Current TargetTransformInfo can support throughput cost model and code size model, but sometimes we also need instruction latency cost model in different optimizations. Hal suggested we need a single public interface to query the different cost of an instruction. So I proposed following interface:
enum TargetCostKind {
TCK_RecipThroughput, ///< Reciprocal throughput.
TCK_Latency, ///< The latency of instruction.
TCK_CodeSize ///< Instruction code size.
};
int getInstructionCost(const Instruction *I, enum TargetCostKind kind) const;
All clients should mainly use this function to query the cost of an instruction, parameter <kind> specifies the desired cost model.
This patch also provides a simple default implementation of getInstructionLatency.
The default getInstructionLatency provides latency numbers for only small number of instruction classes, those latency numbers are only reasonable for modern OOO processors. It can be extended in following ways:
Add more detail into this function.
Add getXXXLatency function and call it from here.
Implement target specific getInstructionLatency function.
Differential Revision: https://reviews.llvm.org/D37170
llvm-svn: 312832
SLP vectorizer supports horizontal reductions for Add/FAdd binary
operations. Patch adds support for horizontal min/max reductions.
Function getReductionCost() is split to getArithmeticReductionCost() for
binary operation reductions and getMinMaxReductionCost() for min/max
reductions.
Patch fixes PR26956.
Differential revision: https://reviews.llvm.org/D27846
llvm-svn: 312791
isLegalAddressingMode() has recently gained the extra optional Instruction*
parameter, and therefore it can now do the job that previously only
isFoldableMemAccess() could do.
The SystemZ implementation of isLegalAddressingMode() has gained the
functionality of checking for offsets, which used to be done with
isFoldableMemAccess().
The isFoldableMemAccess() hook has been removed everywhere.
Review: Quentin Colombet, Ulrich Weigand
https://reviews.llvm.org/D35933
llvm-svn: 310463
This patch makes LSR generate better code for SystemZ in the cases of memory
intrinsics, Load->Store pairs or comparison of immediate with memory.
In order to achieve this, the following common code changes were made:
* New TTI hook: LSRWithInstrQueries(), which defaults to false. Controls if
LSR should do instruction-based addressing evaluations by calling
isLegalAddressingMode() with the Instruction pointers.
* In LoopStrengthReduce: handle address operands of memset, memmove and memcpy
as address uses, and call isFoldableMemAccessOffset() for any LSRUse::Address,
not just loads or stores.
SystemZ changes:
* isLSRCostLess() implemented with Insns first, and without ImmCost.
* New function supportedAddressingMode() that is a helper for TTI methods
looking at Instructions passed via pointers.
Review: Ulrich Weigand, Quentin Colombet
https://reviews.llvm.org/D35262https://reviews.llvm.org/D35049
llvm-svn: 308729
Now, getUserCost() only checks the src and dst types of EXT to decide it is free
or not. This change first checks the types, then calls isExtFreeImpl(), and
check if EXT can form ExtLoad at last. Currently, only AArch64 has customized
implementation of isExtFreeImpl() to check if EXT can be folded into its use.
Differential Revision: https://reviews.llvm.org/D34458
llvm-svn: 308076
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
I did this a long time ago with a janky python script, but now
clang-format has built-in support for this. I fed clang-format every
line with a #include and let it re-sort things according to the precise
LLVM rules for include ordering baked into clang-format these days.
I've reverted a number of files where the results of sorting includes
isn't healthy. Either places where we have legacy code relying on
particular include ordering (where possible, I'll fix these separately)
or where we have particular formatting around #include lines that
I didn't want to disturb in this patch.
This patch is *entirely* mechanical. If you get merge conflicts or
anything, just ignore the changes in this patch and run clang-format
over your #include lines in the files.
Sorry for any noise here, but it is important to keep these things
stable. I was seeing an increasing number of patches with irrelevant
re-ordering of #include lines because clang-format was used. This patch
at least isolates that churn, makes it easy to skip when resolving
conflicts, and gets us to a clean baseline (again).
llvm-svn: 304787
Summary:
The motivation example is like below which has 13 cases but only 2 distinct targets
```
lor.lhs.false2: ; preds = %if.then
switch i32 %Status, label %if.then27 [
i32 -7012, label %if.end35
i32 -10008, label %if.end35
i32 -10016, label %if.end35
i32 15000, label %if.end35
i32 14013, label %if.end35
i32 10114, label %if.end35
i32 10107, label %if.end35
i32 10105, label %if.end35
i32 10013, label %if.end35
i32 10011, label %if.end35
i32 7008, label %if.end35
i32 7007, label %if.end35
i32 5002, label %if.end35
]
```
which is compiled into a balanced binary tree like this on AArch64 (similar on X86)
```
.LBB853_9: // %lor.lhs.false2
mov w8, #10012
cmp w19, w8
b.gt .LBB853_14
// BB#10: // %lor.lhs.false2
mov w8, #5001
cmp w19, w8
b.gt .LBB853_18
// BB#11: // %lor.lhs.false2
mov w8, #-10016
cmp w19, w8
b.eq .LBB853_23
// BB#12: // %lor.lhs.false2
mov w8, #-10008
cmp w19, w8
b.eq .LBB853_23
// BB#13: // %lor.lhs.false2
mov w8, #-7012
cmp w19, w8
b.eq .LBB853_23
b .LBB853_3
.LBB853_14: // %lor.lhs.false2
mov w8, #14012
cmp w19, w8
b.gt .LBB853_21
// BB#15: // %lor.lhs.false2
mov w8, #-10105
add w8, w19, w8
cmp w8, #9 // =9
b.hi .LBB853_17
// BB#16: // %lor.lhs.false2
orr w9, wzr, #0x1
lsl w8, w9, w8
mov w9, #517
and w8, w8, w9
cbnz w8, .LBB853_23
.LBB853_17: // %lor.lhs.false2
mov w8, #10013
cmp w19, w8
b.eq .LBB853_23
b .LBB853_3
.LBB853_18: // %lor.lhs.false2
mov w8, #-7007
add w8, w19, w8
cmp w8, #2 // =2
b.lo .LBB853_23
// BB#19: // %lor.lhs.false2
mov w8, #5002
cmp w19, w8
b.eq .LBB853_23
// BB#20: // %lor.lhs.false2
mov w8, #10011
cmp w19, w8
b.eq .LBB853_23
b .LBB853_3
.LBB853_21: // %lor.lhs.false2
mov w8, #14013
cmp w19, w8
b.eq .LBB853_23
// BB#22: // %lor.lhs.false2
mov w8, #15000
cmp w19, w8
b.ne .LBB853_3
```
However, the inline cost model estimates the cost to be linear with the number
of distinct targets and the cost of the above switch is just 2 InstrCosts.
The function containing this switch is then inlined about 900 times.
This change use the general way of switch lowering for the inline heuristic. It
etimate the number of case clusters with the suitability check for a jump table
or bit test. Considering the binary search tree built for the clusters, this
change modifies the model to be linear with the size of the balanced binary
tree. The model is off by default for now :
-inline-generic-switch-cost=false
This change was originally proposed by Haicheng in D29870.
Reviewers: hans, bmakam, chandlerc, eraman, haicheng, mcrosier
Reviewed By: hans
Subscribers: joerg, aemerson, llvm-commits, rengolin
Differential Revision: https://reviews.llvm.org/D31085
llvm-svn: 301649
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
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
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
Currently when cost of scalar operations is evaluated the vector type is
used for scalar operations. Patch fixes this issue and fixes evaluation
of the vector operations cost.
Several test showed that vector cost model is too optimistic. It
allowed vectorization of 8 or less add/fadd operations, though scalar
code is faster. Actually, only for 16 or more operations vector code
provides better performance.
Differential Revision: https://reviews.llvm.org/D26277
llvm-svn: 288398
Summary:
Unrolled Loop Size calculations moved to a function.
Constant representing number of optimized instructions
when "back edge" becomes "fall through" replaced with
variable.
Some comments added.
Reviewers: mzolotukhin
Differential Revision: http://reviews.llvm.org/D21719
From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 286389
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
Reappy r284044 after revert in r284051. Krzysztof fixed the error in r284049.
The original summary:
This patch tries to fully unroll loops having break statement like this
for (int i = 0; i < 8; i++) {
if (a[i] == value) {
found = true;
break;
}
}
GCC can fully unroll such loops, but currently LLVM cannot because LLVM only
supports loops having exact constant trip counts.
The upper bound of the trip count can be obtained from calling
ScalarEvolution::getMaxBackedgeTakenCount(). Part of the patch is the
refactoring work in SCEV to prevent duplicating code.
The feature of using the upper bound is enabled under the same circumstance
when runtime unrolling is enabled since both are used to unroll loops without
knowing the exact constant trip count.
llvm-svn: 284053
This patch tries to fully unroll loops having break statement like this
for (int i = 0; i < 8; i++) {
if (a[i] == value) {
found = true;
break;
}
}
GCC can fully unroll such loops, but currently LLVM cannot because LLVM only
supports loops having exact constant trip counts.
The upper bound of the trip count can be obtained from calling
ScalarEvolution::getMaxBackedgeTakenCount(). Part of the patch is the
refactoring work in SCEV to prevent duplicating code.
The feature of using the upper bound is enabled under the same circumstance
when runtime unrolling is enabled since both are used to unroll loops without
knowing the exact constant trip count.
Differential Revision: https://reviews.llvm.org/D24790
llvm-svn: 284044
Summary:
In clang commit r268509 we started to invoke loop-unroll pass from the
driver even under -Os. However, we happen to not initialize optsize
thresholds properly, which si fixed with this change.
r268509 led to some big compile time regressions, because we started to
unroll some loops that we didn't unroll before. With this change I hope
to recover most of the regressions. We still are slightly slower than
before, because we do some checks here and there in loop-unrolling
before we bail out, but at least the slowdown is not that huge now.
Reviewers: hfinkel, chandlerc
Subscribers: mzolotukhin, llvm-commits
Differential Revision: https://reviews.llvm.org/D23388
llvm-svn: 279585
Refactored so that a LSRUse owns its fixups, as oppsed to letting the
LSRInstance own them. This makes it easier to rate formulas for
LSRUses, since the fixups are available directly. The Offsets vector
has been removed since it was no longer necessary.
New target hook isFoldableMemAccessOffset(), which is used during formula
rating.
For SystemZ, this is useful to express that loads and stores with
float or vector types with a big/negative offset should be avoided in
loops. Without this, LSR will generate a lot of negative offsets that
would require extra instructions for loading the address.
Updated tests:
test/CodeGen/SystemZ/loop-01.ll
Reviewed by: Quentin Colombet and Ulrich Weigand.
https://reviews.llvm.org/D19152
llvm-svn: 278927
Summary:
TargetBaseAlign is no longer required since LSV checks if target allows misaligned accesses.
A constant defining a base alignment is still needed for stack accesses where alignment can be adjusted.
Previous patch (D22936) was reverted because tests were failing. This patch also fixes the cause of those failures:
- x86 failing tests either did not have the right target, or the right alignment.
- NVPTX failing tests did not have the right alignment.
- AMDGPU failing test (merge-stores) should allow vectorization with the given alignment but the target info
considers <3xi32> a non-standard type and gives up early. This patch removes the condition and only checks
for a maximum size allowed and relies on the next condition checking for %4 for correctness.
This should be revisited to include 3xi32 as a MVT type (on arsenm's non-immediate todo list).
Note that checking the sizeInBits for a MVT is undefined (leads to an assertion failure),
so we need to create an EVT, hence the interface change in allowsMisaligned to include the Context.
Reviewers: arsenm, jlebar, tstellarAMD
Subscribers: jholewinski, arsenm, mzolotukhin, llvm-commits
Differential Revision: https://reviews.llvm.org/D23068
llvm-svn: 277735
The cost model should not assume vector casts get completely scalarized, since
on targets that have vector support, the common case is a partial split up to
the legal vector size. So, when a vector cast gets split, the resulting casts
end up legal and cheap.
Instead of pessimistically assuming scalarization, base TTI can use the costs
the concrete TTI provides for the split vector, plus a fudge factor to account
for the cost of the split itself. This fudge factor is currently 1 by default,
except on AMDGPU where inserts and extracts are considered free.
Differential Revision: http://reviews.llvm.org/D21251
llvm-svn: 274642
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
This reapplies commit r272385 with a fix. The build was failing when compiled
with gcc, but not with clang. With the fix, we now get the data layout from the
current TTI implementation, which will hopefully solve the issue.
llvm-svn: 272395
This patch refines the default cost for interleaved load groups having gaps. If
a load group has gaps, the legalized instructions corresponding to the unused
elements will be dead. Thus, we don't need to account for them in the cost
model. Instead, we only need to account for the fraction of legalized loads
that will actually be used.
Differential Revision: http://reviews.llvm.org/D20873
llvm-svn: 272385
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:
InlineCost's threshold is multiplied by this value. This lets us adjust
the inlining threshold up or down on a per-target basis. For example,
we might want to increase the threshold on targets where calls are
unusually expensive.
Reviewers: chandlerc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D18560
llvm-svn: 266405
The behavior of {MIN,MAX}NAN differs from that of {MIN,MAX}NUM when only
one of the inputs is NaN: -NUM will return the non-NaN argument while
-NAN would return NaN.
It is desirable to lower to @llvm.{min,max}num to -NAN if they don't
have a native instruction for -NUM. Notably, ARMv7 NEON's vmin has the
-NAN semantics.
N.B. Of course, it is only safe to do this if the intrinsic call is
marked nnan.
llvm-svn: 266279
PPC has a vector popcount, this lets the vectorizer use the correct cost
for it. Tweak X86 test to use an intrinsic that's actually scalarized (we
have a somewhat efficient lowering for vector popcount using SSE, the
cost model finds that now).
llvm-svn: 265005
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
Currently in LLVM's cost model, a vectorized arithmetic instruction will have
high cost if its type is split into multiple registers. However, this
punishment is too heavy and unnecessary. The overhead of the split should not
be on arithmetic instructions but instructions that implement the split. Note
that during vectorization we have calculated the register pressure, and we
only choose proper interleaving factor (and also vectorization factor) so
that we don't use more registers than the maximum number.
Here is a very simple example: if a vadd has the cost 1, and if we double VF
so that we need two registers to perform it, then its cost will become 4 with
the current implementation, which will prevent us to use larger VF.
Differential revision: http://reviews.llvm.org/D15159
llvm-svn: 254671
