Bug 1863497 Comment 4 Edit History

Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.

Original comment by
on 
Some benchmark results with sccache direct/processor mode:
```
hyperfine --warmup 1 "bash bench-without.sh" "bash bench-with.sh"
Benchmark 1: bash bench-without.sh
  Time (mean ± σ):     69.406 s ±  4.229 s    [User: 275.969 s, System: 53.104 s]
  Range (min … max):   63.897 s … 75.104 s    10 runs

Benchmark 2: bash bench-with.sh
  Time (mean ± σ):     30.454 s ±  1.541 s    [User: 271.775 s, System: 51.225 s]
  Range (min … max):   29.026 s … 33.550 s    10 runs

Summary
  bash bench-with.sh ran
    2.28 ± 0.18 times faster than bash bench-without.sh
```

My ugly scripts

_bench-without.sh:_
```
cd ~/.mozbuild/sccache/
rm -f *
ln -s ~/.cargo/bin/sccache
cd -

killall sccache

~/.cargo/bin/sccache --start-server

fd '\.o$' obj-* | xargs rm && time ./mach build
~/.cargo/bin/sccache --show-adv-stats
```

_bench-with.sh_
```
cd ~/.mozbuild/sccache/
rm -f *
ln -s ~/.cargo/bin/sccache
cd -

killall sccache

SCCACHE_DIRECT=true ~/.cargo/bin/sccache --start-server

fd '\.o$' obj-* | xargs rm && time ./mach build
~/.cargo/bin/sccache --show-adv-stats
```
Revision 1 by
on 
Some benchmark results with sccache direct/processor mode (took 19m to run this benchmark) :
```
hyperfine --warmup 1 "bash bench-without.sh" "bash bench-with.sh"
Benchmark 1: bash bench-without.sh
  Time (mean ± σ):     69.406 s ±  4.229 s    [User: 275.969 s, System: 53.104 s]
  Range (min … max):   63.897 s … 75.104 s    10 runs

Benchmark 2: bash bench-with.sh
  Time (mean ± σ):     30.454 s ±  1.541 s    [User: 271.775 s, System: 51.225 s]
  Range (min … max):   29.026 s … 33.550 s    10 runs

Summary
  bash bench-with.sh ran
    2.28 ± 0.18 times faster than bash bench-without.sh
```

My ugly scripts

_bench-without.sh:_
```
cd ~/.mozbuild/sccache/
rm -f *
ln -s ~/.cargo/bin/sccache
cd -

killall sccache

~/.cargo/bin/sccache --start-server

fd '\.o$' obj-* | xargs rm && time ./mach build
~/.cargo/bin/sccache --show-adv-stats
```

_bench-with.sh_
```
cd ~/.mozbuild/sccache/
rm -f *
ln -s ~/.cargo/bin/sccache
cd -

killall sccache

SCCACHE_DIRECT=true ~/.cargo/bin/sccache --start-server

fd '\.o$' obj-* | xargs rm && time ./mach build
~/.cargo/bin/sccache --show-adv-stats
```
Revision 2 by
on 
with
```
cargo install sccache
```

Some benchmark results with sccache direct/processor mode (took 19m to run this benchmark) :
```
hyperfine --warmup 1 "bash bench-without.sh" "bash bench-with.sh"
Benchmark 1: bash bench-without.sh
  Time (mean ± σ):     69.406 s ±  4.229 s    [User: 275.969 s, System: 53.104 s]
  Range (min … max):   63.897 s … 75.104 s    10 runs

Benchmark 2: bash bench-with.sh
  Time (mean ± σ):     30.454 s ±  1.541 s    [User: 271.775 s, System: 51.225 s]
  Range (min … max):   29.026 s … 33.550 s    10 runs

Summary
  bash bench-with.sh ran
    2.28 ± 0.18 times faster than bash bench-without.sh
```

My ugly scripts

_bench-without.sh:_
```
cd ~/.mozbuild/sccache/
rm -f *
ln -s ~/.cargo/bin/sccache
cd -

killall sccache

~/.cargo/bin/sccache --start-server

fd '\.o$' obj-* | xargs rm && time ./mach build
~/.cargo/bin/sccache --show-adv-stats
```

_bench-with.sh_
```
cd ~/.mozbuild/sccache/
rm -f *
ln -s ~/.cargo/bin/sccache
cd -

killall sccache

SCCACHE_DIRECT=true ~/.cargo/bin/sccache --start-server

fd '\.o$' obj-* | xargs rm && time ./mach build
~/.cargo/bin/sccache --show-adv-stats
```

Back to Bug 1863497 Comment 4