The previous patch (D148980) didn't set the InstrIdxForVirtReg correctly
in genAlternativeDpCodeSequence(). It causes vnni lit test failure when
LLVM_ENABLE_EXPENSIVE_CHECKS is on.
"vpmaddwd + vpaddd" can be combined to vpdpwssd and the latency is
reduced after combination. However when vpdpwssd is in a critical path
the combination get less ILP. It happens when vpdpwssd is in a loop, the
vpmaddwd can be executed in parallel in multi-iterations while vpdpwssd
has data dependency for each iterations. If vpaddd is in a critical path
while vpmaddwd is not, it is profitable to split vpdpwssd into "vpmaddwd
+ vpaddd".
This patch is based on the machine combiner framework to acheive decision
on "vpmaddwd + vpaddd" combination. The typical example code is as
below.
```
__m256i foo(int cnt, __m256i c, __m256i b, __m256i *p) {
for (int i = 0; i < cnt; ++i) {
__m256i a = p[i];
__m256i m = _mm256_madd_epi16 (b, a);
c = _mm256_add_epi32(m, c);
}
return c;
}
```
Differential Revision: https://reviews.llvm.org/D148980
This change initializes the members TSI, LI, DT, PSI, and ORE pointer feilds of the SelectOptimize class to nullptr.
Reviewed By: LuoYuanke
Differential Revision: https://reviews.llvm.org/D148303
For in-order cores MachineCombiner makes better decisions when the critical path
is calculated only for the current basic block and does not take into account
other blocks from the trace.
This patch adds a virtual method to TargetInstrInfo to allow each target decide
which strategy to use.
Depends on D140541
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D140542
We are about to allow different trace strategies for MachineCombiner. Make
the name of the ensemble strategy-neutral.
Depends on D140540
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D140541
Make forward declaration possible to reduce amount of dependencies and reduce
re-compilation burden caused by further patches.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D140539
This change adjusts the cost modeling used when the target does not have a schedule model with individual instruction latencies. After this change, we use the default latency information available from TargetSchedule. The default latency information essentially ends up treating most instructions as latency 1, with a few "expensive" ones getting a higher cost.
Previously, we unconditionally applied the first legal pattern - without any consideration of profitability. As a result, this change both prevents some patterns being applied, and changes which patterns are exercised. (i.e. previously the first pattern was applied, afterwards, maybe the second one is because the first wasn't profitable.)
The motivation here is two fold.
First, this brings the default behavior in line with the behavior when -mcpu or -mtune is specified. This improves test coverage, and generally makes it less likely we will have bad surprises when providing more information to the compiler.
Second, this enables some reassociation for ILP by default. Despite being unconditionally enabled, the prior code tended to "reassociate" repeatedly through an entire chain and simply moving the first operand to the end. The result was still a serial chain, just a different one. With this change, one of the intermediate transforms is unprofitable and we end up with a partially flattened tree.
Note that the resulting code diffs show significant room for improvement in the basic algorithm. I am intentionally excluding those from this patch.
For the test diffs, I don't seen any concerning regressions. I took a fairly close look at the RISCV ones, but only skimmed the x86 (particularly vector x86) changes.
Differential Revision: https://reviews.llvm.org/D141017
Use deduction guides instead of helper functions.
The only non-automatic changes have been:
1. ArrayRef(some_uint8_pointer, 0) needs to be changed into ArrayRef(some_uint8_pointer, (size_t)0) to avoid an ambiguous call with ArrayRef((uint8_t*), (uint8_t*))
2. CVSymbol sym(makeArrayRef(symStorage)); needed to be rewritten as CVSymbol sym{ArrayRef(symStorage)}; otherwise the compiler is confused and thinks we have a (bad) function prototype. There was a few similar situation across the codebase.
3. ADL doesn't seem to work the same for deduction-guides and functions, so at some point the llvm namespace must be explicitly stated.
4. The "reference mode" of makeArrayRef(ArrayRef<T> &) that acts as no-op is not supported (a constructor cannot achieve that).
Per reviewers' comment, some useless makeArrayRef have been removed in the process.
This is a follow-up to https://reviews.llvm.org/D140896 that introduced
the deduction guides.
Differential Revision: https://reviews.llvm.org/D140955
This patch adds tranformation of fmul+fadd/fsub chains to fused multiply
instructions:
* fmul+fadd->fmadd
* fmul+fsub->fmsub/fnmsub
We also will try to combine these instructions if the fmul has more than one use
and cannot be deleted. However, removing the dependence between fmul and fadd can
still be profitable, and we rely on machine combiner approximations of scheduling.
Differential Revision: https://reviews.llvm.org/D136764
If an MI will not generate a target instruction, we should not compute its
latency. Then we can compute more precise instruction sequence cost, and get
better result.
Differential Revision: https://reviews.llvm.org/D129615
Add a new pattern A - (B + C) ==> (A - B) - C to give machine combiner a chance
to evaluate which instruction sequence has lower latency.
Differential Revision: https://reviews.llvm.org/D124564
This reverts commit 7f230feeeac8a67b335f52bd2e900a05c6098f20.
Breaks CodeGenCUDA/link-device-bitcode.cu in check-clang,
and many LLVM tests, see comments on https://reviews.llvm.org/D121169
Expanding on D109750.
Since `DBG_VALUE` instructions have final register validity determined in
`LDVImpl::handleDebugValue`, there is no apparent reason to immediately prune
unused register operands as their defs are erased. Consequently, this renders
`MachineInstr::eraseFromParentAndMarkDBGValuesForRemoval` moot; gaining a
substantial performance improvement.
The only necessary changes involve making relevant passes consider invalid
DBG_VALUE vregs uses as valid.
Reviewed By: MatzeB
Differential Revision: https://reviews.llvm.org/D112852
Reassociating some patterns to generate more fma instructions to
reduce register pressure.
Reviewed By: jsji
Differential Revision: https://reviews.llvm.org/D92071
Reassociating some patterns to generate more fma instructions to
reduce register pressure.
Reviewed By: jsji
Differential Revision: https://reviews.llvm.org/D92071
add a new goal MustReduceRegisterPressure for machine combiner pass.
PowerPC will use this new goal to do some register pressure related optimization.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D92068
This patch tries to reassociate two patterns related to FMA to expose
more ILP on PowerPC.
// Pattern 1:
// A = FADD X, Y (Leaf)
// B = FMA A, M21, M22 (Prev)
// C = FMA B, M31, M32 (Root)
// -->
// A = FMA X, M21, M22
// B = FMA Y, M31, M32
// C = FADD A, B
// Pattern 2:
// A = FMA X, M11, M12 (Leaf)
// B = FMA A, M21, M22 (Prev)
// C = FMA B, M31, M32 (Root)
// -->
// A = FMUL M11, M12
// B = FMA X, M21, M22
// D = FMA A, M31, M32
// C = FADD B, D
Reviewed By: jsji
Differential Revision: https://reviews.llvm.org/D80175
Summary:
Split off of D67120.
Add the profile guided size optimization instrumentation / queries in the code
gen or target passes. This doesn't enable the size optimizations in those passes
yet as they are currently disabled in shouldOptimizeForSize (for non-IR pass
queries).
A second try after reverted D71072.
Reviewers: davidxl
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71149
Summary:
Split off of D67120.
Add the profile guided size optimization instrumentation / queries in the code
gen or target passes. This doesn't enable the size optimizations in those passes
yet as they are currently disabled in shouldOptimizeForSize (for non-IR pass
queries).
Reviewers: davidxl
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71072
This file lists every pass in LLVM, and is included by Pass.h, which is
very popular. Every time we add, remove, or rename a pass in LLVM, it
caused lots of recompilation.
I found this fact by looking at this table, which is sorted by the
number of times a file was changed over the last 100,000 git commits
multiplied by the number of object files that depend on it in the
current checkout:
recompiles touches affected_files header
342380 95 3604 llvm/include/llvm/ADT/STLExtras.h
314730 234 1345 llvm/include/llvm/InitializePasses.h
307036 118 2602 llvm/include/llvm/ADT/APInt.h
213049 59 3611 llvm/include/llvm/Support/MathExtras.h
170422 47 3626 llvm/include/llvm/Support/Compiler.h
162225 45 3605 llvm/include/llvm/ADT/Optional.h
158319 63 2513 llvm/include/llvm/ADT/Triple.h
140322 39 3598 llvm/include/llvm/ADT/StringRef.h
137647 59 2333 llvm/include/llvm/Support/Error.h
131619 73 1803 llvm/include/llvm/Support/FileSystem.h
Before this change, touching InitializePasses.h would cause 1345 files
to recompile. After this change, touching it only causes 550 compiles in
an incremental rebuild.
Reviewers: bkramer, asbirlea, bollu, jdoerfert
Differential Revision: https://reviews.llvm.org/D70211
Over a year ago, MachineInstr gained a fourth boolean parameter that occurs
before the TII pointer. When this happened, several places started accidentally
passing TII into this boolean parameter instead of the TII parameter.
llvm-svn: 362312
This patch removes hidden codegen flag -print-schedule effectively reverting the
logic originally committed as r300311
(https://llvm.org/viewvc/llvm-project?view=revision&revision=300311).
Flag -print-schedule was originally introduced by r300311 to address PR32216
(https://bugs.llvm.org/show_bug.cgi?id=32216). That bug was about adding "Better
testing of schedule model instruction latencies/throughputs".
These days, we can use llvm-mca to test scheduling models. So there is no longer
a need for flag -print-schedule in LLVM. The main use case for PR32216 is
now addressed by llvm-mca.
Flag -print-schedule is mainly used for debugging purposes, and it is only
actually used by x86 specific tests. We already have extensive (latency and
throughput) tests under "test/tools/llvm-mca" for X86 processor models. That
means, most (if not all) existing -print-schedule tests for X86 are redundant.
When flag -print-schedule was first added to LLVM, several files had to be
modified; a few APIs gained new arguments (see for example method
MCAsmStreamer::EmitInstruction), and MCSubtargetInfo/TargetSubtargetInfo gained
a couple of getSchedInfoStr() methods.
Method getSchedInfoStr() had to originally work for both MCInst and
MachineInstr. The original implmentation of getSchedInfoStr() introduced a
subtle layering violation (reported as PR37160 and then fixed/worked-around by
r330615).
In retrospect, that new API could have been designed more optimally. We can
always query MCSchedModel to get the latency and throughput. More importantly,
the "sched-info" string should not have been generated by the subtarget.
Note, r317782 fixed an issue where "print-schedule" didn't work very well in the
presence of inline assembly. That commit is also reverted by this change.
Differential Revision: https://reviews.llvm.org/D57244
llvm-svn: 353043
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
The DEBUG() macro is very generic so it might clash with other projects.
The renaming was done as follows:
- git grep -l 'DEBUG' | xargs sed -i 's/\bDEBUG\s\?(/LLVM_DEBUG(/g'
- git diff -U0 master | ../clang/tools/clang-format/clang-format-diff.py -i -p1 -style LLVM
- Manual change to APInt
- Manually chage DOCS as regex doesn't match it.
In the transition period the DEBUG() macro is still present and aliased
to the LLVM_DEBUG() one.
Differential Revision: https://reviews.llvm.org/D43624
llvm-svn: 332240
The TargetSchedModel is always initialized using the TargetSubtargetInfo's
MCSchedModel and TargetInstrInfo, so we don't need to extract those and
pass 3 parameters to init().
Differential Revision: https://reviews.llvm.org/D44789
llvm-svn: 329540
In D41587, @mssimpso discovered that the order of some patterns for
AArch64 was sub-optimal. I thought a bit about how we could avoid that
case in the future. I do not think there is a need for evaluating all
patterns for now. But this patch adds an extra (expensive) check, that
evaluates the latencies of all patterns, and ensures that the latency
saved decreases for subsequent patterns.
This catches the sub-optimal order fixed in D41587, but I am not
entirely happy with the check, as it only applies to sub-optimal
patterns seen while building with EXPENSIVE_CHECKS on. It did not
discover any other sub-optimal pattern ordering.
Reviewers: Gerolf, spatel, mssimpso
Reviewed By: Gerolf, mssimpso
Differential Revision: https://reviews.llvm.org/D41766
llvm-svn: 323873
Summary:
When calculating the RootLatency, we add up all the latencies of the
deleted instructions. But for NewRootLatency we only add the latency of
the new root instructions, ignoring the latencies of the other
instructions inserted. This leads the combiner to underestimate the cost
of patterns which add multiple instructions. This patch fixes that by
summing up the latencies of all new instructions. For NewRootNode, the
more complex getLatency function is used.
Note that we may be slightly more precise than just summing up
all latencies. For example, consider a pattern like
r1 = INS1 ..
r2 = INS2 ..
r3 = INS3 r1, r2
I think in some other places, the total latency of the pattern would be
estimated as lat(INS3) + max(lat(INS1), lat(INS2)). If you consider
that worth changing, I think it would be best to do in a follow-up
patch.
Reviewers: Gerolf, sebpop, spop, fhahn
Reviewed By: fhahn
Subscribers: evandro, llvm-commits
Differential Revision: https://reviews.llvm.org/D40307
llvm-svn: 319951
All these headers already depend on CodeGen headers so moving them into
CodeGen fixes the layering (since CodeGen depends on Target, not the
other way around).
llvm-svn: 318490
This header includes CodeGen headers, and is not, itself, included by
any Target headers, so move it into CodeGen to match the layering of its
implementation.
llvm-svn: 317647
