Re: [PATCH 2/2] for-each-ref: add --count-matches option

Am 27.06.23 um 12:05 schrieb Phillip Wood:

On 27/06/2023 08:30, Jeff King wrote:

quoted

On Mon, Jun 26, 2023 at 03:09:57PM +0000, Derrick Stolee via GitGitGadget wrote:

quoted

+for pattern in "refs/heads/" "refs/tags/" "refs/remotes"
+do
+    test_perf "count $pattern: git for-each-ref | wc -l" "
+        git for-each-ref $pattern | wc -l
+    "
+
+    test_perf "count $pattern: git for-each-ref --count-match" "
+        git for-each-ref --count-matches $pattern
+    "
+done

I don't think this is a very realistic perf test, because for-each-ref
is doing a bunch of work to generate its default format, only to have
"wc" throw most of it away. Doing:

   git for-each-ref --format='%(refname)' | wc -l

That's a good point. I wondered if using a short fixed format string was even better so I tried

git init test
cd test
git commit --allow-empty -m initial
seq 0 100000 | sed "s:\(.*\):create refs/heads/some-prefix/\1 $(git rev-parse HEAD):" | git update-ref --stdin
git pack-refs --all
hyperfine -L fmt "","--format=%\(refname\)","--format=x" 'git for-each-ref {fmt} refs/heads/ | wc -l'

Which gives

Benchmark 1: git for-each-ref  refs/heads/ | wc -l
  Time (mean ± σ):      1.150 s ±  0.010 s    [User: 0.494 s, System: 0.637 s]
  Range (min … max):    1.140 s …  1.170 s    10 runs

Benchmark 2: git for-each-ref --format=%\(refname\) refs/heads/ | wc -l
  Time (mean ± σ):      66.0 ms ±   0.3 ms    [User: 58.9 ms, System: 9.5 ms]
  Range (min … max):    65.2 ms …  67.1 ms    43 runs

Benchmark 3: git for-each-ref --format=x refs/heads/ | wc -l
  Time (mean ± σ):      63.0 ms ±   0.5 ms    [User: 54.3 ms, System: 9.6 ms]
  Range (min … max):    62.3 ms …  65.4 ms    44 runs

Summary
  git for-each-ref --format=x refs/heads/ | wc -l ran
    1.05 ± 0.01 times faster than git for-each-ref --format=%\(refname\) refs/heads/ | wc -l
   18.25 ± 0.20 times faster than git for-each-ref  refs/heads/ | wc -l

You don't need no "x", by the way; using the empty string saves some
cycles for me.  In my Git clone, no special setup, 9931 refs:

Benchmark 1: git for-each-ref --format=x | wc -l
  Time (mean ± σ):      25.1 ms ±   0.1 ms    [User: 8.7 ms, System: 16.8 ms]
  Range (min … max):    24.9 ms …  25.6 ms    109 runs

Benchmark 2: git for-each-ref --format= | wc -l
  Time (mean ± σ):      24.6 ms ±   0.1 ms    [User: 8.3 ms, System: 16.8 ms]
  Range (min … max):    24.4 ms …  25.3 ms    110 runs

Summary
  git for-each-ref --format= | wc -l ran
    1.02 ± 0.01 times faster than git for-each-ref --format=x | wc -l

René