| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
| |
Reviewers: sanjoy, anna, reames, apilipenko, igor-laevsky, mkuper
Subscribers: jholewinski, arsenm, mzolotukhin, nemanjai, nhaehnle, javed.absar, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D34531
llvm-svn: 306554
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
Similar to X86, it should be safe to inline callees if their target-features
are a subset of the caller. This change matches GCC's inlining behavior
with respect to attributes [1].
[1] https://gcc.gnu.org/onlinedocs/gcc/AArch64-Function-Attributes.html#AArch64-Function-Attributes
Reviewers: kristof.beyls, javed.absar, rengolin, t.p.northover
Reviewed By: t.p.northover
Subscribers: aemerson, eraman, llvm-commits
Differential Revision: https://reviews.llvm.org/D34698
llvm-svn: 306478
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary: The method TargetTransformInfo::getRegisterBitWidth() is declared const, but the type erasing implementation classes (TargetTransformInfo::Concept & TargetTransformInfo::Model) that were introduced by Chandler in https://reviews.llvm.org/D7293 do not have the method declared const. This is an NFC to tidy up the const consistency between TTI and its implementation.
Reviewers: chandlerc, rnk, reames
Reviewed By: reames
Subscribers: reames, jfb, arsenm, dschuff, nemanjai, nhaehnle, javed.absar, sbc100, jgravelle-google, llvm-commits
Differential Revision: https://reviews.llvm.org/D33903
llvm-svn: 305189
|
| |
|
|
|
|
|
| |
The issue was that the AArch64 TTI hook allowed unpacked integer cmp reductions
which didn't have a lowering.
llvm-svn: 303211
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
ARM Neon has native support for half-sized vector registers (64 bits). This
is beneficial for example for 2D and 3D graphics. This patch adds the option
to lower MinVecRegSize from 128 via a TTI in the SLP Vectorizer.
*** Performance Analysis
This change was motivated by some internal benchmarks but it is also
beneficial on SPEC and the LLVM testsuite.
The results are with -O3 and PGO. A negative percentage is an improvement.
The testsuite was run with a sample size of 4.
** SPEC
* CFP2006/482.sphinx3 -3.34%
A pretty hot loop is SLP vectorized resulting in nice instruction reduction.
This used to be a +22% regression before rL299482.
* CFP2000/177.mesa -3.34%
* CINT2000/256.bzip2 +6.97%
My current plan is to extend the fix in rL299482 to i16 which brings the
regression down to +2.5%. There are also other problems with the codegen in
this loop so there is further room for improvement.
** LLVM testsuite
* SingleSource/Benchmarks/Misc/ReedSolomon -10.75%
There are multiple small SLP vectorizations outside the hot code. It's a bit
surprising that it adds up to 10%. Some of this may be code-layout noise.
* MultiSource/Benchmarks/VersaBench/beamformer/beamformer -8.40%
The opt-viewer screenshot can be seen at F3218284. We start at a colder store
but the tree leads us into the hottest loop.
* MultiSource/Applications/lambda-0.1.3/lambda -2.68%
* MultiSource/Benchmarks/Bullet/bullet -2.18%
This is using 3D vectors.
* SingleSource/Benchmarks/Shootout-C++/Shootout-C++-lists +6.67%
Noise, binary is unchanged.
* MultiSource/Benchmarks/Ptrdist/anagram/anagram +4.90%
There is an additional SLP in the cold code. The test runs for ~1sec and
prints out over 2000 lines. This is most likely noise.
* MultiSource/Applications/aha/aha +1.63%
* MultiSource/Applications/JM/lencod/lencod +1.41%
* SingleSource/Benchmarks/Misc/richards_benchmark +1.15%
Differential Revision: https://reviews.llvm.org/D31965
llvm-svn: 303116
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This caused PR33053.
Original commit message:
> The new experimental reduction intrinsics can now be used, so I'm enabling this
> for AArch64. We will need this for SVE anyway, so it makes sense to do this for
> NEON reductions as well.
>
> The existing code to match shufflevector patterns are replaced with a direct
> lowering of the reductions to AArch64-specific nodes. Tests updated with the
> new, simpler, representation.
>
> Differential Revision: https://reviews.llvm.org/D32247
llvm-svn: 303115
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The new experimental reduction intrinsics can now be used, so I'm enabling this
for AArch64. We will need this for SVE anyway, so it makes sense to do this for
NEON reductions as well.
The existing code to match shufflevector patterns are replaced with a direct
lowering of the reductions to AArch64-specific nodes. Tests updated with the
new, simpler, representation.
Differential Revision: https://reviews.llvm.org/D32247
llvm-svn: 302678
|
| |
|
|
|
|
|
|
|
| |
This pass uses a new target hook to decide whether or not to expand a particular
intrinsic to the shuffevector sequence.
Differential Revision: https://reviews.llvm.org/D32245
llvm-svn: 302631
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The AArch64 instruction set has a few "widening" instructions (e.g., uaddl,
saddl, uaddw, etc.) that take one or more doubleword operands and produce
quadword results. The operands are automatically sign- or zero-extended as
appropriate. However, in LLVM IR, these extends are explicit. This patch
updates TTI to consider these widening instructions as single operations whose
cost is attached to the arithmetic instruction. It marks extends that are part
of a widening operation "free" and applies a sub-target specified overhead
(zero by default) to the arithmetic instructions.
Differential Revision: https://reviews.llvm.org/D32706
llvm-svn: 302582
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
getArithmeticInstrCost(), getShuffleCost(), getCastInstrCost(),
getCmpSelInstrCost(), getVectorInstrCost(), getMemoryOpCost(),
getInterleavedMemoryOpCost() implemented.
Interleaved access vectorization enabled.
BasicTTIImpl::getCastInstrCost() improved to check for legal extending loads,
in which case the cost of the z/sext instruction becomes 0.
Review: Ulrich Weigand, Renato Golin.
https://reviews.llvm.org/D29631
llvm-svn: 300052
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
Move the aarch64-type-promotion pass within the existing type promotion framework in CGP.
This change also support forking sexts when a new sext is required for promotion.
Note that change is based on D27853 and I am submitting this out early to provide a better idea on D27853.
Reviewers: jmolloy, mcrosier, javed.absar, qcolombet
Reviewed By: qcolombet
Subscribers: llvm-commits, aemerson, rengolin, mcrosier
Differential Revision: https://reviews.llvm.org/D28680
llvm-svn: 299379
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Refactoring to remove duplications of this method.
New method getOperandsScalarizationOverhead() that looks at the present unique
operands and add extract costs for them. Old behaviour was to just add extract
costs for one operand of the type always, which still happens in
getArithmeticInstrCost() if no operands are provided by the caller.
This is a good start of improving on this, but there are more places
that can be improved by using getOperandsScalarizationOverhead().
Review: Hal Finkel
https://reviews.llvm.org/D29017
llvm-svn: 293155
|
| |
|
|
|
|
|
|
|
|
|
|
| |
updated instructions:
pmulld, pmullw, pmulhw, mulsd, mulps, mulpd, divss, divps, divsd, divpd, addpd and subpd.
special optimization case which replaces pmulld with pmullw\pmulhw\pshuf seq.
In case if the real operands bitwidth <= 16.
Differential Revision: https://reviews.llvm.org/D28104
llvm-svn: 291657
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
stride seems to be 'complex' and need some extra cost for address computation handling.
This code seems to be target dependent which may not be the same for all targets.
Passed the decision whether the given stride is complex or not to the target by sending stride information via SCEV to getAddressComputationCost instead of 'IsComplex'.
Specifically at X86 targets we dont see any significant address computation cost in case of the strided access in general.
Differential Revision: https://reviews.llvm.org/D27518
llvm-svn: 291106
|
| |
|
|
|
|
|
|
|
|
| |
All of these existed because MSVC 2013 was unable to synthesize default
move ctors. We recently dropped support for it so all that error-prone
boilerplate can go.
No functionality change intended.
llvm-svn: 284721
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This change adds a new hook for estimating the cost of vector extracts followed
by zero- and sign-extensions. The motivating example for this change is the
SMOV and UMOV instructions on AArch64. These instructions move data from vector
to general purpose registers while performing the corresponding extension
(sign-extend for SMOV and zero-extend for UMOV) at the same time. For these
operations, TargetTransformInfo can assume the extensions are free and only
report the cost of the vector extract. The SLP vectorizer has been updated to
make use of the new hook.
Differential Revision: http://reviews.llvm.org/D18523
llvm-svn: 267725
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
It can hurt performance to prefetch ahead too much. Be conservative for
now and don't prefetch ahead more than 3 iterations on Cyclone.
Reviewers: hfinkel
Subscribers: llvm-commits, mzolotukhin
Differential Revision: http://reviews.llvm.org/D17949
llvm-svn: 263772
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
And use this TTI for Cyclone. As it was explained in the original RFC
(http://thread.gmane.org/gmane.comp.compilers.llvm.devel/92758), the HW
prefetcher work up to 2KB strides.
I am also adding tests for this and the previous change (D17943):
* Cyclone prefetching accesses with a large stride
* Cyclone not prefetching accesses with a small stride
* Generic Aarch64 subtarget not prefetching either
Reviewers: hfinkel
Subscribers: aemerson, rengolin, llvm-commits, mzolotukhin
Differential Revision: http://reviews.llvm.org/D17945
llvm-svn: 263771
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
This wires up the pass for Cyclone but keeps it off for now because we
need a few more TTIs.
The getPrefetchMinStride value is not very well tuned right now but it
works well with CFP2006/433.milc which motivated this.
Tests will be added as part of the upcoming large-stride prefetching
patch.
Reviewers: t.p.northover
Subscribers: llvm-commits, aemerson, hfinkel, rengolin
Differential Revision: http://reviews.llvm.org/D17943
llvm-svn: 263770
|
| |
|
|
|
|
| |
propagate to all callers/users/etc.
llvm-svn: 247864
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
This change turns on by default interleaved access vectorization
for AArch64.
We also clean up some tests which were spedifically enabling this
behaviour.
Reviewers: rengolin
Subscribers: aemerson, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D12149
llvm-svn: 246542
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
rather than 'unsigned' for their costs.
For something like costs in particular there is a natural "negative"
value, that of savings or saved cost. As a consequence, there is a lot
of code that subtracts or creates negative values based on cost, all of
which is prone to awkwardness or bugs when dealing with an unsigned
type. Similarly, we *never* want these values to wrap, as that would
cause Very Bad code generation (likely percieved as an infinite loop as
we try to emit over 2^32 instructions or some such insanity).
All around 'int' seems a much better fit for these basic metrics. I've
added asserts to ensure that at least the TTI interface never returns
negative numbers here. If we ever have a use case for negative numbers,
we can remove this, but this way a bug where someone used '-1' to
produce a 'very large' cost will be caught by the assert.
This passes all tests, and is also UBSan clean.
No functional change intended.
Differential Revision: http://reviews.llvm.org/D11741
llvm-svn: 244080
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
DataLayout is no longer optional. It was initialized with or without
a DataLayout, and the DataLayout when supplied could have been the
one from the TargetMachine.
Summary:
This change is part of a series of commits dedicated to have a single
DataLayout during compilation by using always the one owned by the
module.
Reviewers: echristo
Subscribers: jholewinski, llvm-commits, rafael, yaron.keren
Differential Revision: http://reviews.llvm.org/D11021
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 241774
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
patch also adds a function to calculate the cost of interleaved memory accesses.
E.g. Lower an interleaved load:
%wide.vec = load <8 x i32>, <8 x i32>* %ptr
%v0 = shuffle %wide.vec, undef, <0, 2, 4, 6>
%v1 = shuffle %wide.vec, undef, <1, 3, 5, 7>
into:
%ld2 = { <4 x i32>, <4 x i32> } call llvm.aarch64.neon.ld2(%ptr)
%vec0 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 0
%vec1 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 1
E.g. Lower an interleaved store:
%i.vec = shuffle <8 x i32> %v0, <8 x i32> %v1, <0, 4, 8, 1, 5, 9, 2, 6, 10, 3, 7, 11>
store <12 x i32> %i.vec, <12 x i32>* %ptr
into:
%sub.v0 = shuffle <8 x i32> %v0, <8 x i32> v1, <0, 1, 2, 3>
%sub.v1 = shuffle <8 x i32> %v0, <8 x i32> v1, <4, 5, 6, 7>
%sub.v2 = shuffle <8 x i32> %v0, <8 x i32> v1, <8, 9, 10, 11>
call void llvm.aarch64.neon.st3(%sub.v0, %sub.v1, %sub.v2, %ptr)
Differential Revision: http://reviews.llvm.org/D10533
llvm-svn: 240754
|
| |
|
|
|
|
| |
AArch64 and ARM backend.
llvm-svn: 239713
|
| |
|
|
|
|
|
| |
Re-commit after adding "-aarch64-neon-syntax=generic" to fix the failure on OS X.
This patch was firstly committed in r239514, then reverted in r239544 because of a syntax incompatible failure on OS X.
llvm-svn: 239711
|
| |
|
|
|
|
|
|
|
| |
Revert "[AArch64] Match interleaved memory accesses into ldN/stN instructions."
Revert "Fixing MSVC 2013 build error."
The test/CodeGen/AArch64/aarch64-interleaved-accesses.ll test was failing on OS X.
llvm-svn: 239544
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a pass AArch64InterleavedAccess to identify and match interleaved memory accesses. This pass transforms an interleaved load/store into ldN/stN intrinsic. As Loop Vectorizor disables optimization on interleaved accesses by default, this optimization is also disabled by default. To enable it by "-aarch64-interleaved-access-opt=true"
E.g. Transform an interleaved load (Factor = 2):
%wide.vec = load <8 x i32>, <8 x i32>* %ptr
%v0 = shuffle %wide.vec, undef, <0, 2, 4, 6> ; Extract even elements
%v1 = shuffle %wide.vec, undef, <1, 3, 5, 7> ; Extract odd elements
Into:
%ld2 = { <4 x i32>, <4 x i32> } call aarch64.neon.ld2(%ptr)
%v0 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 0
%v1 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 1
E.g. Transform an interleaved store (Factor = 2):
%i.vec = shuffle %v0, %v1, <0, 4, 1, 5, 2, 6, 3, 7> ; Interleaved vec
store <8 x i32> %i.vec, <8 x i32>* %ptr
Into:
%v0 = shuffle %i.vec, undef, <0, 1, 2, 3>
%v1 = shuffle %i.vec, undef, <4, 5, 6, 7>
call void aarch64.neon.st2(%v0, %v1, %ptr)
llvm-svn: 239514
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The patch disabled unrolling in loop vectorization pass when VF==1 on x86 architecture,
by setting MaxInterleaveFactor to 1. Unrolling in loop vectorization pass may introduce
the cost of overflow check, memory boundary check and extra prologue/epilogue code when
regular unroller will unroll the loop another time. Disable it when VF==1 remove the
unnecessary cost on x86. The same can be done for other platforms after verifying
interleaving/memory bound checking to be not perf critical on those platforms.
Differential Revision: http://reviews.llvm.org/D9515
llvm-svn: 236613
|
| |
|
|
|
|
| |
preferences interface on TTI now that all of TTI is per-function.
llvm-svn: 227741
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
now that we have a correct and cached subtarget specific to the
function.
Also, finish providing a cached per-function subtarget in the core
LLVMTargetMachine -- that layer hadn't switched over yet.
The only use of the TargetMachine was to re-lookup a subtarget for
a particular function to work around the fact that TTI was immutable.
Now that it is per-function and we haved a cached subtarget, use it.
This still leaves a few interfaces with real warts on them where we were
passing Function objects through the TTI interface. I'll remove these
and clean their usage up in subsequent commits now that this isn't
necessary.
llvm-svn: 227738
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
intermediate TTI implementation template and instead query up to the
derived class for both the TargetMachine and the TargetLowering.
Most of the derived types had a TLI cached already and there is no need
to store a less precisely typed target machine pointer.
This will in turn make it much cleaner to look up the TLI via
a per-function subtarget instead of the generic subtarget, and it will
pave the way toward pulling the subtarget used for unroll preferences
into the same form once we are *always* using the function to look up
the correct subtarget.
llvm-svn: 227737
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
TargetIRAnalysis access path directly rather than implementing getTTI.
This even removes getTTI from the interface. It's more efficient for
each target to just register a precise callback that creates their
specific TTI.
As part of this, all of the targets which are building their subtargets
individually per-function now build their TTI instance with the function
and thus look up the correct subtarget and cache it. NVPTX, R600, and
XCore currently don't leverage this functionality, but its trivial for
them to add it now.
llvm-svn: 227735
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
null.
For some reason some of the original TTI code supported a null target
machine. This seems to have been legacy, and I made matters worse when
refactoring this code by spreading that pattern further through the
various targets.
The TargetMachine can't actually be null, and it doesn't make sense to
support that use case. I've now consistently removed it and removed all
of the code trying to cope with that situation. This is probably good,
as several targets *didn't* cope with it being null despite the null
default argument in their constructors. =]
llvm-svn: 227734
|
|
|
base which it adds a single analysis pass to, to instead return the type
erased TargetTransformInfo object constructed for that TargetMachine.
This removes all of the pass variants for TTI. There is now a single TTI
*pass* in the Analysis layer. All of the Analysis <-> Target
communication is through the TTI's type erased interface itself. While
the diff is large here, it is nothing more that code motion to make
types available in a header file for use in a different source file
within each target.
I've tried to keep all the doxygen comments and file boilerplate in line
with this move, but let me know if I missed anything.
With this in place, the next step to making TTI work with the new pass
manager is to introduce a really simple new-style analysis that produces
a TTI object via a callback into this routine on the target machine.
Once we have that, we'll have the building blocks necessary to accept
a function argument as well.
llvm-svn: 227685
|
