[all-commits] [llvm/llvm-project] 78eaff: [llvm-exegesis] Loop unrolling for loop snippet re...

[Roman Lebedev via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 78eaff2ef8a984859a04f944522280360ee825aa
      https://github.com/llvm/llvm-project/commit/78eaff2ef8a984859a04f944522280360ee825aa
  Author: Roman Lebedev <lebedev.ri at gmail.com>
  Date:   2021-05-25 (Tue, 25 May 2021)

  Changed paths:
    M llvm/docs/CommandGuide/llvm-exegesis.rst
    M llvm/tools/llvm-exegesis/lib/BenchmarkResult.h
    M llvm/tools/llvm-exegesis/lib/BenchmarkRunner.cpp
    M llvm/tools/llvm-exegesis/lib/BenchmarkRunner.h
    M llvm/tools/llvm-exegesis/lib/SnippetRepetitor.cpp
    M llvm/tools/llvm-exegesis/lib/SnippetRepetitor.h
    M llvm/tools/llvm-exegesis/llvm-exegesis.cpp
    M llvm/unittests/tools/llvm-exegesis/X86/SnippetRepetitorTest.cpp

  Log Message:
  -----------
  [llvm-exegesis] Loop unrolling for loop snippet repetitor mode

I really needed this, like, factually, yesterday,
when verifying dependency breaking idioms for AMD Zen 3 scheduler model.

Consider the following example:
```
$ ./bin/llvm-exegesis --mode=inverse_throughput --snippets-file=/tmp/snippet.s --num-repetitions=1000000 --repetition-mode=duplicate
Check generated assembly with: /usr/bin/objdump -d /tmp/snippet-4a7e50.o
---
mode:            inverse_throughput
key:
  instructions:
    - 'VPXORYrr YMM0 YMM0 YMM0'
  config:          ''
  register_initial_values: []
cpu_name:        znver3
llvm_triple:     x86_64-unknown-linux-gnu
num_repetitions: 1000000
measurements:
  - { key: inverse_throughput, value: 0.31025, per_snippet_value: 0.31025 }
error:           ''
info:            ''
assembled_snippet: C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C5FDEFC0C3
...

```
What does it tell us?
So wait, it can only execute ~3 x86 AVX YMM PXOR zero-idioms per cycle?
That doesn't seem right. That's even less than there are pipes supporting this type of op.

Now, second example:
```
$ ./bin/llvm-exegesis --mode=inverse_throughput --snippets-file=/tmp/snippet.s --num-repetitions=1000000 --repetition-mode=loop
Check generated assembly with: /usr/bin/objdump -d /tmp/snippet-2418b5.o
---
mode:            inverse_throughput
key:
  instructions:
    - 'VPXORYrr YMM0 YMM0 YMM0'
  config:          ''
  register_initial_values: []
cpu_name:        znver3
llvm_triple:     x86_64-unknown-linux-gnu
num_repetitions: 1000000
measurements:
  - { key: inverse_throughput, value: 1.00011, per_snippet_value: 1.00011 }
error:           ''
info:            ''
assembled_snippet: 49B80800000000000000C5FDEFC0C5FDEFC04983C0FF75F2C3
...
```
Now that's just worse. Due to the looping, the throughput completely plummeted,
and now we can only do a single instruction/cycle!?

That's not great.
And final example:
```
$ ./bin/llvm-exegesis --mode=inverse_throughput --snippets-file=/tmp/snippet.s --num-repetitions=1000000 --repetition-mode=loop --loop-body-size=1000
Check generated assembly with: /usr/bin/objdump -d /tmp/snippet-c402e2.o
---
mode:            inverse_throughput
key:
  instructions:
    - 'VPXORYrr YMM0 YMM0 YMM0'
  config:          ''
  register_initial_values: []
cpu_name:        znver3
llvm_triple:     x86_64-unknown-linux-gnu
num_repetitions: 1000000
measurements:
  - { key: inverse_throughput, value: 0.167087, per_snippet_value: 0.167087 }
error:           ''
info:            ''
assembled_snippet: 49B80800000000000000C5FDEFC0C5FDEFC04983C0FF75F2C3
...
```

So if we merge the previous two approaches, do duplicate this single-instruction snippet 1000x
(loop-body-size/instruction count in snippet), and run a loop with 1000 iterations
over that duplicated/unrolled snippet, the measured throughput goes through the roof,
up to 5.9 instructions/cycle, which finally tells us that this idiom is zero-cycle!

Reviewed By: courbet

Differential Revision: https://reviews.llvm.org/D102522