-flags be a property of an FP value
because the alternative (having them on things like fcmp) leads to logical
inconsistency such as:
https://bugs.llvm.org/show_bug.cgi?id=38086
The earlier patch for select made almost no practical difference because most
unoptimized conditional code begins life as a phi (based on what I see in clang).
Similarly, I don't expect this patch to do much on its own either because
SimplifyCFG promptly drops the flags when converting to select on a minimal
example like:
https://bugs.llvm.org/show_bug.cgi?id=39535
But once we have this plumbing in place, we should be able to wire up the FMF
propagation and start solving cases like that.
The change to RecurrenceDescriptor::AddReductionVar() is required to prevent a
regression in a LoopVectorize test. We are intersecting the FMF of any
FPMathOperator there, so if a phi is not properly annotated, new math
instructions may not be either. Once we fix the propagation in SimplifyCFG, it
may be safe to remove that hack.
Differential Revision: https://reviews.llvm.org/D67564
llvm-svn: 372866
Summary:
Currently InductionBinOps are only saved for FP induction variables, the PR extends it with non FP induction variable, so user of IVDescriptors can query the InductionBinOps for integer induction variables.
The changes in hasUnsafeAlgebra() and getUnsafeAlgebraInst() are required for the existing LIT test cases to pass. As described in the comment of the two functions, one of the requirement to return true is it is a FP induction variable. The checks was not needed because InductionBinOp was not set on non FP cases before.
https://reviews.llvm.org/D60565 depends on the patch.
Committed on behalf of @Whitney (Whitney Tsang).
Reviewers: jdoerfert, kbarton, fhahn, hfinkel, dmgreen, Meinersbur
Reviewed By: jdoerfert
Subscribers: mgorny, hiraditya, jsji, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61329
llvm-svn: 360671
Change from original commit: move test (that uses an X86 triple) into the X86
subdirectory.
Original description:
Gating vectorizing reductions on *all* fastmath flags seems unnecessary;
`reassoc` should be sufficient.
Reviewers: tvvikram, mkuper, kristof.beyls, sdesmalen, Ayal
Reviewed By: sdesmalen
Subscribers: dcaballe, huntergr, jmolloy, mcrosier, jlebar, bixia, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D57728
llvm-svn: 355889
DomTreeUpdater depends on headers from Analysis, but is in IR. This is a
layering violation since Analysis depends on IR. Relocate this code from IR
to Analysis to fix the layering violation.
llvm-svn: 353265
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
Adding a new reduction pattern match for vectorizing code similar
to TSVC s3111:
for (int i = 0; i < N; i++)
if (a[i] > b)
sum += a[i];
This patch adds support for fadd, fsub and fmull, as well as multiple
branches and different (but compatible) instructions (ex. add+sub) in
different branches.
The difference from the previous patch(https://reviews.llvm.org/D49168)
is as follows:
- Added check of fast-math property of fp-instruction to the
previous patch
- Fix/add some pattern for if-reduction.ll
Differential Revision: https://reviews.llvm.org/D54464
Patch by Takahiro Miyoshi <takahiro.miyoshi@linaro.org>
and Masakazu Ueno <masakazu.ueno@linaro.org>
llvm-svn: 347989
Adding a new reduction pattern match for vectorizing code similar to TSVC s3111:
for (int i = 0; i < N; i++)
if (a[i] > b)
sum += a[i];
This patch adds support for fadd, fsub and fmull, as well as multiple
branches and different (but compatible) instructions (ex. add+sub) in
different branches.
I have forwarded to trunk, added fsub and fmul functionality and
additional tests, but the credit goes to Takahiro, who did most of the
actual work.
Differential Revision: https://reviews.llvm.org/D49168
Patch by Takahiro Miyoshi <takahiro.miyoshi@linaro.org>.
llvm-svn: 344172
This fixes a layering violation:
Analysis/IVDescrtors.cpp can't include Transforms/Utils/BasicBlockUtils.h,
since TransformUtils depends on Analysis.
llvm-svn: 342024
Summary:
The InductionDescriptor and RecurrenceDescriptor classes basically analyze the IR to identify the respective IVs. So, it is better to have them in the "Analysis" directory instead of the "Transforms" directory.
The rationale for this is to make the Induction and Recurrence descriptor classes available for analysis passes. Currently including them in an analysis pass produces link error (http://lists.llvm.org/pipermail/llvm-dev/2018-July/124456.html).
Induction and Recurrence descriptors are moved from Transforms/Utils/LoopUtils.h|cpp to Analysis/IVDescriptors.h|cpp.
Reviewers: dmgreen, llvm-commits, hfinkel
Reviewed By: dmgreen
Subscribers: mgorny
Differential Revision: https://reviews.llvm.org/D51153
llvm-svn: 342016
David Green