Account for subvector extraction/insertion, helps prevent the vectorizers from selecting 256-bit vectors that will have to be split anyhow on AVX1 targets.
llvm-svn: 302378
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:
LSV wants to know the maximum size that can be loaded to a vector register.
On X86, this always matches the maximum register width. Implement this
accordingly and add a test to make sure that LSV can vectorize up to the
maximum permissible width on X86.
Reviewers: delena, arsenm
Reviewed By: arsenm
Subscribers: wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D31504
llvm-svn: 299589
getIntrinsicInstrCost() used to only compute scalarization cost based on types.
This patch improves this so that the actual arguments are checked when they are
available, in order to handle only unique non-constant operands.
Tests updates:
Analysis/CostModel/X86/arith-fp.ll
Transforms/LoopVectorize/AArch64/interleaved_cost.ll
Transforms/LoopVectorize/ARM/interleaved_cost.ll
The improvement in getOperandsScalarizationOverhead() to differentiate on
constants made it necessary to update the interleaved_cost.ll tests even
though they do not relate to intrinsics.
Review: Hal Finkel
https://reviews.llvm.org/D29540
llvm-svn: 297705
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
The 'fast' costs should only work for shifts by uniform constants (uniform non-constant are lowered using the slow default implementation).
Logical shifts were not taking into account that we must mask the psrlw result, so the costs needed to be doubled.
Added missing AVX2/AVX512BW costs as well.
llvm-svn: 291391
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
X86 target does not provide any target specific cost calculation for interleave patterns.It uses the common target-independent calculation, which gives very high numbers. As a result, the scalar version is chosen in many cases. The situation on AVX-512 is even worse, since we have 3-src shuffles that significantly reduce the cost.
In this patch I calculate the cost on AVX-512. It will allow to compare interleave pattern with gather/scatter and choose a better solution (PR31426).
* Shiffle-broadcast cost will be changed in Simon's upcoming patch.
Differential Revision: https://reviews.llvm.org/D28118
llvm-svn: 290810
Add explicit v16i16/v32i8 ADD/SUB costs, matching the costs of v4i64/v8i32 - they were missing for some reason.
This has side effects on the LV max bandwidth tests (AVX1 now prefers 128-bit vectors vs AVX2 which still prefers 256-bit)
llvm-svn: 286832
This patch avoids scalarization of CTLZ by instead expanding to use CTPOP (ref: "Hacker's Delight") when the necessary operations are available.
This also adds the necessary cost models for X86 SSE2 targets (the main beneficiary) to ensure vectorization only happens when its useful.
Differential Revision: https://reviews.llvm.org/D25910
llvm-svn: 286233
There is a bug describing poor cost model for floating point operations:
Bug 29083 - [X86][SSE] Improve costs for floating point operations. This
patch is the second one in series of patches dealing with cost model.
Differential Revision: https://reviews.llvm.org/D25722
llvm-svn: 285564
With DQI but without VLX, lower v2i64 and v4i64 MUL operations with v8i64 MUL (vpmullq).
Updated cost table accordingly.
Differential Revision: https://reviews.llvm.org/D26011
llvm-svn: 285304
We were defaulting to SSE2 costs which weren't taking into account the availability of PBLENDW/PBLENDVB to improve merging of per-element shift results.
llvm-svn: 284939
We weren't accounting for legal types on every subtarget, meaning that many of the costs were using defaults.
We still don't correctly cost (or test) the 512-bit sdiv/udiv by uniform const cases, nor the power-of-2 cases.
llvm-svn: 284744
As discussed on PR28461 we currently miss the chance to lower "fptosi <2 x double> %arg to <2 x i32>" to cvttpd2dq due to its use of illegal types.
This patch adds support for fptosi to 2i32 from both 2f64 and 2f32.
It also recognises that cvttpd2dq zeroes the upper 64-bits of the xmm result (similar to D23797) - we still don't do this for the cvttpd2dq/cvttps2dq intrinsics - this can be done in a future patch.
Differential Revision: https://reviews.llvm.org/D23808
llvm-svn: 284459
The current Cost Model implementation is very inaccurate and has to be
updated, improved, re-implemented to be able to take into account the
concrete CPU models and the concrete targets where this Cost Model is
being used. For example, the Latency Cost Model should be differ from
Code Size Cost Model, etc.
This patch is the first step to launch the developing and implementation
of a new Cost Model generation.
Differential Revision: https://reviews.llvm.org/D25186
llvm-svn: 284012
This is a mechanical change of comments in switches like fallthrough,
fall-through, or fall-thru to use the LLVM_FALLTHROUGH macro instead.
llvm-svn: 278902
This reverts commit r278048. Something changed between the last time I
built this--it takes awhile on my ridiculously slow and ancient
computer--and now that broke this.
llvm-svn: 278053
Summary:
Based on two patches by Michael Mueller.
This is a target attribute that causes a function marked with it to be
emitted as "hotpatchable". This particular mechanism was originally
devised by Microsoft for patching their binaries (which they are
constantly updating to stay ahead of crackers, script kiddies, and other
ne'er-do-wells on the Internet), but is now commonly abused by Windows
programs to hook API functions.
This mechanism is target-specific. For x86, a two-byte no-op instruction
is emitted at the function's entry point; the entry point must be
immediately preceded by 64 (32-bit) or 128 (64-bit) bytes of padding.
This padding is where the patch code is written. The two byte no-op is
then overwritten with a short jump into this code. The no-op is usually
a `movl %edi, %edi` instruction; this is used as a magic value
indicating that this is a hotpatchable function.
Reviewers: majnemer, sanjoy, rnk
Subscribers: dberris, llvm-commits
Differential Revision: https://reviews.llvm.org/D19908
llvm-svn: 278048
Shifts with a uniform but non-constant count were considered very expensive to
vectorize, because the splat of the uniform count and the shift would tend to
appear in different blocks. That made the splat invisible to ISel, and we'd
scalarize the shift at codegen time.
Since r201655, CodeGenPrepare sinks those splats to be next to their use, and we
are able to select the appropriate vector shifts. This updates the cost model to
to take this into account by making shifts by a uniform cheap again.
Differential Revision: https://reviews.llvm.org/D23049
llvm-svn: 277782
This patch adds costs for the vectorized implementations of CTPOP, the default values were seriously underestimating the cost of these and was encouraging vectorization on targets where serialized use of POPCNT would be much better.
Differential Revision: https://reviews.llvm.org/D22456
llvm-svn: 276104
Make some AVX and AVX512 cast costs more precise.
Based on part of a patch by Elena Demikhovsky (D15604).
Differential Revision: http://reviews.llvm.org/D22064
llvm-svn: 275106
This is "cvtdq2ps" which does not appear to be particularly slow on any CPU
according to Agner's tables. Choosing "5" as a cost here as suggested in:
https://llvm.org/bugs/show_bug.cgi?id=21356
...but it seems very conservative given that the instruction is fully pipelined,
and I think these costs are supposed to model throughput.
Note that related costs are also most likely too high, but this fixes PR21356
and partly fixes PR28434.
llvm-svn: 274658
Cast cost tables are now sorted, for each cast type, lexicographically on
[source base type, source vector width, dest base type, base vector width].
llvm-svn: 274653
The BSWAP of vector types is quite efficiently implemented using vector shuffles on SSE/AVX targets, we should reflect the typical cost of this to encourage vectorization.
Differential Revision: http://reviews.llvm.org/D21521
llvm-svn: 273217
The costs are somewhat hand-wavy, but should be much closer to the truth
than what we get from BasicTTI.
Differential Revision: http://reviews.llvm.org/D21156
llvm-svn: 272406
By making pointer extraction from a vector more expensive in the cost model,
we avoid the vectorization of a loop that is very likely to be memory-bound:
https://llvm.org/bugs/show_bug.cgi?id=27826
There are still bugs related to this, so we may need a more general solution
to avoid vectorizing obviously memory-bound loops when we don't have HW gather
support.
Differential Revision: http://reviews.llvm.org/D20601
llvm-svn: 270729
As discussed on PR24888, until SSE42 we don't have access to PCMPGTQ for v2i64 comparisons, but the cost models don't reflect this, resulting in over-optimistic vectorizaton.
This patch adds SSE2 'base level' costs that match what a typical target is capable of and only reduces the v2i64 costs at SSE42.
Technically SSE41 provides a PCMPEQQ v2i64 equality test, but as getCmpSelInstrCost doesn't give us a way to discriminate between comparison test types we can't easily make use of this, otherwise we could split the cost of integer equality and greater-than tests to give better costings of each.
Differential Revision: http://reviews.llvm.org/D20057
llvm-svn: 268972
Summary:
rL256194 transforms truncations between vectors of integers into PACKUS/PACKSS
operations during DAG combine. This generates better code for truncate, so cost
of truncate needs to be changed but looks like it got changed only in SSE2 table
Whereas this change is also applicable for SSE4.1, so the cost of truncate needs
to be changed for that as well. Cost of “TRUNCATE v16i32 to v16i8” & “TRUNCATE
v16i16 to v16i8” should be same in SSE4.1 & SSE2 table. Removing their cost from
SSE4.1, so it will fall back to SSE2.
Reviewers: Simon Pilgrim
llvm-svn: 267123
This code was creating a new type in the global context, regardless
of which context the user is sitting in, what can possibly go wrong?
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 266275
The irony of this patch is that one CPU that is affected is AMD Jaguar, and Jaguar
has a completely double-pumped AVX implementation. But getting the cost model to
reflect that is a much bigger problem. The small goal here is simply to improve on
the lie that !AVX2 == SandyBridge.
Differential Revision: http://reviews.llvm.org/D18000
llvm-svn: 263069
Use AVX1 FP instructions (vmaskmovps/pd) in place of the AVX2 int instructions (vpmaskmovd/q).
Differential Revision: http://reviews.llvm.org/D16528
llvm-svn: 258675
The cost is calculated for all X86 targets. When gather/scatter instruction
is not supported we calculate the cost of scalar sequence.
Differential revision: http://reviews.llvm.org/D15677
llvm-svn: 256519
This patch transforms truncation between vectors of integers into
X86ISD::PACKUS/PACKSS operations during DAG combine. We don't do it in
lowering phase because after type legalization, the original truncation
will be turned into a BUILD_VECTOR with each element that is extracted
from a vector and then truncated, and from them it is difficult to do
this optimization. This greatly improves the performance of truncations
on some specific types.
Cost table is updated accordingly.
Differential revision: http://reviews.llvm.org/D14588
llvm-svn: 256194
This allows "icmp ugt %a, 4294967295" and "icmp uge %a, 4294967296" to be optimized into right shifts by 32 which can fold the immediate into the shift instruction. These patterns show up with some regularity in real code.
Unfortunately, since getImmCost can't see the icmp predicate we can't be tell if we're only catching these specific cases.
llvm-svn: 256126
Previously in the conversion cost table there are no entries for integer-integer
conversions on SSE2. This will result in imprecise costs for certain vectorized
operations. This patch adds those entries for SSE2 and SSE4.1. The cost numbers
are counted from the result of running llc on the new test case in this patch.
Differential revision: http://reviews.llvm.org/D15132
llvm-svn: 255315
I checked and updated the cost of AVX-512 conversion operations. Added cost of conversion operations in DQ mode.
Conversion of illegal types that requires vector split is not calculated right now (like for other X86 targets).
Differential Revision: http://reviews.llvm.org/D15074
llvm-svn: 254494
The masked intrinsics support all integer and floating point data types. I added the pointer type to this list.
Added tests for CodeGen and for Loop Vectorizer.
Updated the Language Reference.
Differential Revision: http://reviews.llvm.org/D14150
llvm-svn: 253544
This also lets us remove the versions of the functions that took a statically sized array as we can rely on ArrayRef implicit conversion now.
llvm-svn: 251490
This avoid mentioning the table name an extra time and allows the lookup to be done directly in the ifs by relying on the bool conversion of the pointer.
While there make use of ArrayRef and std::find_if.
llvm-svn: 251382
When the target does not support these intrinsics they should be converted to a chain of scalar load or store operations.
If the mask is not constant, the scalarizer will build a chain of conditional basic blocks.
I added isLegalMaskedGather() isLegalMaskedScatter() APIs.
Differential Revision: http://reviews.llvm.org/D13722
llvm-svn: 251237
Clang runtime failure was reported.
Assertion failed: (isExtended() && "Type is not extended!"), function getTypeForEVT
I'll need to add a proper handling for PointerType in masked load/store intrinsics.
llvm-svn: 250995
Originally I planned to use the same interface for masked gather/scatter and set isConsecutive to "false" in this case.
Now I'm implementing masked gather/scatter and see that the interface is inconvenient. I want to add interfaces isLegalMaskedGather() / isLegalMaskedScatter() instead of using the "Consecutive" parameter in the existing interfaces.
Differential Revision: http://reviews.llvm.org/D13850
llvm-svn: 250686
Most importantly, this keeps constant hoisting from preventing instruction selections ability to turn an AND with 0xffffffff into a move into a 32-bit subregister.
llvm-svn: 249370
The XOP shifts just have logical/arithmetic versions and the left/right shifts are controlled by whether the value is positive/negative. Because of this I've added new X86ISD nodes instead of trying to force them to use the existing shift nodes.
Additionally Excavator cores (bdver4) support XOP and AVX2 - meaning that it should use the AVX2 shifts when it can and fall back to XOP in other cases.
Differential Revision: http://reviews.llvm.org/D8690
llvm-svn: 248878
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
This patch vectorizes the v2i64/v4i64 ASHR shift operations - the last remaining integer vector shifts that are still being transferred to/from the scalar unit to be completed.
Differential Revision: http://reviews.llvm.org/D11439
llvm-svn: 243569
Reordered the data tables at the top and placed the lookups after. The first stage in the yak shaving necessary to get more accurate costs for a variety of targets given the recent improvements to SINT_TO_FP/UINT_TO_FP/SIGN_EXTEND vector lowering.
llvm-svn: 242643
While the v4i32 shl operation is already vectorized using a cvttps2dq/pmulld pattern, the lshr/ashr opeations are still scalarized.
This patch adds vectorization support for non-uniform v4i32 shift operations - it splats constant shift amounts to allow them to use the immediate sse shift instructions, or extracts/zero-extends non-constant shift amounts. The individual results are then blended together.
Differential Revision: http://reviews.llvm.org/D11063
llvm-svn: 241989
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, ted, yaron.keren, rafael, llvm-commits
Differential Revision: http://reviews.llvm.org/D11028
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 241775
This patch adds vectorization support for uniform constant i64 arithmetic shift right operators.
Differential Revision: http://reviews.llvm.org/D9645
llvm-svn: 241514
This checks subtarget feature compatibility for inlining by verifying
that the callee is a strict subset of the caller's features. This includes
the cpu as part of the subtarget we can get via the incoming functions as
the backend takes CPUs as feature sets.
This allows us to inline things like:
int foo() { return baz(); }
int __attribute__((target("sse4.2"))) bar() {
return foo();
}
so that generic code can be inlined into specialized functions.
llvm-svn: 241221
This patch ensures that SHL/SRL/SRA shifts for i8 and i16 vectors avoid scalarization. It builds on the existing i8 SHL vectorized implementation of moving the shift bits up to the sign bit position and separating the 4, 2 & 1 bit shifts with several improvements:
1 - SSE41 targets can use (v)pblendvb directly with the sign bit instead of performing a comparison to feed into a VSELECT node.
2 - pre-SSE41 targets were masking + comparing with an 0x80 constant - we avoid this by using the fact that a set sign bit means a negative integer which can be compared against zero to then feed into VSELECT, avoiding the need for a constant mask (zero generation is much cheaper).
3 - SRA i8 needs to be unpacked to the upper byte of a i16 so that the i16 psraw instruction can be correctly used for sign extension - we have to do more work than for SHL/SRL but perf tests indicate that this is still beneficial.
The i16 implementation is similar but simpler than for i8 - we have to do 8, 4, 2 & 1 bit shifts but less shift masking is involved. SSE41 use of (v)pblendvb requires that the i16 shift amount is splatted to both bytes however.
Tested on SSE2, SSE41 and AVX machines.
Differential Revision: http://reviews.llvm.org/D9474
llvm-svn: 239509
Part of D9474, this patch extends AVX2 v16i16 types to 2 x 8i32 vectors and uses i32 shift variable shifts before packing back to i16.
Adds AVX2 tests for v8i16 and v16i16
llvm-svn: 238149
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
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
type erased interface and a single analysis pass rather than an
extremely complex analysis group.
The end result is that the TTI analysis can contain a type erased
implementation that supports the polymorphic TTI interface. We can build
one from a target-specific implementation or from a dummy one in the IR.
I've also factored all of the code into "mix-in"-able base classes,
including CRTP base classes to facilitate calling back up to the most
specialized form when delegating horizontally across the surface. These
aren't as clean as I would like and I'm planning to work on cleaning
some of this up, but I wanted to start by putting into the right form.
There are a number of reasons for this change, and this particular
design. The first and foremost reason is that an analysis group is
complete overkill, and the chaining delegation strategy was so opaque,
confusing, and high overhead that TTI was suffering greatly for it.
Several of the TTI functions had failed to be implemented in all places
because of the chaining-based delegation making there be no checking of
this. A few other functions were implemented with incorrect delegation.
The message to me was very clear working on this -- the delegation and
analysis group structure was too confusing to be useful here.
The other reason of course is that this is *much* more natural fit for
the new pass manager. This will lay the ground work for a type-erased
per-function info object that can look up the correct subtarget and even
cache it.
Yet another benefit is that this will significantly simplify the
interaction of the pass managers and the TargetMachine. See the future
work below.
The downside of this change is that it is very, very verbose. I'm going
to work to improve that, but it is somewhat an implementation necessity
in C++ to do type erasure. =/ I discussed this design really extensively
with Eric and Hal prior to going down this path, and afterward showed
them the result. No one was really thrilled with it, but there doesn't
seem to be a substantially better alternative. Using a base class and
virtual method dispatch would make the code much shorter, but as
discussed in the update to the programmer's manual and elsewhere,
a polymorphic interface feels like the more principled approach even if
this is perhaps the least compelling example of it. ;]
Ultimately, there is still a lot more to be done here, but this was the
huge chunk that I couldn't really split things out of because this was
the interface change to TTI. I've tried to minimize all the other parts
of this. The follow up work should include at least:
1) Improving the TargetMachine interface by having it directly return
a TTI object. Because we have a non-pass object with value semantics
and an internal type erasure mechanism, we can narrow the interface
of the TargetMachine to *just* do what we need: build and return
a TTI object that we can then insert into the pass pipeline.
2) Make the TTI object be fully specialized for a particular function.
This will include splitting off a minimal form of it which is
sufficient for the inliner and the old pass manager.
3) Add a new pass manager analysis which produces TTI objects from the
target machine for each function. This may actually be done as part
of #2 in order to use the new analysis to implement #2.
4) Work on narrowing the API between TTI and the targets so that it is
easier to understand and less verbose to type erase.
5) Work on narrowing the API between TTI and its clients so that it is
easier to understand and less verbose to forward.
6) Try to improve the CRTP-based delegation. I feel like this code is
just a bit messy and exacerbating the complexity of implementing
the TTI in each target.
Many thanks to Eric and Hal for their help here. I ended up blocked on
this somewhat more abruptly than I expected, and so I appreciate getting
it sorted out very quickly.
Differential Revision: http://reviews.llvm.org/D7293
llvm-svn: 227669
I'm recommiting the codegen part of the patch.
The vectorizer part will be send to review again.
Masked Vector Load and Store Intrinsics.
Introduced new target-independent intrinsics in order to support masked vector loads and stores. The loop vectorizer optimizes loops containing conditional memory accesses by generating these intrinsics for existing targets AVX2 and AVX-512. The vectorizer asks the target about availability of masked vector loads and stores.
Added SDNodes for masked operations and lowering patterns for X86 code generator.
Examples:
<16 x i32> @llvm.masked.load.v16i32(i8* %addr, <16 x i32> %passthru, i32 4 /* align */, <16 x i1> %mask)
declare void @llvm.masked.store.v8f64(i8* %addr, <8 x double> %value, i32 4, <8 x i1> %mask)
Scalarizer for other targets (not AVX2/AVX-512) will be done in a separate patch.
http://reviews.llvm.org/D6191
llvm-svn: 223348
This reverts commit r222632 (and follow-up r222636), which caused a host
of LNT failures on an internal bot. I'll respond to the commit on the
list with a reproduction of one of the failures.
Conflicts:
lib/Target/X86/X86TargetTransformInfo.cpp
llvm-svn: 222936
Introduced new target-independent intrinsics in order to support masked vector loads and stores. The loop vectorizer optimizes loops containing conditional memory accesses by generating these intrinsics for existing targets AVX2 and AVX-512. The vectorizer asks the target about availability of masked vector loads and stores.
Added SDNodes for masked operations and lowering patterns for X86 code generator.
Examples:
<16 x i32> @llvm.masked.load.v16i32(i8* %addr, <16 x i32> %passthru, i32 4 /* align */, <16 x i1> %mask)
declare void @llvm.masked.store.v8f64(i8* %addr, <8 x double> %value, i32 4, <8 x i1> %mask)
Scalarizer for other targets (not AVX2/AVX-512) will be done in a separate patch.
http://reviews.llvm.org/D6191
llvm-svn: 222632
AVX2 is available.
According to IACA, the new lowering has a throughput of 8 cycles instead of 13
with the previous one.
Althought this lowering kicks in some SPECs benchmarks, the performance
improvement was within the noise.
Correctness testing has been done for the whole range of uint32_t with the
following program:
uint4 v = (uint4) {0,1,2,3};
uint32_t i;
//Check correctness over entire range for uint4 -> float4 conversion
for( i = 0; i < 1U << (32-2); i++ )
{
float4 t = test(v);
float4 c = correct(v);
if( 0xf != _mm_movemask_ps( t == c ))
{
printf( "Error @ %vx: %vf vs. %vf\n", v, c, t);
return -1;
}
v += 4;
}
Where "correct" is the old lowering and "test" the new one.
The patch adds a test case for the two custom lowering instruction.
It also modifies the vector cost model, which is why cast.ll and uitofp.ll are
modified.
2009-02-26-MachineLICMBug.ll is also modified because we now hoist 7
instructions instead of 4 (3 more constant loads).
rdar://problem/18153096>
llvm-svn: 221657
"Unroll" is not the appropriate name for this variable. Clang already uses
the term "interleave" in pragmas and metadata for this.
Differential Revision: http://reviews.llvm.org/D5066
llvm-svn: 217528
This patch adds support to recognize division by uniform power of 2 and modifies the cost table to vectorize division by uniform power of 2 whenever possible.
Updates Cost model for Loop and SLP Vectorizer.The cost table is currently only updated for X86 backend.
Thanks to Hal, Andrea, Sanjay for the review. (http://reviews.llvm.org/D4971)
llvm-svn: 216371
This lets us experiment with 512-bit vectorization without passing
force-vector-width manually.
The code generated for a simple integer memset loop is properly vectorized.
Disassembly is still broken for it though :(.
llvm-svn: 212634
This patch:
1) Improves the cost model for x86 alternate shuffles (originally
added at revision 211339);
2) Teaches the Cost Model Analysis pass how to analyze alternate shuffles.
Alternate shuffles are a special kind of blend; on x86, we can often
easily lowered alternate shuffled into single blend
instruction (depending on the subtarget features).
The existing cost model didn't take into account subtarget features.
Also, it had a couple of "dead" entries for vector types that are never
legal (example: on x86 types v2i32 and v2f32 are not legal; those are
always either promoted or widened to 128-bit vector types).
The new x86 cost model takes into account what target features we have
before returning the shuffle cost (i.e. the number of instructions
after the blend is lowered/expanded).
This patch also teaches the Cost Model Analysis how to identify and analyze
alternate shuffles (i.e. 'SK_Alternate' shufflevector instructions):
- added function 'isAlternateVectorMask';
- added some logic to check if an instruction is a alternate shuffle and, in
case, call the target specific TTI to get the corresponding shuffle cost;
- added a test to verify the cost model analysis on alternate shuffles.
llvm-svn: 212296
This patch adds support to recognize patterns such as fadd,fsub,fadd,fsub.../add,sub,add,sub... and
vectorizes them as vector shuffles if they are profitable.
These patterns of vector shuffle can later be converted to instructions such as addsubpd etc on X86.
Thanks to Arnold and Hal for the reviews. http://reviews.llvm.org/D4015
llvm-svn: 211339
lib/Target/X86/X86TargetTransformInfo.cpp: In member function ‘virtual unsigned int {anonymous}::X86TTI::getIntImmCost(unsigned int, unsigned int, const llvm::APInt&, llvm::Type*) const’:
lib/Target/X86/X86TargetTransformInfo.cpp:920:60: warning: enumeral and non-enumeral type in conditional expression [enabled by default]
This seems like an unhelpful warning, but there doesnt seem to be a controlling
flag, so add an explicit cast to silence the warning.
llvm-svn: 210806
This improves the X86 cost model for small constants with large types. Before
this commit we would even hoist trivial constants such as i96 2.
This is related to <rdar://problem/17070936>
llvm-svn: 210504
Currently the X86 backend doesn't support types larger than i128 very well. For
example an i192 multiply will assert in codegen when the 2nd argument is a constant and the constant got hoisted.
This fix changes the cost model to never hoist constants for types larger than
i128. Once the codegen issues have been resolved, the cost model can be updated
to allow also larger types.
This is related to <rdar://problem/16954938>
llvm-svn: 209162
The old method used by X86TTI to determine partial-unrolling thresholds was
messy (because it worked by testing target features), and also would not
correctly identify the target CPU if certain target features were disabled.
After some discussions on IRC with Chandler et al., it was decided that the
processor scheduling models were the right containers for this information
(because it is often tied to special uop dispatch-buffer sizes).
This does represent a small functionality change:
- For generic x86-64 (which uses the SB model and, thus, will get some
unrolling).
- For AMD cores (because they still currently use the SB scheduling model)
- For Haswell (based on benchmarking by Louis Gerbarg, it was decided to bump
the default threshold to 50; we're working on a test case for this).
Otherwise, nothing has changed for any other targets. The logic, however, has
been moved into BasicTTI, so other targets may now also opt-in to this
functionality simply by setting LoopMicroOpBufferSize in their processor
model definitions.
llvm-svn: 208289
The loop stream detector (LSD) on modern Intel cores, which optimizes the
execution of small loops, has limits on the number of taken branches in
addition to uop-count limits (modern AMD cores have similar limits).
Unfortunately, at the IR level, estimating the number of branches that will be
taken is difficult. For one thing, it strongly depends on later passes (block
placement, etc.). The original implementation took a conservative approach and
limited the maximal BB DFS depth of the loop. However, fairly-extensive
benchmarking by several of us has revealed that this is the wrong approach. In
fact, there are zero known cases where the branch limit prevents a detrimental
unrolling (but plenty of cases where it does prevent beneficial unrolling).
While we could improve the current branch counting logic by incorporating
branch probabilities, this further complication seems unjustified without a
motivating regression. Instead, unless and until a regression appears, the
branch counting will be removed.
llvm-svn: 208255
There is no need to check if we want to hoist the immediate value of an
shift instruction. Simply return TCC_Free right away.
This change is like r206101, but for X86.
rdar://problem/16190769
llvm-svn: 207692
This provides an initial implementation of getUnrollingPreferences for x86.
getUnrollingPreferences is used by the generic (concatenation) unroller, which
is distinct from the unrolling done by the loop vectorizer. Many modern x86
cores have some kind of uop cache and loop-stream detector (LSD) used to
efficiently dispatch small loops, and taking full advantage of this requires
unrolling small loops (small here means 10s of uops).
These caches also have limits on the number of taken branches in the loop, and
so we also cap the loop unrolling factor based on the maximum "depth" of the
loop. This is currently calculated with a partial DFS traversal (partial
because it will stop early if the path length grows too much). This is still an
approximation, and one that is both conservative (because it does not account
for branches eliminated via block placement) and optimistic (because it is only
recording the maximum depth over minimum paths). Nevertheless, because the
loops that fit in these uop caches are so small, it is not clear how much the
details matter.
The original set of patches posted for review produced the following test-suite
performance results (from the TSVC benchmark) at that time:
ControlLoops-dbl - 13% speedup
ControlLoops-flt - 15% speedup
Reductions-dbl - 7.5% speedup
llvm-svn: 205348
Simon Pilgrim