CallInst::updateProfWeight() creates branch_weights with i64 instead of i32.
To be more consistent everywhere and remove lots of casts from uint64_t
to uint32_t, use i64 for branch_weights.
Reviewed By: davidxl
Differential Revision: https://reviews.llvm.org/D88609
Use -0.0 instead of 0.0 as the start value. The previous use of 0.0
was fine for all existing uses of this function though, as it is
always generated with fast flags right now, and thus nsz.
When trying to prove that a memory access touches only dereferenceable memory across all iterations of a loop, use the maximum exit count rather than an exact one. In many cases we can't prove exact exit counts whereas we can prove an upper bound.
The test included is for a single exit loop with a min(C,V) exit count, but the true motivation is support for multiple exits loops. It's just really hard to write a test case for multiple exits because the vectorizer (the primary user of this API), bails far before this. For multiple exits, this allows a mix of analyzeable and unanalyzable exits when only analyzeable exits are needed to prove deref.
-Oz normally does not allow loop header duplication so this loop wouldn't be
vectorized. However the vectorization pragma should override this and allow
for loop rotation.
rdar://problem/49281061
Original patch by Adam Nemet.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D59832
CallInst::updateProfWeight() creates branch_weights with i64 instead of i32.
To be more consistent everywhere and remove lots of casts from uint64_t
to uint32_t, use i64 for branch_weights.
Reviewed By: davidxl
Differential Revision: https://reviews.llvm.org/D88609
The warning would fire when calling isDereferenceableAndAlignedInLoop
with a scalable load. Calling isDereferenceableAndAlignedInLoop with a
scalable load would result in the use of the now deprecated implicit
cast of TypeSize to uint64_t through the overloaded operator.
This patch fixes this issue by:
- no longer considering vector loads as candidates in
canVectorizeWithIfConvert. This doesn't make sense in the context of
identifying scalar loads to vectorize.
- making use of getFixedSize inside isDereferenceableAndAlignedInLoop --
this removes the dependency on the deprecated interface, and will
trigger an assertion error if the function is ever called with a
scalable type.
Reviewed By: sdesmalen
Differential Revision: https://reviews.llvm.org/D89798
We should first try to constant fold the add expression and only
strengthen nowrap flags afterwards. This allows us to determine
stronger flags if e.g. only two operands are left after constant
folding (and thus "guaranteed no wrap region" code applies) or the
resulting operands are non-negative and thus nsw->nuw strengthening
applies.
LV fails with assertion checking that UF > 0. We already set UF to 1 if it is 0 except the case when IC > MaxInterleaveCount. The fix is to set UF to 1 for that case as well.
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D87679
This expands upon the inloop reductions added in e9761688e41cb9e976,
allowing them to be inserted into tail folded loops. Reductions are
generates with the form:
x = select(mask, vecop, zero)
v = vecreduce.add(x)
c = add chain, v
Where zero here is chosen as the identity value for add reductions. The
backend is then expected to fold the select and the vecreduce into a
single predicated instruction.
Most of the code is fairly straight forward, except for the creation of
blockmasks which need to ensure they are created in dominance order. The
order they are added is altered to be after any phis, keeping the
requirements for the underlying IR.
Differential Revision: https://reviews.llvm.org/D84451
We currently collect the ICmp and Add from an induction variable,
marking them as dead so that vplan values are not created for them. This
extends that to include any single use trunk from the ICmp, which allows
the Add to more readily be removed too.
This can help with costing vplan nodes, as the ICmp and Add are more
reliably removed and are not double-counted.
Differential Revision: https://reviews.llvm.org/D88873
Currently LAA uses getScalarSizeInBits to compute the size of an element
when computing the end bound of an access.
This does not work as expected for pointers to pointers, because
getScalarSizeInBits will return 0 for pointer types.
By using DataLayout to get the size of the element we can also correctly
handle pointer element types.
Note the changes to the existing test, which seems to also use the wrong
offset for the end.
Fixes PR47751.
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D88953
Regarding this bug I posted earlier: https://bugs.llvm.org/show_bug.cgi?id=47035
After reading through LLVM source code and getting familiar with VPlan I was able to vectorize the code using by enabling VPlan native path. After talking with @fhahn he suggested that I contribute this as a test case. So here it is. I tried to follow the available guides how to do this best I could. I modified IR code by hand to have more clear variable names instead of numbers.
One thing what I'd like to get input from someone is that is current CHECK lines sufficient enough to verify that the inner loop has been vectorized properly?
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D87564
We have been running tests/benchmarks downstream with tail-predication enabled
for some time now and this behaves as expected: we are not aware of any
correctness issues, and this performs better across the board than with
tail-predication disabled. Time to flip the switch!
Differential Revision: https://reviews.llvm.org/D88093
For some expressions, we can use information from loop guards when
we are looking for a maximum. This patch applies information from
loop guards to the expression used to compute the maximum backedge
taken count in howFarToZero. It currently replaces an unknown
expression X with UMin(X, Y), if the loop is guarded by
X ult Y.
This patch is minimal in what conditions it applies, and there
are a few TODOs to generalize.
This partly addresses PR40961. We will also need an update to
LV to address it completely.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D67178
Although LLVM supports vectorization of loops containing log10/sqrt, it did not support using SVML implementation of it. Added support so that when clang is invoked with -fveclib=SVML now an appropriate SVML library log2 implementation will be invoked.
Follow up on: https://reviews.llvm.org/D77114
Tests:
Added unit tests to svml-calls.ll, svml-calls-finite.ll. Can be run with llvm-lint.
Created a simple c++ file that tests log10/sqrt, and used clang+ to build it, and output final assembly.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D87169
This allows the backend to tell the vectorizer to produce inloop
reductions through a TTI hook.
For the moment on ARM under MVE this means allowing integer add
reductions of the correct size. In the future this can include integer
min/max too, under -Os.
Differential Revision: https://reviews.llvm.org/D75512
Although LLVM supports vectorization of loops containing log2, it did not support using SVML implementation of it. Added support so that when clang is invoked with -fveclib=SVML now an appropriate SVML library log2 implementation will be invoked.
Follow up on: https://reviews.llvm.org/D77114
Tests:
Added unit tests to svml-calls.ll, svml-calls-finite.ll. Can be run with llvm-lint.
Created a simple c++ file that tests log2, and used clang+ to build it, and output final assembly.
Reviewed By: wenlei, craig.topper
Differential Revision: https://reviews.llvm.org/D86730
Interleave for small loops that have reductions inside,
which breaks dependencies and expose.
This gives very significant performance improvements for some benchmarks.
Because small loops could be in very hot functions in real applications.
Differential Revision: https://reviews.llvm.org/D81416
addRuntimeChecks uses SCEVExpander, which relies on the DT/LoopInfo to
be up-to-date. Changing the CFG afterwards may invalidate some inserted
instructions, especially LCSSA phis.
Reorder the code to first update the CFG and then create the runtime
checks. This should not have any impact on the generated code, as we
adjust the CFG and generate runtime checks together.
Fixes PR47343.
The original take 1 was 6102310d81,
which taught InstSimplify to do that, which seemed better at time,
since we got EarlyCSE support for free.
However, it was proven that we can not do that there,
the simplified-to PHI would not be reachable from the original PHI,
and that is not something InstSimplify is allowed to do,
as noted in the commit ed90f15efb
that reverted it:
> It appears to cause compilation non-determinism and caused stage3 mismatches.
Then there was take 2 3e69871ab5,
which was InstCombine-specific, but it again showed stage2-stage3 differences,
and reverted in bdaa3f86a0.
This is quite alarming.
Here, let's try to change how we find existing PHI candidate:
due to the worklist order, and the way PHI nodes are inserted
(it may be inserted as the first one, or maybe not), let's look at *all*
PHI nodes in the block.
Effects on vanilla llvm test-suite + RawSpeed:
```
| statistic name | baseline | proposed | Δ | % | \|%\| |
|----------------------------------------------------|-----------|-----------|-------:|---------:|---------:|
| asm-printer.EmittedInsts | 7942329 | 7942457 | 128 | 0.00% | 0.00% |
| assembler.ObjectBytes | 254295632 | 254312480 | 16848 | 0.01% | 0.01% |
| correlated-value-propagation.NumPhis | 18412 | 18347 | -65 | -0.35% | 0.35% |
| early-cse.NumCSE | 2183283 | 2183267 | -16 | 0.00% | 0.00% |
| early-cse.NumSimplify | 550105 | 541842 | -8263 | -1.50% | 1.50% |
| instcombine.NumAggregateReconstructionsSimplified | 73 | 4506 | 4433 | 6072.60% | 6072.60% |
| instcombine.NumCombined | 3640311 | 3644419 | 4108 | 0.11% | 0.11% |
| instcombine.NumDeadInst | 1778204 | 1783205 | 5001 | 0.28% | 0.28% |
| instcombine.NumPHICSEs | 0 | 22490 | 22490 | 0.00% | 0.00% |
| instcombine.NumWorklistIterations | 2023272 | 2024400 | 1128 | 0.06% | 0.06% |
| instcount.NumCallInst | 1758395 | 1758802 | 407 | 0.02% | 0.02% |
| instcount.NumInvokeInst | 59478 | 59502 | 24 | 0.04% | 0.04% |
| instcount.NumPHIInst | 330557 | 330545 | -12 | 0.00% | 0.00% |
| instcount.TotalBlocks | 1077138 | 1077220 | 82 | 0.01% | 0.01% |
| instcount.TotalFuncs | 101442 | 101441 | -1 | 0.00% | 0.00% |
| instcount.TotalInsts | 8831946 | 8832606 | 660 | 0.01% | 0.01% |
| simplifycfg.NumHoistCommonCode | 24186 | 24187 | 1 | 0.00% | 0.00% |
| simplifycfg.NumInvokes | 4300 | 4410 | 110 | 2.56% | 2.56% |
| simplifycfg.NumSimpl | 1019813 | 999767 | -20046 | -1.97% | 1.97% |
```
So it fires 22490 times, which is less than ~24k the take 1 did,
but more than what take 2 did (22228 times)
.
It allows foldAggregateConstructionIntoAggregateReuse() to actually work
after PHI-of-extractvalue folds did their thing. Previously SimplifyCFG
would have done this PHI CSE, of all places. Additionally, allows some
more `invoke`->`call` folds to happen (+110, +2.56%).
All in all, expectedly, this catches less things overall,
but all the motivational cases are still caught, so all good.
While the original variant with doing this in InstSimplify (rightfully)
caused questions and ultimately was detected to be a culprit
of stage2-stage3 mismatch, it was expected that
InstCombine-based implementation would be fine.
But apparently it's not, as
http://lab.llvm.org:8011/builders/clang-with-thin-lto-ubuntu/builds/24095/steps/compare-compilers/logs/stdio
suggests.
Which suggests that somewhere in InstCombine there is a loop
over nondeterministically sorted container, which causes
different worklist ordering.
This reverts commit 3e69871ab5.
The original take was 6102310d81,
which taught InstSimplify to do that, which seemed better at time,
since we got EarlyCSE support for free.
However, it was proven that we can not do that there,
the simplified-to PHI would not be reachable from the original PHI,
and that is not something InstSimplify is allowed to do,
as noted in the commit ed90f15efb
that reverted it :
> It appears to cause compilation non-determinism and caused stage3 mismatches.
However InstCombine already does many different optimizations,
so it should be a safe place to do it here.
Note that we still can't just compare incoming values ranges,
because there is no guarantee that these PHI's we'd simplify to
were already re-visited and sorted.
However coming up with a test is problematic.
Effects on vanilla llvm test-suite + RawSpeed:
```
| statistic name | baseline | proposed | Δ | % | |%| |
|----------------------------------------------------|-----------|-----------|-------:|---------:|---------:|
| instcombine.NumPHICSEs | 0 | 22228 | 22228 | 0.00% | 0.00% |
| asm-printer.EmittedInsts | 7942329 | 7942456 | 127 | 0.00% | 0.00% |
| assembler.ObjectBytes | 254295632 | 254313792 | 18160 | 0.01% | 0.01% |
| early-cse.NumCSE | 2183283 | 2183272 | -11 | 0.00% | 0.00% |
| early-cse.NumSimplify | 550105 | 541842 | -8263 | -1.50% | 1.50% |
| instcombine.NumAggregateReconstructionsSimplified | 73 | 4506 | 4433 | 6072.60% | 6072.60% |
| instcombine.NumCombined | 3640311 | 3666911 | 26600 | 0.73% | 0.73% |
| instcombine.NumDeadInst | 1778204 | 1783318 | 5114 | 0.29% | 0.29% |
| instcount.NumCallInst | 1758395 | 1758804 | 409 | 0.02% | 0.02% |
| instcount.NumInvokeInst | 59478 | 59502 | 24 | 0.04% | 0.04% |
| instcount.NumPHIInst | 330557 | 330549 | -8 | 0.00% | 0.00% |
| instcount.TotalBlocks | 1077138 | 1077221 | 83 | 0.01% | 0.01% |
| instcount.TotalFuncs | 101442 | 101441 | -1 | 0.00% | 0.00% |
| instcount.TotalInsts | 8831946 | 8832611 | 665 | 0.01% | 0.01% |
| simplifycfg.NumInvokes | 4300 | 4410 | 110 | 2.56% | 2.56% |
| simplifycfg.NumSimpl | 1019813 | 999740 | -20073 | -1.97% | 1.97% |
```
So it fires ~22k times, which is less than ~24k the take 1 did.
It allows foldAggregateConstructionIntoAggregateReuse() to actually work
after PHI-of-extractvalue folds did their thing. Previously SimplifyCFG
would have done this PHI CSE, of all places. Additionally, allows some
more `invoke`->`call` folds to happen (+110, +2.56%).
All in all, expectedly, this catches less things overall,
but all the motivational cases are still caught, so all good.
Apparently, we don't do this, neither in EarlyCSE, nor in InstSimplify,
nor in (old) GVN, but do in NewGVN and SimplifyCFG of all places..
While i could teach EarlyCSE how to hash PHI nodes,
we can't really do much (anything?) even if we find two identical
PHI nodes in different basic blocks, same-BB case is the interesting one,
and if we teach InstSimplify about it (which is what i wanted originally,
https://reviews.llvm.org/D86530), we get EarlyCSE support for free.
So i would think this is pretty uncontroversial.
On vanilla llvm test-suite + RawSpeed, this has the following effects:
```
| statistic name | baseline | proposed | Δ | % | \|%\| |
|----------------------------------------------------|-----------|-----------|-------:|---------:|---------:|
| instsimplify.NumPHICSE | 0 | 23779 | 23779 | 0.00% | 0.00% |
| asm-printer.EmittedInsts | 7942328 | 7942392 | 64 | 0.00% | 0.00% |
| assembler.ObjectBytes | 273069192 | 273084704 | 15512 | 0.01% | 0.01% |
| correlated-value-propagation.NumPhis | 18412 | 18539 | 127 | 0.69% | 0.69% |
| early-cse.NumCSE | 2183283 | 2183227 | -56 | 0.00% | 0.00% |
| early-cse.NumSimplify | 550105 | 542090 | -8015 | -1.46% | 1.46% |
| instcombine.NumAggregateReconstructionsSimplified | 73 | 4506 | 4433 | 6072.60% | 6072.60% |
| instcombine.NumCombined | 3640264 | 3664769 | 24505 | 0.67% | 0.67% |
| instcombine.NumDeadInst | 1778193 | 1783183 | 4990 | 0.28% | 0.28% |
| instcount.NumCallInst | 1758401 | 1758799 | 398 | 0.02% | 0.02% |
| instcount.NumInvokeInst | 59478 | 59502 | 24 | 0.04% | 0.04% |
| instcount.NumPHIInst | 330557 | 330533 | -24 | -0.01% | 0.01% |
| instcount.TotalInsts | 8831952 | 8832286 | 334 | 0.00% | 0.00% |
| simplifycfg.NumInvokes | 4300 | 4410 | 110 | 2.56% | 2.56% |
| simplifycfg.NumSimpl | 1019808 | 999607 | -20201 | -1.98% | 1.98% |
```
I.e. it fires ~24k times, causes +110 (+2.56%) more `invoke` -> `call`
transforms, and counter-intuitively results in *more* instructions total.
That being said, the PHI count doesn't decrease that much,
and looking at some examples, it seems at least some of them
were previously getting PHI CSE'd in SimplifyCFG of all places..
I'm adjusting `Instruction::isIdenticalToWhenDefined()` at the same time.
As a comment in `InstCombinerImpl::visitPHINode()` already stated,
there are no guarantees on the ordering of the operands of a PHI node,
so if we just naively compare them, we may false-negatively say that
the nodes are not equal when the only difference is operand order,
which is especially important since the fold is in InstSimplify,
so we can't rely on InstCombine sorting them beforehand.
Fixing this for the general case is costly (geomean +0.02%),
and does not appear to catch anything in test-suite, but for
the same-BB case, it's trivial, so let's fix at least that.
As per http://llvm-compile-time-tracker.com/compare.php?from=04879086b44348cad600a0a1ccbe1f7776cc3cf9&to=82bdedb888b945df1e9f130dd3ac4dd3c96e2925&stat=instructions
this appears to cause geomean +0.03% compile time increase (regression),
but geomean -0.01%..-0.04% code size decrease (improvement).
This implements 2 different vectorisation fallback strategies if tail-folding
fails: 1) don't vectorise at all, or 2) vectorise using a scalar epilogue. This
can be controlled with option -prefer-predicate-over-epilogue, that has been
changed to take a numeric value corresponding to the tail-folding preference
and preferred fallback.
Patch by: Pierre van Houtryve, Sjoerd Meijer.
Differential Revision: https://reviews.llvm.org/D79783
MVE Gather scatter codegeneration is looking a lot better than it used
to, but still has some issues. The instructions we currently model as 1
cycle per element, which is a bit low for some cases. Increasing the
cost by the MVECostFactor brings them in-line with our other instruction
costs. This will have the effect of only generating then when the extra
benefit is more likely to overcome some of the issues. Notably in
running out of registers and vectorizing loops that could otherwise be
SLP vectorized.
In the short-term whilst we look at other ways of dealing with those
more directly, we can increase the costs of gathers to make them more
likely to be beneficial when created.
Differential Revision: https://reviews.llvm.org/D86444
The legacy LoopVectorize has a dependency on InjectTLIMappingsLegacy.
That cannot be expressed in the new PM since they are both normal
passes. Explicitly add -inject-tli-mappings as a pass.
Follow-up to https://reviews.llvm.org/D86492.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D86561
This adapts LV to the new semantics of get.active.lane.mask as discussed in
D86147, which means that the LV now emits intrinsic get.active.lane.mask with
the loop tripcount instead of the backedge-taken count as its second argument.
The motivation for this is described in D86147.
Differential Revision: https://reviews.llvm.org/D86304
If gather/scatters are enabled, ARMTargetTransformInfo now allows
tail predication for loops with a much wider range of strides, up
to anything that is loop invariant.
Differential Revision: https://reviews.llvm.org/D85410
As part of D84741, this adds a target hook for the
preferPredicatedReductionSelect option and makes use
of it under MVE, allowing us to tail predicate most
reduction loops.
Differential Revision: https://reviews.llvm.org/D85980
The normal scheme for tail folding reductions is to use:
loop:
p = phi(0, a)
mask = ...
x = masked_load(..., mask)
a = add(x, p)
s = select(mask, a, p)
This means we need to keep the register p and a alive out of the loop, plus
the mask. On a target with predicated operations we can instead generate
the phi as p = phi(0, s). This ensures the select in the loop and we can
fold select(m, add(a, b), c) to something like a vaddt c, a, b using the
m predicate. This in turn allows us to tail predicate the entire loop.
Differential Revision: https://reviews.llvm.org/D84741
D81345 appears to accidentally disables vectorization when explicitly
enabled. As PGSO isn't currently accessible from LoopAccessInfo, revert back to
the vectorization with versioning-for-unit-stride for PGSO.
Differential Revision: https://reviews.llvm.org/D85784
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
No widening decisions will be computed for instructions outside the
loop. Do not try to get a widening decision. The load/store will be just
a scalar load, so treating at as normal should be fine I think.
Fixes PR46950.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D85087
If an analysis is actually invalidated, there's already a log statement
for that: 'Invalidating analysis: FooAnalysis'.
Otherwise the statement is not very useful.
Reviewed By: asbirlea, ychen
Differential Revision: https://reviews.llvm.org/D84981
This removes some unneeded block masks when we don't have any
reductions. It should not have any effect on codegen as the values
created are dead anyway.
Differential Revision: https://reviews.llvm.org/D81415
This patch uses the feature added in D79162 to fix the cost of a
sext/zext of a masked load, or a trunc for a masked store.
Previously, those were considered cheap or even free, but it's
not the case as we cannot split the load in the same way we would for
normal loads.
This updates the costs to better reflect reality, and adds a test for it
in test/Analysis/CostModel/ARM/cast.ll.
It also adds a vectorizer test that showcases the improvement: in some
cases, the vectorizer will now choose a smaller VF when
tail-predication is enabled, which results in better codegen. (Because
if it were to use a higher VF in those cases, the code we see above
would be generated, and the vmovs would block tail-predication later in
the process, resulting in very poor codegen overall)
Original Patch by Pierre van Houtryve
Differential Revision: https://reviews.llvm.org/D79163
Summary: To match NewPM name. Also the new name is clearer and more consistent.
Subscribers: jvesely, nhaehnle, hiraditya, asbirlea, kerbowa, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D84542
It was getting difficult to see which test was in which file, so this
reorganises the test files so that now all filenames start with tail-folding-*
followed by a more descriptive name what that group of tests check.
This patch enables the LoopVectorizer to build a phi of pointer
type and provide the vector loads and stores with vector type
getelementptrs built from the pointer induction variable, which
produces much less instructions than the previous approach of
creating scalar getelementpointers and glue them together to a
vector.
Differential Revision: https://reviews.llvm.org/D81267
If a vector body has live-out values, it is probably a reduction, which needs a
final reduction step after the loop. MVE has a VADDV instruction to reduce
integer vectors, but doesn't have an equivalent one for float vectors. A
live-out value that is not recognised as reduction later in the optimisation
pipeline will result in the tail-predicated loop to be reverted to a
non-predicated loop and this is very expensive, i.e. it has a significant
performance impact, which is what we hope to avoid with fine tuning the ARM TTI
hook preferPredicateOverEpilogue implementation.
Differential Revision: https://reviews.llvm.org/D82953
This refactors option -disable-mve-tail-predication to take different arguments
so that we have 1 option to control tail-predication rather than several
different ones.
This is also a prep step for D82953, in which we want to reject reductions
unless that is requested with this option.
Differential Revision: https://reviews.llvm.org/D83133
This patch fixes D81345 and PR46652.
If a loop with a small trip count is compiled w/o -Os/-Oz, Loop Access Analysis
still generates runtime checks for unit strides that will version the loop.
In such cases, the loop vectorizer should either re-run the analysis or bail-out
from vectorizing the loop, as done prior to D81345. The latter is applied for
now as the former requires refactoring.
Differential Revision: https://reviews.llvm.org/D83470
Currently the DomTree is not kept up to date for additional blocks
generated in the vector loop, for example when vectorizing with
predication. SCEVExpander relies on dominance checks when looking for
existing instructions to re-use and in some cases that can lead to the
expander picking instructions that do not actually dominate their insert
point (e.g. as in PR46525).
Unfortunately keeping the DT up-to-date is a bit tricky, because the CFG
is only patched up after generating code for a block. For now, we can
just use the vector loop header, as this ensures the inserted
instructions dominate all uses in the vector loop. There should be no
noticeable impact on the generated code, as other passes should sink
those instructions, if profitable.
Fixes PR46525.
Reviewers: Ayal, gilr, mkazantsev, dmgreen
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D83288
If a loop is in a function marked OptSize, Loop Access Analysis should refrain
from generating runtime checks for unit strides that will version the loop.
If a loop is in a function marked OptSize and its vectorization is enabled, it
should be vectorized w/o any versioning.
Fixes PR46228.
Differential Revision: https://reviews.llvm.org/D81345
This adjusts the MVE fp16 cost model, similar to how we already do for
integer casts. It uses the base cost of 1 per cvt for most fp extend /
truncates, but adjusts it for loads and stores where we know that a
extending load has been used to get the load into the correct lane, and
only an MVE VCVTB is then needed.
Differential Revision: https://reviews.llvm.org/D81813
This alters getMemoryOpCost to use the Base TargetTransformInfo version
that includes some additional checks for whether extending loads are
legal. This will generally have the effect of making <2 x ..> and some
<4 x ..> loads/stores more expensive, which in turn should help favour
larger vector factors.
Notably it alters the cost of a <4 x half>, which with the current
codegen will be expensive if it is not extended.
Differential Revision: https://reviews.llvm.org/D82456
This patch enables the LoopVectorizer to build a phi of pointer
type and provide the vector loads and stores with vector type
getelementptrs built from the pointer induction variable, which
produces much less instructions than the previous approach of
creating scalar getelementpointers and glue them together to a
vector.
Differential Revision: https://reviews.llvm.org/D81267
D79164/2596da31740f changed getCFInstrCost to return 1 per default.
AArch64 did not have its own implementation, hence the throughput cost
of CFI instructions is overestimated. On most cores, most branches should
be predicated and essentially free throughput wise.
This restores a 9% performance regression on a SPEC2006 benchmark on
AArch64 with -O3 LTO & PGO.
This patch effectively restores pre 2596da3174 behavior for AArch64
and undoes the AArch64 test changes of the patch.
Reviewers: samparker, dmgreen, anemet
Reviewed By: samparker
Differential Revision: https://reviews.llvm.org/D82755
This emits new IR intrinsic @llvm.get.active.mask for tail-folded vectorised
loops if the intrinsic is supported by the backend, which is checked by
querying TargetTransform hook emitGetActiveLaneMask.
This intrinsic creates a mask representing active and inactive vector lanes,
which is used by the masked load/store instructions that are created for
tail-folded loops. The semantics of @llvm.get.active.mask are described here in
LangRef:
https://llvm.org/docs/LangRef.html#llvm-get-active-lane-mask-intrinsics
This intrinsic is also used to provide a hint to the backend. That is, the
second argument of the intrinsic represents the back-edge taken count of the
loop. For MVE, for example, we use that to set up tail-predication, which is a
new form of predication in MVE for vector loops that implicitely predicates the
last vector loop iteration by implicitely setting active/inactive lanes, i.e.
the tail loop is predicated. In order to set up a tail-predicated vector loop,
we need to know the number of data elements processed by the vector loop, which
corresponds the the tripcount of the scalar loop, which we can now reconstruct
using @llvm.get.active.mask.
Differential Revision: https://reviews.llvm.org/D79100
Have BasicTTI call the base implementation so that both agree on the
default behaviour, which the default being a cost of '1'. This has
required an X86 specific implementation as it seems to be very
reliant on those instructions being free. Changes are also made to
AMDGPU so that their implementations distinguish between cost kinds,
so that the unrolling isn't affected. PowerPC also has its own
implementation to prevent changes to the reg-usage vectorizer test.
The cost model test changes now reflect that ret instructions are not
generally free.
Differential Revision: https://reviews.llvm.org/D79164
Alternative approach to D80570.
canCheckPtrAtRT already contains checks the figure out for which alias
sets runtime checks are needed. But it currently sets CanDoRT to false
for alias sets for which we cannot do RT checks but also do not need
any.
If we know that we do not need RT checks based on the number of
reads/writes in the alias set, we can skip processing the AS.
This patch also adds an assertion to ensure that DepCands does not
contain more than one write from the alias set.
Reviewers: Ayal, anemet, hfinkel, dmgreen
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D80622
Similar to a recent change to the X86 backend, this changes things so
that we always produce a reduction intrinsics for all reduction types,
not just the legal ones. This gives a better chance in the backend to
custom lower them to something more suitable for MVE. Especially for
something like fadd the in-order reduction produced during DAG lowering
is already better than the shuffles produced in the midend, and we can
do even better with a bit of custom lowering.
Differential Revision: https://reviews.llvm.org/D81398
getOrCreateTripCount is used to generate code for the outer loop, but it
requires a computable backedge taken counts. Check that in the VPlan
native path.
Reviewers: Ayal, gilr, rengolin, sguggill
Reviewed By: sguggill
Differential Revision: https://reviews.llvm.org/D81088
Motivating examples are seen in the PhaseOrdering tests based on:
https://bugs.llvm.org/show_bug.cgi?id=43953#c2 - if we have
intrinsics there, some pass can fold them.
The intrinsics are still named "experimental" at this point, but
if there is no fallout from this patch, that will be a good
indicator that it is safe to finalize them.
Differential Revision: https://reviews.llvm.org/D80867
LV currently only supports power of 2 vectorization factors, which has
been made explicit with the assertion added in
840450549c.
However, if the widest type is not a power-of-2 the computed MaxVF won't
be a power-of-2 either. This patch updates computeFeasibleMaxVF to
ensure the returned value is a power-of-2 by rounding down to the
nearest power-of-2.
Fixes PR46139.
Reviewers: Ayal, gilr, rengolin
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D80870
The -reassociate pass tends to transform this kind of pattern into
something that is worse for vectorization and codegen. See PR43953:
https://bugs.llvm.org/show_bug.cgi?id=43953
Follows-up the FP version of the same transform:
rGa0ce2338a083
This intrinsic implements IEEE-754 operation roundToIntegralTiesToEven,
and performs rounding to the nearest integer value, rounding halfway
cases to even. The intrinsic represents the missed case of IEEE-754
rounding operations and now llvm provides full support of the rounding
operations defined by the standard.
Differential Revision: https://reviews.llvm.org/D75670
If a loop has a constant trip count known to be a multiple of MaxVF (times user
UF), LV infers that no tail will be generated for any chosen VF. This relies on
the chosen VF's being powers of 2 bound by MaxVF, and assumes MaxVF is a power
of 2. Make sure the latter holds, in particular when MaxVF is set by a memory
dependence distance which may not be a power of 2.
Differential Revision: https://reviews.llvm.org/D80491
Currently we unconditionally get the first lane of the condition
operand, even if we later use the full vector condition. This can result
in some unnecessary instructions being generated.
Suggested as follow-up in D80219.
Summary:
When handling loops whose VF is 1, fold-tail vectorization sets the
backedge taken count of the original loop with a vector of a single
element. This causes type-mismatch during instruction generartion.
The purpose of this patch is toto address the case of VF==1.
Reviewer: Ayal (Ayal Zaks), bmahjour (Bardia Mahjour), fhahn (Florian Hahn), gilr (Gil Rapaport), rengolin (Renato Golin)
Reviewed By: Ayal (Ayal Zaks), bmahjour (Bardia Mahjour), fhahn (Florian Hahn)
Subscribers: Ayal (Ayal Zaks), rkruppe (Hanna Kruppe), bmahjour (Bardia Mahjour), rogfer01 (Roger Ferrer Ibanez), vkmr (Vineet Kumar), bollu (Siddharth Bhat), hiraditya (Aditya Kumar), llvm-commits (Mailing List llvm-commits)
Tag: LLVM
Differential Revision: https://reviews.llvm.org/D79976
LV considers an internally computed MaxVF to decide if a constant trip-count is
a multiple of any subsequently chosen VF, and conclude that no scalar remainder
iterations (tail) will be left for Fold Tail to handle. If an external VF is
provided via -force-vector-width, it must be considered instead of the internal
MaxVF.
If an external UF is provided via -force-vector-interleave, it too must be
considered in addition to MaxVF or user VF.
Fixes PR45679.
Differential Revision: https://reviews.llvm.org/D80085
For IR generated by a compiler, this is really simple: you just take the
datalayout from the beginning of the file, and apply it to all the IR
later in the file. For optimization testcases that don't care about the
datalayout, this is also really simple: we just use the default
datalayout.
The complexity here comes from the fact that some LLVM tools allow
overriding the datalayout: some tools have an explicit flag for this,
some tools will infer a datalayout based on the code generation target.
Supporting this properly required plumbing through a bunch of new
machinery: we want to allow overriding the datalayout after the
datalayout is parsed from the file, but before we use any information
from it. Therefore, IR/bitcode parsing now has a callback to allow tools
to compute the datalayout at the appropriate time.
Not sure if I covered all the LLVM tools that want to use the callback.
(clang? lli? Misc IR manipulation tools like llvm-link?). But this is at
least enough for all the LLVM regression tests, and IR without a
datalayout is not something frontends should generate.
This change had some sort of weird effects for certain CodeGen
regression tests: if the datalayout is overridden with a datalayout with
a different program or stack address space, we now parse IR based on the
overridden datalayout, instead of the one written in the file (or the
default one, if none is specified). This broke a few AVR tests, and one
AMDGPU test.
Outside the CodeGen tests I mentioned, the test changes are all just
fixing CHECK lines and moving around datalayout lines in weird places.
Differential Revision: https://reviews.llvm.org/D78403
This was reverted because of a miscompilation. At closer inspection, the
problem was actually visible in a changed llvm regression test too. This
one-line follow up fix/recommit will splat the IV, which is what we are trying
to avoid if unnecessary in general, if tail-folding is requested even if all
users are scalar instructions after vectorisation. Because with tail-folding,
the splat IV will be used by the predicate of the masked loads/stores
instructions. The previous version omitted this, which caused the
miscompilation. The original commit message was:
If tail-folding of the scalar remainder loop is applied, the primary induction
variable is splat to a vector and used by the masked load/store vector
instructions, thus the IV does not remain scalar. Because we now mark
that the IV does not remain scalar for these cases, we don't emit the vector IV
if it is not used. Thus, the vectoriser produces less dead code.
Thanks to Ayal Zaks for the direction how to fix this.
- Specifically check for sext/zext users which have 'long' form NEON
instructions.
- Add more entries to the table for sext/zexts so that we can report
more accurately the number of vmovls required for NEON.
- Pass the instruction to the pass implementation.
Differential Revision: https://reviews.llvm.org/D79561
If tail-folding of the scalar remainder loop is applied, the primary induction
variable is splat to a vector and used by the masked load/store vector
instructions, thus the IV does not remain scalar. Because we now mark
that the IV does not remain scalar for these cases, we don't emit the vector IV
if it is not used. Thus, the vectoriser produces less dead code.
Thanks to Ayal Zaks for the direction how to fix this.
Differential Revision: https://reviews.llvm.org/D78911
The improvements to the x86 vector insert/extract element costs in D74976 resulted in the estimated costs for vector initialization and scalarization increasing higher than should be expected. This is particularly noticeable on pre-SSE4 targets where the available of legal INSERT_VECTOR_ELT ops is more limited.
This patch does 2 things:
1 - it implements X86TTIImpl::getScalarizationOverhead to more accurately represent the typical costs of a ISD::BUILD_VECTOR pattern.
2 - it adds a DemandedElts mask to getScalarizationOverhead to permit the SLP's BoUpSLP::getGatherCost to be rewritten to use it directly instead of accumulating raw vector insertion costs.
This fixes PR45418 where a v4i8 (zext'd to v4i32) was no longer vectorizing.
A future patch should extend X86TTIImpl::getScalarizationOverhead to tweak the EXTRACT_VECTOR_ELT scalarization costs as well.
Reviewed By: @craig.topper
Differential Revision: https://reviews.llvm.org/D78216
The crash that caused the original revert has been fixed in
a3c964a278. I also added a reduced version of the crash reproducer.
This reverts the revert commit 2107af9ccf.
When folding tail, branch taken count is computed during initial VPlan execution
and recorded to be used by the compare computing the loop's mask. This recording
should directly set the State, instead of reusing Value2VPValue mapping which
serves original Values present prior to vectorization.
The branch taken count may be a constant Value, which may be used elsewhere in
the loop; trying to employ Value2VPValue for both leads to the issue reported in
https://reviews.llvm.org/D76992#inline-721028
Differential Revision: https://reviews.llvm.org/D78847
One of transforms the loop vectorizer makes is LCSSA formation. In some cases it
is the only transform it makes. We should not drop CFG analyzes if only LCSSA was
formed and no actual CFG changes was made.
We should think of expanding this logic to other passes as well, and maybe make
it a part of PM framework.
Reviewed By: Florian Hahn
Differential Revision: https://reviews.llvm.org/D78360
This will allow us to use the datalayout to disambiguate other
constructs in IR, like load alignment. Split off from D78403.
Differential Revision: https://reviews.llvm.org/D78413
First-order recurrences require special treatment when they are live-out;
such treatment is provided by fixFirstOrderRecurrence(), so they should be
included in AllowedExit set.
(Should probably have been included originally in D16197.)
Fixes PR45526: AllowedExit set is used by prepareToFoldTailByMasking() to
check whether the treatment for live-outs also holds when folding the tail,
which is not (yet) the case for first-order recurrences.
Differential Revision: https://reviews.llvm.org/D78210
Cost-modeling decisions are tied to the compute interleave groups
(widening decisions, scalar and uniform values). When invalidating the
interleave groups, those decisions also need to be invalidated.
Otherwise there is a mis-match during VPlan construction.
VPWidenMemoryRecipes created initially are left around w/o converting them
into VPInterleave recipes. Such a conversion indeed should not take place,
and these gather/scatter recipes may in fact be right. The crux is leaving around
obsolete CM_Interleave (and dependent) markings of instructions along with
their costs, instead of recalculating decisions, costs, and recipes.
Alternatively to forcing a complete recompute later on, we could try
to selectively invalidate the decisions connected to the interleave
groups. But we would likely need to run the uniform/scalar value
detection parts again anyways and the extra complexity is probably not
worth it.
Fixes PR45572.
Reviewers: gilr, rengolin, Ayal, hsaito
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D78298
Fix an assert introduced in 41ed5d856c1: a phi with a single predecessor and a
mask is a valid case which is already supported by the code.
Differential Revision: https://reviews.llvm.org/D78115
D77635 added support to recognise primary induction variables for counting-down
loops. This allows us to fold the scalar tail loop into the main vector body,
which we need for MVE tail-predication. This adds some ARM tail-folding test
cases that we want to support.
This test was extracted from D76838, which implemented a different approach to
reverse and thus find a primary induction variable.
Introduce a new VPWidenCanonicalIVRecipe to generate a canonical vector
induction for use in fold-tail-with-masking, if a primary induction is absent.
The canonical scalar IV having start = 0 and step = VF*UF, created during code
-gen to control the vector loop, is widened into a canonical vector IV having
start = {<Part*VF, Part*VF+1, ..., Part*VF+VF-1> for 0 <= Part < UF} and
step = <VF*UF, VF*UF, ..., VF*UF>.
Differential Revision: https://reviews.llvm.org/D77635
Summary:
Add mapping from exp2 math functions
to corresponding SVML calls.
This is a follow up and extension for llvm diff
https://reviews.llvm.org/D19544
Test Plan:
- update test case and run ninja check.
- run tests locally
Reviewers: wenlei, hoyFB, mmasten, mzolotukhin, spatel
Reviewed By: spatel
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77114
In InnerLoopVectorizer::getOrCreateTripCount, when the backedge taken
count is a SCEV add expression, its type is defined by the type of the
last operand of the add expression.
In the test case from PR45259, this last operand happens to be a
pointer, which (according to llvm::Type) does not have a primitive size
in bits. In this case, LoopVectorize fails to truncate the SCEV and
crashes as a result.
Uing ScalarEvolution::getTypeSizeInBits makes the truncation work as expected.
https://bugs.llvm.org/show_bug.cgi?id=45259
Differential Revision: https://reviews.llvm.org/D76669
Also adds a force-reduction-intrinsics option for testing, for forcing
the generation of reduction intrinsics even when the backend is not
requesting them.
This tries to improve the accuracy of extract/insert element costs by accounting for subvector extraction/insertion for >128-bit vectors and the shuffling of elements to/from the 0'th index.
It also adds INSERTPS for f32 types and PINSR/PEXTR costs for integer types (at the moment we assume the same cost as MOVD/MOVQ - which isn't always true).
Differential Revision: https://reviews.llvm.org/D74976
I added test cases that rely on the availability of the PPC target into
the general directory for the loop vectorizer. This causes failures on
bots that don't build the PPC target. Moving them to the PowerPC directory
to fix this.
A recent commit
(https://reviews.llvm.org/rG66c120f02560ef528a60924104ead66f330190f1) changed
the cost for calls to functions that have a vector version for some
vectorization factor. However, no check is performed for whether the
vectorization factor matches the current one being cost modeled. This leads to
attempts to widen call instructions to a vectorization factor for which such a
function does not exist, which in turn leads to an assertion failure.
This patch adds the check for vectorization factor (i.e. not just that the
called function has a vector version for some VF, but that it has a vector
version for this VF).
Differential revision: https://reviews.llvm.org/D74944
Summary:
Previosly we simply always said that `SCEVMinMaxExpr` is too costly to expand.
But this isn't really true, it expands into just a comparison+swap pair.
And again much like with add/mul, there will be one less such pair
than the number of operands. And we need to count the cost of operands themselves.
This does change a number of testcases, and as far as i can tell,
all of these changes are improvements, in the sense that
we fixed up more latches to do the [in]equality comparison.
This concludes cost-modelling changes, no other SCEV expressions exist as of now.
This is a part of addressing [[ https://bugs.llvm.org/show_bug.cgi?id=44668 | PR44668 ]].
Reviewers: reames, mkazantsev, wmi, sanjoy
Reviewed By: mkazantsev
Subscribers: hiraditya, javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73744
Summary:
Currently, `SCEVExpander::isHighCostExpansionHelper()` has the following logic:
```
if (auto *UDivExpr = dyn_cast<SCEVUDivExpr>(S)) {
// If the divisor is a power of two and the SCEV type fits in a native
// integer (and the LHS not expensive), consider the division cheap
// irrespective of whether it occurs in the user code since it can be
// lowered into a right shift.
if (auto *SC = dyn_cast<SCEVConstant>(UDivExpr->getRHS()))
if (SC->getAPInt().isPowerOf2()) {
if (isHighCostExpansionHelper(UDivExpr->getLHS(), L, At,
BudgetRemaining, TTI, Processed))
return true;
const DataLayout &DL =
L->getHeader()->getParent()->getParent()->getDataLayout();
unsigned Width = cast<IntegerType>(UDivExpr->getType())->getBitWidth();
return DL.isIllegalInteger(Width);
}
```
Since this test does not have a datalayout specified,
`SCEVExpander::isHighCostExpansionHelper()` says that
`[[TMP2:%.*]] = lshr exact i64 [[TMP1]], 5` is high-cost, and didn't perform it.
But future patches will change that logic to solely rely on cost-model,
without any such datalayout checks, so i think it is best to show
that that change is ephemeral, and can already happen without costmodel changes.
Reviewers: reames, fhahn, sanjoy, craig.topper, RKSimon
Reviewed By: RKSimon
Subscribers: javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73717
bo (splat X), (bo Y, OtherOp) --> bo (splat (bo X, Y)), OtherOp
This patch depends on the splat analysis enhancement in D73549.
See the test with comment:
; Negative test - mismatched splat elements
...as the motivation for that first patch.
The motivating case for reassociating splatted ops is shown in PR42174:
https://bugs.llvm.org/show_bug.cgi?id=42174
In that example, a slight change in order-of-associative math results
in a big difference in IR and codegen. This patch gets all of the
unnecessary shuffles out of the way, but doesn't address the potential
scalarization (see D50992 or D73480 for that).
Differential Revision: https://reviews.llvm.org/D73703
This patch adds initial support for a DemandedElts mask to the internal computeKnownBits/ComputeNumSignBits methods, matching the SelectionDAG and GlobalISel equivalents.
So far only a couple of instructions have been setup to handle the DemandedElts, the remainder still using the existing 'all elements' default. The plan is to extend support as we have test coverage.
Differential Revision: https://reviews.llvm.org/D73435
Dead instructions do not need to be sunk. Currently we try and record
the recipies for them, but there are no recipes emitted for them and
there's nothing to sink. They can be removed from SinkAfter while
marking them for recording.
Fixes PR44634.
Reviewers: rengolin, hsaito, fhahn, Ayal, gilr
Reviewed By: gilr
Differential Revision: https://reviews.llvm.org/D73423
from DenseMap to MapVector
The iteration order of LoopVectorizationLegality::Reductions matters for the
final code generation, so we better use MapVector instead of DenseMap for it
to remove the nondeterminacy. reduction-order.ll in the patch is an example
reduced from the case we saw. In the output of opt command, the order of the
select instructions in the vector.body block keeps changing from run to run
currently.
Differential Revision: https://reviews.llvm.org/D73490
The codegen for splitting a llvm.vector.reduction intrinsic into parts
will be better than the codegen for the generic reductions. This will
only directly effect when vectorization factors are specified by the
user.
Also added tests to make sure the codegen for larger reductions is OK.
Differential Revision: https://reviews.llvm.org/D72257
Currently due to the edge caching, we create wrong predicates for
branches with matching true and false successors. We will cache the
condition for the edge from the true successor, and then lookup the same
edge (src and dst are the same) for the edge to the false successor.
If both successors match, the condition should always be true. At the
moment, we cannot really create constant VPValues, but we can just
create a true condition as X | !X. Later passes will clean that up.
Fixes PR44488.
Reviewers: rengolin, hsaito, fhahn, Ayal, dorit, gilr
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D73079
This patch adds a custom implementation of isLegalNTStore to AArch64TTI
that supports vector types that can be directly stored by STNP. Note
that the implementation may not catch all valid cases (e.g. because the
vector is a multiple of 256 and could be broken down to multiple valid 256 bit
stores), but it is good enough for LV to vectorize loops with NT stores,
as LV only passes in a vector with 2 elements to check. LV seems to also
be the only user of isLegalNTStore.
We should also do the same for NT loads, but before that we need to
ensure that we properly lower LDNP of vectors, similar to D72919.
Reviewers: dmgreen, samparker, t.p.northover, ab
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D73158
Summary: Vectorized loop processes VFxUF number of elements in one iteration thus total number of iterations decreases proportionally. In addition epilog loop may not have more than VFxUF - 1 iterations. This patch updates profile information accordingly.
Reviewers: hsaito, Ayal, fhahn, reames, silvas, dcaballe, SjoerdMeijer, mkuper, DaniilSuchkov
Reviewed By: Ayal, DaniilSuchkov
Subscribers: fedor.sergeev, hiraditya, rkruppe, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67905
Summary:
This commits is a rework of the patch in
https://reviews.llvm.org/D67572.
The rework was requested to prevent out-of-tree performance regression
when vectorizing out-of-tree IR intrinsics. The vectorization of such
intrinsics is enquired via the static function `isTLIScalarize`. For
detail see the discussion in https://reviews.llvm.org/D67572.
Reviewers: uabelho, fhahn, sdesmalen
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D72734
The assume intrinsic is intentionally marked as may reading/writing
memory, to avoid passes moving them around. When flattening the CFG
for predicated blocks, we have to drop the assume calls, as they
are control-flow dependent.
There are some cases where we can do better (when control flow is
preserved), but that is follow-up work.
Fixes PR43620.
Reviewers: hsaito, rengolin, dcaballe, Ayal
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D68814
Architecturally, it's allowed to have MVE-I without an FPU, thus
-mfpu=none should not disable MVE-I, or moves to/from FP-registers.
This patch removes `+/-fpregs` from features unconditionally added to
target feature list, depending on FPU and moves the logic to Clang
driver, where the negative form (`-fpregs`) is conditionally added to
the target features list for the cases of `-mfloat-abi=soft`, or
`-mfpu=none` without either `+mve` or `+mve.fp`. Only the negative
form is added by the driver, the positive one is derived from other
features in the backend.
Differential Revision: https://reviews.llvm.org/D71843
This addresses a vectorisation regression for tail-folded loops that are
counting down, e.g. loops as simple as this:
void foo(char *A, char *B, char *C, uint32_t N) {
while (N > 0) {
*C++ = *A++ + *B++;
N--;
}
}
These are loops that can be vectorised, but when tail-folding is requested, it
can't find a primary induction variable which we do need for predicating the
loop. As a result, the loop isn't vectorised at all, which it is able to do
when tail-folding is not attempted. So, this adds a check for the primary
induction variable where we decide how to lower the scalar epilogue. I.e., when
there isn't a primary induction variable, a scalar epilogue loop is allowed
(i.e. don't request tail-folding) so that vectorisation could still be
triggered.
Having this check for the primary induction variable make sense anyway, and in
addition, in a follow-up of this I will look into discovering earlier the
primary induction variable for counting down loops, so that this can also be
tail-folded.
Differential revision: https://reviews.llvm.org/D72324
Don't overwrite existing target-cpu attributes.
I've often found the replacement behavior annoying, and this is
inconsistent with how the fast math command line flags interact with
the function attributes.
Does not yet change target-features, since I think that should behave
as a concatenation.
In https://reviews.llvm.org/D67148, we use isFloatTy to test floating
point type, otherwise we return GPRRC.
So 'double' will be classified as GPRRC, which is not accurate.
This patch covers other floating point types.
Reviewed By: #powerpc, nemanjai
Differential Revision: https://reviews.llvm.org/D71946
A sequence of additions or multiplications that is known not to wrap, may wrap
if it's order is changed (i.e., reassociated). Therefore when vectorizing
integer sum or product reductions, their no-wrap flags need to be removed.
Fixes PR43828
Patch by Denis Antrushin
Differential Revision: https://reviews.llvm.org/D69563
We somehow missed doing this when we were working on Power9 exploitation.
This just adds the missing legalization and cost for producing the vector
intrinsics.
Differential revision: https://reviews.llvm.org/D70436
With the extra optimisations we have done, these should now be fine to
enable by default. Which is what this patch does.
Differential Revision: https://reviews.llvm.org/D70968
This attempts to teach the cost model in Arm that code such as:
%s = shl i32 %a, 3
%a = and i32 %s, %b
Can under Arm or Thumb2 become:
and r0, r1, r2, lsl #3
So the cost of the shift can essentially be free. To do this without
trying to artificially adjust the cost of the "and" instruction, it
needs to get the users of the shl and check if they are a type of
instruction that the shift can be folded into. And so it needs to have
access to the actual instruction in getArithmeticInstrCost, which if
available is added as an extra parameter much like getCastInstrCost.
We otherwise limit it to shifts with a single user, which should
hopefully handle most of the cases. The list of instruction that the
shift can be folded into include ADC, ADD, AND, BIC, CMP, EOR, MVN, ORR,
ORN, RSB, SBC and SUB. This translates to Add, Sub, And, Or, Xor and
ICmp.
Differential Revision: https://reviews.llvm.org/D70966
This adds some extra cost model tests for shifts, and does some minor
adjustments to some Neon code to make it clear as to what it applies to.
Both NFC.
Alas, using half the available vector registers in a single instruction
is just too much for the register allocator to handle. The mve-vldst4.ll
test here fails when these instructions are enabled at present. This
patch disables the generation of VLD4 and VST4 by adding a
mve-max-interleave-factor option, which we currently default to 2.
Differential Revision: https://reviews.llvm.org/D71109
Currently we fail to pick the right insertion point when
PreviousLastPart of a first-order-recurrence is a PHI node not in the
LoopVectorBody. This can happen when PreviousLastPart is produce in a
predicated block. In that case, we should pick the insertion point in
the BB the PHI is in.
Fixes PR44020.
Reviewers: hsaito, fhahn, Ayal, dorit
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D71071
Fix PR40816: avoid considering scalar-with-predication instructions as also
uniform-after-vectorization.
Instructions identified as "scalar with predication" will be "vectorized" using
a replicating region. If such instructions are also optimized as "uniform after
vectorization", namely when only the first of VF lanes is used, such a
replicating region becomes erroneous - only the first instance of the region can
and should be formed. Fix such cases by not considering such instructions as
"uniform after vectorization".
Differential Revision: https://reviews.llvm.org/D70298
rL341831 moved one-use check higher up, restricting a few folds
that produced a single instruction from two instructions to the case
where the inner instruction would go away.
Original commit message:
> InstCombine: move hasOneUse check to the top of foldICmpAddConstant
>
> There were two combines not covered by the check before now,
> neither of which actually differed from normal in the benefit analysis.
>
> The most recent seems to be because it was just added at the top of the
> function (naturally). The older is from way back in 2008 (r46687)
> when we just didn't put those checks in so routinely, and has been
> diligently maintained since.
From the commit message alone, there doesn't seem to be a
deeper motivation, deeper problem that was trying to solve,
other than 'fixing the wrong one-use check'.
As i have briefly discusses in IRC with Tim, the original motivation
can no longer be recovered, too much time has passed.
However i believe that the original fold was doing the right thing,
we should be performing such a transformation even if the inner `add`
will not go away - that will still unchain the comparison from `add`,
it will no longer need to wait for `add` to compute.
Doing so doesn't seem to break any particular idioms,
as least as far as i can see.
References https://bugs.llvm.org/show_bug.cgi?id=44100
I'm not sure what the effect of this change will be on all of the affected
tests or a larger benchmark, but it fixes the horizontal add/sub problems
noted here:
https://reviews.llvm.org/D59710?vs=227972&id=228095&whitespace=ignore-most#toc
The costs are based on reciprocal throughput numbers in Agner's tables for
PEXTR*; these appear to be very slow ops on Silvermont.
This is a small step towards the larger motivation discussed in PR43605:
https://bugs.llvm.org/show_bug.cgi?id=43605
Also, it seems likely that insert/extract is the source of perf regressions on
other CPUs (up to 30%) that were cited as part of the reason to revert D59710,
so maybe we'll extend the table-based approach to other subtargets.
Differential Revision: https://reviews.llvm.org/D70607
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
Follow-up of cb47b8783: don't query TTI->preferPredicateOverEpilogue when
option -prefer-predicate-over-epilog is set to false, i.e. when we prefer not
to predicate the loop.
Differential Revision: https://reviews.llvm.org/D70382
Now that we have the intrinsics, we can add VLD2/4 and VST2/4 lowering
for MVE. This works the same way as Neon, recognising the load/shuffles
combination and converting them into intrinsics in a pre-isel pass,
which just calls getMaxSupportedInterleaveFactor, lowerInterleavedLoad
and lowerInterleavedStore.
The main difference to Neon is that we do not have a VLD3 instruction.
Otherwise most of the code works very similarly, with just some minor
differences in the form of the intrinsics to work around. VLD3 is
disabled by making isLegalInterleavedAccessType return false for those
cases.
We may need some other future adjustments, such as VLD4 take up half the
available registers so should maybe cost more. This patch should get the
basics in though.
Differential Revision: https://reviews.llvm.org/D69392
This is a follow up of d90804d, to also flag fmcp instructions as instructions
that we do not support in tail-predicated vector loops.
Differential Revision: https://reviews.llvm.org/D70295
The vectoriser queries TTI->preferPredicateOverEpilogue to determine if
tail-folding is preferred for a loop, but it was not respecting loop hint
'predicate' that can disable this, which has now been added. This showed that
we were incorrectly initialising loop hint 'vectorize.predicate.enable' with 0
(i.e. FK_Disabled) but this should have been FK_Undefined, which has been
fixed.
Differential Revision: https://reviews.llvm.org/D70125
This implements TTI hook 'preferPredicateOverEpilogue' for MVE. This is a
first version and it operates on single block loops only. With this change, the
vectoriser will now determine if tail-folding scalar remainder loops is
possible/desired, which is the first step to generate MVE tail-predicated
vector loops.
This is disabled by default for now. I.e,, this is depends on option
-disable-mve-tail-predication, which is off by default.
I will follow up on this soon with a patch for the vectoriser to respect loop
hint 'vectorize.predicate.enable'. I.e., with this loop hint set to Disabled,
we don't want to tail-fold and we shouldn't query this TTI hook, which is
done in D70125.
Differential Revision: https://reviews.llvm.org/D69845
This recommits 11ed1c0239 (reverted in
9f08ce0d21 for failing an assert) with a fix:
tryToWidenMemory() now first checks if the widening decision is to interleave,
thus maintaining previous behavior where tryToInterleaveMemory() was called
first, giving priority to interleave decisions over widening/scalarization. This
commit adds the test case that exposed this bug as a LIT.
This recommits 100e797adb (reverted in
009e032634 for failing an assert). While the
root cause was independently reverted in eaff300401,
this commit includes a LIT to make sure IVDescriptor's SinkAfter logic does not
try to sink branch instructions.
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
We have two ways to steer creating a predicated vector body over creating a
scalar epilogue. To force this, we have 1) a command line option and 2) a
pragma available. This adds a third: a target hook to TargetTransformInfo that
can be queried whether predication is preferred or not, which allows the
vectoriser to make the decision without forcing it.
While this change behaves as a non-functional change for now, it shows the
required TTI plumbing, usage of this new hook in the vectoriser, and the
beginning of an ARM MVE implementation. I will follow up on this with:
- a complete MVE implementation, see D69845.
- a patch to disable this, i.e. we should respect "vector_predicate(disable)"
and its corresponding loophint.
Differential Revision: https://reviews.llvm.org/D69040
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
Summary:
Getelementptr has vector type if any of its operands are vectors
(the scalar operands being implicitly broadcast to all vector elements).
Extractelement applied to a vector getelementptr can be folded by
applying the extractelement in turn to all of the vector operands.
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69379
Currently we may do iterleaving by more than estimated trip count
coming from the profile or computed maximum trip count. The solution is to
use "best known" trip count instead of exact one in interleaving analysis.
Patch by Evgeniy Brevnov.
Differential Revision: https://reviews.llvm.org/D67948
Add generic DAG combine for extending masked loads.
Allow us to generate sext/zext masked loads which can access v4i8,
v8i8 and v4i16 memory to produce v4i32, v8i16 and v4i32 respectively.
Differential Revision: https://reviews.llvm.org/D68337
llvm-svn: 375085
In loop-vectorize, interleave count and vector factor depend on target register number. Currently, it does not
estimate different register pressure for different register class separately(especially for scalar type,
float type should not be on the same position with int type), so it's not accurate. Specifically,
it causes too many times interleaving/unrolling, result in too many register spills in loop body and hurting performance.
So we need classify the register classes in IR level, and importantly these are abstract register classes,
and are not the target register class of backend provided in td file. It's used to establish the mapping between
the types of IR values and the number of simultaneous live ranges to which we'd like to limit for some set of those types.
For example, POWER target, register num is special when VSX is enabled. When VSX is enabled, the number of int scalar register is 32(GPR),
float is 64(VSR), but for int and float vector register both are 64(VSR). So there should be 2 kinds of register class when vsx is enabled,
and 3 kinds of register class when VSX is NOT enabled.
It runs on POWER target, it makes big(+~30%) performance improvement in one specific bmk(503.bwaves_r) of spec2017 and no other obvious degressions.
Differential revision: https://reviews.llvm.org/D67148
llvm-svn: 374634
When optimising for size and SCEV runtime checks need to be emitted to check
overflow behaviour, the loop vectorizer can run in this assert:
LoopVectorize.cpp:2699: void llvm::InnerLoopVectorizer::emitSCEVChecks(
llvm::Loop *, llvm::BasicBlock *): Assertion `!BB->getParent()->hasOptSize()
&& "Cannot SCEV check stride or overflow when opt
We should not generate predicates while optimising for size because
code will be generated for predicates such as these SCEV overflow runtime
checks.
This should fix PR43371.
Differential Revision: https://reviews.llvm.org/D68082
llvm-svn: 374166
Also Revert "[LoopVectorize] Fix non-debug builds after rL374017"
This reverts commit 9f41deccc0.
This reverts commit 18b6fe07bc.
The patch is breaking PowerPC internal build, checked with author, reverting
on behalf of him for now due to timezone.
llvm-svn: 374091
Simon Pilgrim