Summary:
... from 8.
`VALIGNDZ128rmbik XMM0 XMM0 K1 XMM3 RDI i_0x1 i_0x0 i_0x1` instruction already has 9 components.
It does not matter much in terms of performance, but avoiding allocation seems to come with low cost here..
Reviewers: courbet, gchatelet
Reviewed By: courbet
Subscribers: tschuett, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D57654
llvm-svn: 353022
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
Summary:
SetVector uses both DenseSet and vector, which is time/memory inefficient. The points are represented as natural numbers so we can replace the DenseSet part by indexing into a vector<char> instead.
Don't cargo cult the pseudocode on the wikipedia DBSCAN page. This is a standard BFS style algorithm (the similar loops have been used several times in other LLVM components): every point is processed at most once, thus the queue has at most NumPoints elements. We represent it with a vector and allocate it outside of the loop to avoid allocation in the loop body.
We check `Processed[P]` to avoid enqueueing a point more than once, which also nicely saves us a `ClusterIdForPoint_[Q].isUndef()` check.
Many people hate the oneshot abstraction but some favor it, therefore we make a compromise, use a lambda to abstract away the neighbor adding process.
Delete the comment `assert(Neighbors.capacity() == (Points_.size() - 1));` as it is wrong.
llvm-svn: 350035
Summary:
Use `vector<char> Added + vector<size_t> ToProcess` to replace `SetVector ToProcess`
We also check `Added[P]` to enqueueing a point more than once, which
also saves us a `ClusterIdForPoint_[Q].isUndef()` check.
Reviewers: courbet, RKSimon, gchatelet, john.brawn, lebedev.ri
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D54442
........
Patch wasn't approved and breaks buildbots
llvm-svn: 349139
Summary:
Use `vector<char> Added + vector<size_t> ToProcess` to replace `SetVector ToProcess`
We also check `Added[P]` to enqueueing a point more than once, which
also saves us a `ClusterIdForPoint_[Q].isUndef()` check.
Reviewers: courbet, RKSimon, gchatelet, john.brawn, lebedev.ri
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D54442
llvm-svn: 349136
Apply review comments of https://reviews.llvm.org/D54185 to other target as well, specifically:
1. make anonymous namespaces as small as possible, avoid using static inside anonymous namespaces
2. Add missing header to some files
3. GetLoadImmediateOpcodem-> getLoadImmediateOpcode
4. Fix typo
Differential Revision: https://reviews.llvm.org/D54343
llvm-svn: 347309
Summary:
As it was pointed out in D54388+D54390, the maximal size of `Neighbors` is known,
it will contain at most Points_.size() minus one (the center of the cluster)
While that is the upper bound, meaning in the most cases, the actual count
will be much smaller, since D54390 made the allocation persistent,
we no longer have to worry about overly-optimistically `reserve()`ing.
Old: (D54393)
```
Performance counter stats for './bin/llvm-exegesis -mode=analysis -analysis-epsilon=100000 -benchmarks-file=/tmp/benchmarks.yaml -analysis-inconsistencies-output-file=/tmp/clusters.html' (16 runs):
6553.167456 task-clock (msec) # 1.000 CPUs utilized ( +- 0.21% )
...
6.5547 +- 0.0134 seconds time elapsed ( +- 0.20% )
```
New:
```
Performance counter stats for './bin/llvm-exegesis -mode=analysis -analysis-epsilon=100000 -benchmarks-file=/tmp/benchmarks.yaml -analysis-inconsistencies-output-file=/tmp/clusters.html' (16 runs):
6315.057872 task-clock (msec) # 0.999 CPUs utilized ( +- 0.24% )
...
6.3187 +- 0.0160 seconds time elapsed ( +- 0.25% )
```
And that is another -~4%.
Since this is the last (as of this moment) patch in this patch series,
it is a good time to summarize:
Old: (svn trunk, as stated in D54381)
```
$ time ./bin/llvm-exegesis -mode=analysis -analysis-epsilon=100000 -benchmarks-file=/tmp/benchmarks.yaml -analysis-inconsistencies-output-file=/tmp/clusters.html &> /dev/null
real 0m24.884s
user 0m24.099s
sys 0m0.785s
```
So these patches, on a given benchmark,
has decreased llvm-exegesis analysis time by 74.62%.
There surely is more room for further improvements.
D54514 may improve thins by -11.5% more (relative to this patch).
Parallelization may improve things further significantly, too.
Reviewers: courbet, MaskRay, RKSimon, gchatelet, john.brawn
Reviewed By: courbet, MaskRay
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D54415
llvm-svn: 347204
Summary:
Test data: 500kLOC of benchmark.yaml, 23Mb. (that is a subset of the actual uops benchmark i was trying to analyze!)
Old time: (D54381)
```
$ time ./bin/llvm-exegesis -mode=analysis -analysis-epsilon=100000 -benchmarks-file=/tmp/benchmarks.yaml -analysis-inconsistencies-output-file=/tmp/clusters.html &> /dev/null
real 0m10.487s
user 0m9.745s
sys 0m0.740s
```
New time:
```
$ time ./bin/llvm-exegesis -mode=analysis -analysis-epsilon=100000 -benchmarks-file=/tmp/benchmarks.yaml -analysis-inconsistencies-output-file=/tmp/clusters.html &> /dev/null
real 0m9.599s
user 0m8.824s
sys 0m0.772s
```
Not that much, around -9%. But that is not the good part yet, again.
Old:
* calls to allocation functions: 3347676
* temporary allocations: 277818
* bytes allocated in total (ignoring deallocations): 10.52 GB
New:
* calls to allocation functions: 2109712 (-36%)
* temporary allocations: 33112 (-88%)
* bytes allocated in total (ignoring deallocations): 4.43 GB (-58% *sic*)
Reviewers: courbet, MaskRay, RKSimon, gchatelet, john.brawn
Reviewed By: courbet, MaskRay
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D54382
llvm-svn: 347198
Summary:
Test data: 500kLOC of benchmark.yaml, 23Mb. (that is a subset of the actual uops benchmark i was trying to analyze!)
Old time:
```
$ time ./bin/llvm-exegesis -mode=analysis -analysis-epsilon=100000 -benchmarks-file=/tmp/benchmarks.yaml -analysis-inconsistencies-output-file=/tmp/clusters.html &> /dev/null
real 0m24.884s
user 0m24.099s
sys 0m0.785s
```
New time:
```
$ time ./bin/llvm-exegesis -mode=analysis -analysis-epsilon=100000 -benchmarks-file=/tmp/benchmarks.yaml -analysis-inconsistencies-output-file=/tmp/clusters.html &> /dev/null
real 0m10.469s
user 0m9.797s
sys 0m0.672s
```
So -60%. And that isn't the good bit yet.
Old:
* calls to allocation functions: 106560180 (yes, 107 *million* allocations.)
* bytes allocated in total (ignoring deallocations): 12.17 GB
New:
* calls to allocation functions: 3347676 (-96.86%) (just 3 mil)
* bytes allocated in total (ignoring deallocations): 10.52 GB (~2GB less)
---
Two points i want to raise:
* `std::unordered_set<>` should not have been used there in the first place.
It is banned by the https://llvm.org/docs/ProgrammersManual.html#other-set-like-container-options
* There is no tests, so i'm not fully sure this is correct.
Since it was unordered set, i guess there are zero restrictions on the order, and anything will be ok?
* I tried other containers suggested in https://llvm.org/docs/ProgrammersManual.html#set-like-containers-std-set-smallset-setvector-etc,
this `llvm::SetVector<>` seems to be best here.
Reviewers: courbet, MaskRay, RKSimon, gchatelet, john.brawn
Reviewed By: courbet
Subscribers: kristina, bobsayshilol, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D54381
llvm-svn: 347197
Summary:
This simplifies the code and moves everything to tablegen for consistency. This
also prepares the ground for adding issue counters.
Reviewers: gchatelet, john.brawn, jsji
Subscribers: nemanjai, mgorny, javed.absar, kbarton, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D54297
llvm-svn: 346489
This is patch to add PowerPC target to llvm-exegesis.
The target does just enough to be able to run llvm-exegesis in latency mode for at least some opcodes.
Differential Revision: https://reviews.llvm.org/D54185
llvm-svn: 346411
Summary:
Add unit tests to check the support for each supported format to avoid
regressions such as the one in PR36906.
Reviewers: gchatelet
Subscribers: tschuett, lebedev.ri, llvm-commits
Differential Revision: https://reviews.llvm.org/D54144
llvm-svn: 346330
Summary: They do not lower to actual MCInsts and have no scheduling info.
Reviewers: gchatelet
Subscribers: llvm-commits, tschuett
Differential Revision: https://reviews.llvm.org/D54147
llvm-svn: 346227
MachineModuleInfo can only be used in code using lib/CodeGen, hence we
can keep a more specific reference to LLVMTargetMachine rather than just
TargetMachine around.
llvm-svn: 346182
Summary: SNB is the only one that has P23 as a single proc res.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D53766
llvm-svn: 345480
Summary:
The pfm counters are now in the ExegesisTarget rather than the
MCSchedModel (PR39165).
This also compresses the pfm counter tables (PR37068).
Reviewers: RKSimon, gchatelet
Subscribers: mgrang, llvm-commits
Differential Revision: https://reviews.llvm.org/D52932
llvm-svn: 345243
Summary:
This allows simplifying references of llvm::foo with foo when the needs
come in the future.
Reviewers: courbet, gchatelet
Reviewed By: gchatelet
Subscribers: javed.absar, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D53455
llvm-svn: 344922
Summary:
We only support the first form because we rely on information that is
only available there.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D53430
llvm-svn: 344782
Summary:
This was lost during refactoring in rL342644.
Fix and simplify simplify value size handling: always go through a 80 bit value,
because the value can be 1 byte). Add unit tests.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D53423
llvm-svn: 344779
Summary:
We try to recover gracefully on instructions that would crash the
program.
This includes some refactoring of runMeasurement() implementations.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D53371
llvm-svn: 344695
Summary: LatencyGenerator now computes all possible mode of serial execution for an Instruction upfront and generates CodeTemplate for the ones that give the best results (e.g. no need to generate a two instructions snippet when repeating a single one would do). The next step is to generate even more configurations for cases (e.g. for XOR we should generate "XOR EAX, EAX, EAX" and "XOR EAX, EAX, EBX")
Reviewers: courbet
Reviewed By: courbet
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D53320
llvm-svn: 344689
Summary: This is part one of the change where I simply changed the signature of the functions. More work need to be done to actually produce more than one CodeTemplate per instruction.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D53209
llvm-svn: 344493
Summary: Simplify code by having LLVMState hold the RegisterAliasingTrackerCache.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D53078
llvm-svn: 344143
When fillMachineFunction generates a return on targets without a return opcode
(such as AArch64) it should pass an empty set of registers as the return
registers, not 0 which means register number zero.
Differential Revision: https://reviews.llvm.org/D53074
llvm-svn: 344139
Summary:
This moves checking logic into the accessors and makes the structure smaller.
It will also help when/if Operand are generated from the TD files.
Subscribers: tschuett, courbet, llvm-commits
Differential Revision: https://reviews.llvm.org/D52982
llvm-svn: 344028
deserializeMCOperand - ensure that we at least match the first character of the sscanf pattern before calling
This reduces llvm-exegesis uops analysis of the instructions supported from btver2 from 5m13s to 2m1s on debug builds.
llvm-svn: 343690
Summary: This is prelimineary to moving random functions to SnippetGenerator.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D52718
llvm-svn: 343456
Summary: The key is now the resource name, not the resource id.
Reviewers: gchatelet
Subscribers: tschuett, RKSimon, llvm-commits
Differential Revision: https://reviews.llvm.org/D52607
llvm-svn: 343208
Summary:
THis is a backwards-compatible change (existing files will work as
expected).
See PR39082.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D52546
llvm-svn: 343108
Summary: This is is preparation of exploring value ranges.
Reviewers: courbet
Reviewed By: courbet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D52542
llvm-svn: 343098
Summary: This is a NFC in preparation of exporting the initial registers as part of the YAML dump
Reviewers: courbet
Reviewed By: courbet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D52427
llvm-svn: 342967
Summary:
This is a step towards fixing PR38048.
Note that right now the measurements are given per instruction. We'll
need to give measurements a per code snippet and update the analysis (PR38731).
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D52041
llvm-svn: 342947
Summary:
The `set` statements was incorrectly reading the value of the local variable and
setting the value of the parent variable.
Reviewers: tycho, gchatelet, john.brawn
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D52343
llvm-svn: 342865
Summary:
Added function to set a register to a particular value + tests.
Add EFLAGS test, use new setRegTo instead of setRegToConstant.
Reviewers: courbet, javed.absar
Subscribers: llvm-commits, tschuett, mgorny
Differential Revision: https://reviews.llvm.org/D52297
llvm-svn: 342644
rL342465 is breaking the MSVC buildbots, but I need to revert this dependent revision as well.
Summary:
Added function to set a register to a particular value + tests.
Add EFLAGS test, use new setRegTo instead of setRegToConstant.
Reviewers: courbet, javed.absar
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D51856
llvm-svn: 342489
Summary:
Added function to set a register to a particular value + tests.
Add EFLAGS test, use new setRegTo instead of setRegToConstant.
Reviewers: courbet, javed.absar
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D51856
llvm-svn: 342466
Summary: This will be useful to generate many configurations and test instruction regimes (NaN, Inf, subnormal, normal).
Reviewers: courbet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D51858
llvm-svn: 342369
Summary:
The snippet-generation part goes to the SnippetGenerator class.
This will allow benchmarking arbitrary code (see PR38437).
Reviewers: gchatelet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D51979
llvm-svn: 342117
Some asm has double spaces between operands, the deserializer was keeping these empty split pieces, causing assertions later on:
'ADC16mi RDI i_0x1x i_0x0x i_0x1x'
llvm-svn: 341799
The target does just enough to be able to run llvm-exegesis in latency mode for
at least some opcodes.
Differential Revision: https://reviews.llvm.org/D48780
llvm-svn: 336187
Currently the llvm-exegesis native architecture is determined by comparing the
llvm native architecture with X86, so to add a new target would mean adding a
new check. Change this to building up a list of the targets llvm-exegesis
supports then using that, as this means that when adding a new target you just
add the target to the list of supported targets.
Differential Revision: https://reviews.llvm.org/D48778
llvm-svn: 336105
Currently the cycle counter is taken from the subtarget schedule model, which
isn't any use if the subtarget doesn't have one. Delegate the decision to the
target benchmark runner, as it may know better what to do in that case, with
the default being the current behaviour.
Differential Revision: https://reviews.llvm.org/D48779
llvm-svn: 336099
Summary:
This enables the X86-specific X86FloatingPointStackifierPass, and allow
llvm-exegesis to generate and measure X87 latency/uops for some FP ops.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48592
llvm-svn: 335815
Summary:
This allows targets to override code generation for some instructions.
As an example of override, this also moves ad-hoc instruction filtering
for X86 into the X86 ExegesisTarget.
Reviewers: gchatelet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48587
llvm-svn: 335582
Summary:
This ensures that the snippet always sees the same values for registers,
making measurements reproducible.
This will also allow exploring different values.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48542
llvm-svn: 335465
Summary: Pretty much everything we need is in llvm::TargetMachine.
Reviewers: gchatelet
Subscribers: llvm-commits, tschuett
Differential Revision: https://reviews.llvm.org/D48428
llvm-svn: 335237
Summary: Introducing a Prototype object to capture Variables that must be set but keeps degrees of freedom as Invalid. This allows exploring non constraint variables later on.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48316
llvm-svn: 335105
Summary: This is a step towards implementing memory operands and X87.
Reviewers: gchatelet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48210
llvm-svn: 335038
Summary:
On hover, the whole asm snippet is displayed, including operands.
This requires the actual assembly output instead of just the MCInsts:
This is because some pseudo-instructions get lowered to actual target
instructions during codegen (e.g. ABS_Fp32 -> SSE or X87).
Reviewers: gchatelet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48164
llvm-svn: 334805
Summary:
Get rid of OpcodeName.
To remove the opcode name from an old file:
```
cat old_file | sed '/opcode_name.*/d'
```
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48121
llvm-svn: 334691
Summary: Previous design was relying on the 'mutate' keyword and was quite confusing. This version separate mutable from immutable data and makes it clearer what changes and what doesn't.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48020
llvm-svn: 334596
Summary: This patch moves linking of libpfm from different places to a single one.
Reviewers: courbet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48075
llvm-svn: 334499
This breaks the OpenFlags enumeration into two separate
enumerations: OpenFlags and CreationDisposition. The first
controls the behavior of the API depending on whether or not
the target file already exists, and is not a flags-based
enum. The second controls more flags-like values.
This yields a more easy to understand API, while also allowing
flags to be passed to the openForRead api, where most of the
values didn't make sense before. This also makes the apis more
testable as it becomes easy to enumerate all the configurations
which make sense, so I've added many new tests to exercise all
the different values.
llvm-svn: 334221
Summary: BenchmarkRunner subclasses can now create many configurations - although this patch still generates one.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D47877
llvm-svn: 334197
Summary: This is the first step to have the BenchmarkRunner create and measure many different configurations (different initial values for instance).
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D47826
llvm-svn: 334169
Summary: BenchmarkResult IO functions now return an Error or Expected so caller can deal take proper action.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D47868
llvm-svn: 334167
Moves the Mode field out of the Key. The existing yaml benchmark results can be fixed with the following script:
```
readonly FILE=$1
readonly MODE=latency # Change to uops to fix a uops benchmark.
cat $FILE | \
sed "/^\ \+mode:\ \+$MODE$/d" | \
sed "/^cpu_name.*$/i mode: $MODE"
```
Differential Revision: https://reviews.llvm.org/D47813
Authored by: Guillaume Chatelet
llvm-svn: 334079
We want llvm-exegesis to explore instructions (effect of initial register values, effect of operand selection). To enable this a BenchmarkResult muststore all the relevant data in its key. This patch starts adding such data. Here we simply allow to store the generated instructions, following patches will add operands and initial values for registers.
https://reviews.llvm.org/D47764
Authored by: Guilluame Chatelet
llvm-svn: 334008
Summary:
We now highlight any sched classes whose measurements do not match the
LLVM SchedModel. "bad" clusters are marked in red.
Screenshot in phabricator diff.
Reviewers: gchatelet
Subscribers: tschuett, mgrang, RKSimon, llvm-commits
Differential Revision: https://reviews.llvm.org/D47639
llvm-svn: 333884
Summary:
I'm slowly looking into a new X86 scheduler model,
for AMD Bulldozer CPU, model 2 (bdver2, Piledriver).
And naturally, i have hit that assert :)
I happened to know what it meant, and how to fix it,
but that is not too common knowledge.
Reviewers: courbet, RKSimon
Reviewed By: courbet
Subscribers: tschuett, llvm-commits, craig.topper
Differential Revision: https://reviews.llvm.org/D47572
llvm-svn: 333632
Roman Lebedev