summaryrefslogtreecommitdiff
path: root/t/perf/p7519-fsmonitor.sh
AgeCommit message (Collapse)Author
2021-03-18fsmonitor: add perf test for git diff HEADNipunn Koorapati
Update the xargs call so that if your large repo contains symlinks, test-tool chmtime failure does not end the script. On Linux Test this tree upstream/master --------------------------------------------------------------------------------------------------------- 7519.4: status (fsmonitor=fsmonitor-watchman) 0.52(0.43+0.10) 0.53(0.49+0.05) +1.9% 7519.5: status -uno (fsmonitor=fsmonitor-watchman) 0.21(0.15+0.07) 0.22(0.13+0.09) +4.8% 7519.6: status -uall (fsmonitor=fsmonitor-watchman) 1.65(0.93+0.71) 1.69(1.03+0.65) +2.4% 7519.7: status (dirty) (fsmonitor=fsmonitor-watchman) 11.99(11.34+1.58) 11.95(11.02+1.79) -0.3% 7519.8: diff (fsmonitor=fsmonitor-watchman) 0.25(0.17+0.26) 0.25(0.18+0.26) +0.0% 7519.9: diff HEAD (fsmonitor=fsmonitor-watchman) 0.39(0.25+0.34) 0.89(0.35+0.74) +128.2% 7519.10: diff -- 0_files (fsmonitor=fsmonitor-watchman) 0.16(0.13+0.04) 0.16(0.12+0.05) +0.0% 7519.11: diff -- 10_files (fsmonitor=fsmonitor-watchman) 0.16(0.12+0.05) 0.16(0.12+0.05) +0.0% 7519.12: diff -- 100_files (fsmonitor=fsmonitor-watchman) 0.16(0.12+0.05) 0.16(0.12+0.05) +0.0% 7519.13: diff -- 1000_files (fsmonitor=fsmonitor-watchman) 0.16(0.11+0.06) 0.16(0.12+0.05) +0.0% 7519.14: diff -- 10000_files (fsmonitor=fsmonitor-watchman) 0.18(0.13+0.06) 0.17(0.10+0.08) -5.6% 7519.15: add (fsmonitor=fsmonitor-watchman) 2.25(1.53+0.68) 2.25(1.47+0.74) +0.0% 7519.18: status (fsmonitor=disabled) 0.88(0.73+1.03) 0.89(0.67+1.08) +1.1% 7519.19: status -uno (fsmonitor=disabled) 0.45(0.43+0.89) 0.45(0.34+0.98) +0.0% 7519.20: status -uall (fsmonitor=disabled) 1.88(1.16+1.58) 1.88(1.22+1.51) +0.0% 7519.21: status (dirty) (fsmonitor=disabled) 7.53(7.05+2.11) 7.53(6.98+2.04) +0.0% 7519.22: diff (fsmonitor=disabled) 0.42(0.37+0.92) 0.42(0.38+0.91) +0.0% 7519.23: diff HEAD (fsmonitor=disabled) 0.44(0.41+0.90) 0.44(0.40+0.91) +0.0% 7519.24: diff -- 0_files (fsmonitor=disabled) 0.13(0.09+0.05) 0.13(0.09+0.05) +0.0% 7519.25: diff -- 10_files (fsmonitor=disabled) 0.13(0.10+0.04) 0.13(0.10+0.04) +0.0% 7519.26: diff -- 100_files (fsmonitor=disabled) 0.13(0.09+0.05) 0.13(0.10+0.04) +0.0% 7519.27: diff -- 1000_files (fsmonitor=disabled) 0.13(0.09+0.06) 0.13(0.09+0.05) +0.0% 7519.28: diff -- 10000_files (fsmonitor=disabled) 0.14(0.11+0.05) 0.14(0.10+0.05) +0.0% 7519.29: add (fsmonitor=disabled) 2.43(1.61+1.64) 2.43(1.69+1.57) +0.0% On linux (2.29.2 vs w/ this patch): nipunn@nipunn-dbx:~/src/server3$ strace -f -c git diff 2>&1 | grep lstat 0.04 0.000063 3 20 6 lstat nipunn@nipunn-dbx:~/src/server3$ strace -f -c git diff HEAD 2>&1 | grep lstat 94.98 5.242262 10 523783 13 lstat nipunn@nipunn-dbx:~/src/server3$ strace -f -c ../git/bin-wrappers/git diff 2>&1 | grep lstat 0.38 0.000032 5 7 3 lstat nipunn@nipunn-dbx:~/src/server3$ strace -f -c ../git/bin-wrappers/git diff HEAD 2>&1 | grep lstat 99.44 0.741892 9 81634 10 lstat On mac (2.29.2 vs w/ this patch): nipunn-mbp:server nipunn$ sudo dtruss -L -f -c git diff 2>&1 | grep "^lstat64 " lstat64 8 nipunn-mbp:server nipunn$ sudo dtruss -L -f -c git diff HEAD 2>&1 | grep "^lstat64 " lstat64 120242 nipunn-mbp:server nipunn$ sudo dtruss -L -f -c ../git/bin-wrappers/git diff 2>&1 | grep "^lstat64 " lstat64 4 nipunn-mbp:server nipunn$ sudo dtruss -L -f -c ../git/bin-wrappers/git diff HEAD 2>&1 | grep "^lstat64 " lstat64 4497 There are still a bunch of lstats - on directories, but not every file. Progress! Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-03-01Merge branch 'jh/fsmonitor-prework'Junio C Hamano
Preliminary changes to fsmonitor integration. * jh/fsmonitor-prework: fsmonitor: refactor initialization of fsmonitor_last_update token fsmonitor: allow all entries for a folder to be invalidated fsmonitor: log FSMN token when reading and writing the index fsmonitor: log invocation of FSMonitor hook to trace2 read-cache: log the number of scanned files to trace2 read-cache: log the number of lstat calls to trace2 preload-index: log the number of lstat calls to trace2 p7519: add trace logging during perf test p7519: move watchman cleanup earlier in the test p7519: fix watchman watch-list test on Windows p7519: do not rely on "xargs -d" in test
2021-02-17p7519: add trace logging during perf testJeff Hostetler
Add optional trace logging to allow us to better compare performance of various fsmonitor providers and compare results with non-fsmonitor runs. Currently, this includes Trace2 logging, but may be extended to include other trace targets, such as GIT_TRACE_FSMONITOR if desired. Using this logging helped me explain an odd behavior on MacOS where the kernel was dropping events and causing the hook to Watchman to timeout. Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com> Reviewed-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-02-17p7519: move watchman cleanup earlier in the testJeff Hostetler
Shutdown Watchman after the Watchman-based tests and before the block of "no fsmonitor" tests. This helps ensure that Watchman cannot affect the test results for the other. Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com> Reviewed-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-02-17p7519: fix watchman watch-list test on WindowsJeff Hostetler
Only use the final portion of the test trash directory file name when verifying that Watchman was started. On Windows and under the SDK, $GIT_WORKTREE is a cygwin-style path with forward slashes and a "/c/" drive name. However `watchman watch-list` reports a proper Windows-style pathname with drive letters and backslashes. This causes the grep to fail. Since we don't really care about the full pathname (and we really don't want to bother with normalizaing them), just see if the test-name portion of the path is found. Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com> Reviewed-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-02-17p7519: do not rely on "xargs -d" in testJeff Hostetler
Convert the test to use a more portable method to update the mtime on a large number of files under version control. The Mac version of xargs does not support the "-d" option. Likewise, the "-0" and "--null" options are not portable. Furthermore, use `test-tool chmtime` rather than `touch` to update the mtime to ensure that it is actually updated (especially on file systems with only whole second resolution). Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com> Reviewed-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-01-06p7519: allow running without watchman prereqTaylor Blau
p7519 measures the performance of the fsmonitor code. To do this, it uses the installed copy of Watchman. If Watchman isn't installed, a noop integration script is installed in its place. When in the latter mode, it is expected that the script should not write a "last update token": in fact, it doesn't write anything at all since the script is blank. Commit 33226af42b (t/perf/fsmonitor: improve error message if typoing hook name, 2020-10-26) made sure that running 'git update-index --fsmonitor' did not write anything to stderr, but this is not the case when using the empty Watchman script, since Git will complain that: $ which watchman watchman not found $ cat .git/hooks/fsmonitor-empty $ git -c core.fsmonitor=.git/hooks/fsmonitor-empty update-index --fsmonitor warning: Empty last update token. Prior to 33226af42b, the output wasn't checked at all, which allowed this noop mode to work. But, 33226af42b breaks p7519 when running it without a 'watchman(1)' on your system. Handle this by only checking that the stderr is empty only when running with a real watchman executable. Otherwise, assert that the error message is the expected one when running in the noop mode. Signed-off-by: Taylor Blau <me@ttaylorr.com> Acked-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-11-30perf/fsmonitor: use test_must_be_empty helperNipunn Koorapati
Simplify test and make error messages more clear here. Per feedback from Junio in 33226af42b (t/perf/fsmonitor: improve error message if typoing hook name, 2020-10-26) Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Acked-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: add benchmark for dirty statusNipunn Koorapati
This benchmark covers the git status time for a heavily dirty directory - benchmarking fsmonitor's refresh When running to compare our perl vs rs-git-fsmonitor - we see that the perl script incurs significant overhead - further motivation to provide a faster implementation within git. 7519.7: status (dirty) (fsmonitor=query-watchman) 10.05(7.78+1.56) 7519.20: status (dirty) (fsmonitor=rs-git-fsmonitor) 6.72(4.37+1.64) 7519.33: status (dirty) (fsmonitor=disabled) 5.62(4.24+2.03) Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: perf comparison of multiple fsmonitor integrationsNipunn Koorapati
Allows for simple perf comparison of different integrations. I ran it to compare our perl script w/ rs-git-fsmonitor and found 20-30ms of overhead on every command. Output looks like this (extra newlines added for readability) Test this tree --------------------------------------------------------------------------- 7519.4: status (fsmonitor=query-watchman) 0.42(0.37+0.05) 7519.5: status -uno (fsmonitor=query-watchman) 0.19(0.12+0.07) 7519.6: status -uall (fsmonitor=query-watchman) 1.36(0.73+0.62) 7519.7: diff (fsmonitor=query-watchman) 0.14(0.09+0.05) 7519.8: diff -- 0_files (fsmonitor=query-watchman) 0.14(0.11+0.03) 7519.9: diff -- 10_files (fsmonitor=query-watchman) 0.14(0.10+0.04) 7519.10: diff -- 100_files (fsmonitor=query-watchman) 0.14(0.09+0.05) 7519.11: diff -- 1000_files (fsmonitor=query-watchman) 0.14(0.08+0.06) 7519.12: diff -- 10000_files (fsmonitor=query-watchman) 0.14(0.09+0.05) 7519.13: add (fsmonitor=query-watchman) 2.04(1.32+0.66) 7519.16: status (fsmonitor=rs-git-fsmonitor) 0.39(0.32+0.08) 7519.17: status -uno (fsmonitor=rs-git-fsmonitor) 0.17(0.11+0.06) 7519.18: status -uall (fsmonitor=rs-git-fsmonitor) 1.33(0.71+0.61) 7519.19: diff (fsmonitor=rs-git-fsmonitor) 0.11(0.07+0.04) 7519.20: diff -- 0_files (fsmonitor=rs-git-fsmonitor) 0.11(0.09+0.03) 7519.21: diff -- 10_files (fsmonitor=rs-git-fsmonitor) 0.11(0.09+0.03) 7519.22: diff -- 100_files (fsmonitor=rs-git-fsmonitor) 0.11(0.07+0.04) 7519.23: diff -- 1000_files (fsmonitor=rs-git-fsmonitor) 0.11(0.06+0.06) 7519.24: diff -- 10000_files (fsmonitor=rs-git-fsmonitor) 0.11(0.06+0.06) 7519.25: add (fsmonitor=rs-git-fsmonitor) 2.03(1.28+0.69) 7519.28: status (fsmonitor=disabled) 0.77(0.59+0.99) 7519.29: status -uno (fsmonitor=disabled) 0.42(0.33+0.85) 7519.30: status -uall (fsmonitor=disabled) 1.59(1.02+1.34) 7519.31: diff (fsmonitor=disabled) 0.35(0.30+0.81) 7519.32: diff -- 0_files (fsmonitor=disabled) 0.11(0.08+0.04) 7519.33: diff -- 10_files (fsmonitor=disabled) 0.11(0.07+0.04) 7519.34: diff -- 100_files (fsmonitor=disabled) 0.11(0.08+0.03) 7519.35: diff -- 1000_files (fsmonitor=disabled) 0.11(0.10+0.02) 7519.36: diff -- 10000_files (fsmonitor=disabled) 0.12(0.07+0.06) 7519.37: add (fsmonitor=disabled) 2.24(1.48+1.44) Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: initialize test with git resetNipunn Koorapati
Previously, the git add of the previous suiterun would pollute the numbers in the second run Before: Test this tree ----------------------------------------------------------------------------- 7519.4: status (fsmonitor=fsmonitor-watchman) 0.40(0.36+0.04) 7519.5: status -uno (fsmonitor=fsmonitor-watchman) 0.19(0.12+0.07) 7519.6: status -uall (fsmonitor=fsmonitor-watchman) 1.36(0.74+0.61) 7519.7: diff (fsmonitor=fsmonitor-watchman) 0.14(0.10+0.04) 7519.8: diff -- 0_files (fsmonitor=fsmonitor-watchman) 0.14(0.10+0.04) 7519.9: diff -- 10_files (fsmonitor=fsmonitor-watchman) 0.14(0.09+0.05) 7519.10: diff -- 100_files (fsmonitor=fsmonitor-watchman) 0.14(0.10+0.04) 7519.11: diff -- 1000_files (fsmonitor=fsmonitor-watchman) 0.14(0.08+0.06) 7519.12: diff -- 10000_files (fsmonitor=fsmonitor-watchman) 0.14(0.10+0.04) 7519.13: add (fsmonitor=fsmonitor-watchman) 2.03(1.28+0.69) 7519.16: status (fsmonitor=disabled) 0.64(0.49+0.90) 7519.17: status -uno (fsmonitor=disabled) 1.15(0.92+1.00) 7519.18: status -uall (fsmonitor=disabled) 2.32(1.46+1.55) 7519.19: diff (fsmonitor=disabled) 1.44(1.12+1.76) 7519.20: diff -- 0_files (fsmonitor=disabled) 0.11(0.07+0.05) 7519.21: diff -- 10_files (fsmonitor=disabled) 0.11(0.06+0.05) 7519.22: diff -- 100_files (fsmonitor=disabled) 0.11(0.08+0.03) 7519.23: diff -- 1000_files (fsmonitor=disabled) 0.11(0.08+0.04) 7519.24: diff -- 10000_files (fsmonitor=disabled) 0.12(0.06+0.07) 7519.25: add (fsmonitor=disabled) 2.25(1.47+1.47) After: Test this tree ----------------------------------------------------------------------------- 7519.4: status (fsmonitor=fsmonitor-watchman) 0.41(0.33+0.09) 7519.5: status -uno (fsmonitor=fsmonitor-watchman) 0.20(0.14+0.07) 7519.6: status -uall (fsmonitor=fsmonitor-watchman) 1.37(0.78+0.58) 7519.7: diff (fsmonitor=fsmonitor-watchman) 0.14(0.10+0.04) 7519.8: diff -- 0_files (fsmonitor=fsmonitor-watchman) 0.14(0.08+0.06) 7519.9: diff -- 10_files (fsmonitor=fsmonitor-watchman) 0.14(0.09+0.05) 7519.10: diff -- 100_files (fsmonitor=fsmonitor-watchman) 0.14(0.10+0.05) 7519.11: diff -- 1000_files (fsmonitor=fsmonitor-watchman) 0.14(0.11+0.04) 7519.12: diff -- 10000_files (fsmonitor=fsmonitor-watchman) 0.14(0.09+0.05) 7519.13: add (fsmonitor=fsmonitor-watchman) 2.04(1.27+0.71) 7519.16: status (fsmonitor=disabled) 0.78(0.59+0.99) 7519.17: status -uno (fsmonitor=disabled) 0.43(0.32+0.88) 7519.18: status -uall (fsmonitor=disabled) 1.58(0.96+1.38) 7519.19: diff (fsmonitor=disabled) 0.36(0.31+0.79) 7519.20: diff -- 0_files (fsmonitor=disabled) 0.11(0.08+0.03) 7519.21: diff -- 10_files (fsmonitor=disabled) 0.11(0.07+0.04) 7519.22: diff -- 100_files (fsmonitor=disabled) 0.11(0.08+0.04) 7519.23: diff -- 1000_files (fsmonitor=disabled) 0.11(0.07+0.05) 7519.24: diff -- 10000_files (fsmonitor=disabled) 0.12(0.08+0.05) 7519.25: add (fsmonitor=disabled) 2.25(1.48+1.47) Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: factor setup for fsmonitor into functionNipunn Koorapati
This prepares for it being called multiple times when testing different hooks Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: silence initial git commitNipunn Koorapati
It is extremely verbose, printing >10K non-useful lines Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: shorten DESC to basenameNipunn Koorapati
The full name is lengthy and makes it hard to read Before: 7519.3: status (fsmonitor=/home/nipunn/src/server/.git/hooks/rs-git-fsmonitor) 0.02(0.01+0.00) After 7519.3: status (fsmonitor=rs-git-fsmonitor) 0.03(0.02+0.00) Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: factor description out for readabilityNipunn Koorapati
There was much duplication here. Prepares for making changes to the description. Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: improve error message if typoing hook nameNipunn Koorapati
Previously - it would silently run the perf suite w/o using fsmonitor - fsmonitor errors are not hard failures. Now it errors loudly. GIT_PERF_7519_FSMONITOR="$HOME/rs-git-fsmonitorr" ./p7519-fsmonitor.sh -i -v fatal: cannot run /home/nipunn/rs-git-fsmonitorr: No such file or directory not ok 2 - setup for fsmonitor Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: move watchman setup to one-time-repo-setupNipunn Koorapati
It is only required to be set up once. This prepares for testing multiple hooks in one invocation. Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-26t/perf/fsmonitor: separate one time repo initializationNipunn Koorapati
In preparation for testing multiple fsmonitor hooks Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-20p7519-fsmonitor: add a git add benchmarkNipunn Koorapati
Test v2.29.0-rc1 this tree ----------------------------------------------------------------------------------------------------------------- 7519.2: status (fsmonitor=.git/hooks/fsmonitor-watchman) 1.48(0.79+0.67) 1.48(0.79+0.67) +0.0% 7519.3: status -uno (fsmonitor=.git/hooks/fsmonitor-watchman) 0.16(0.11+0.05) 0.17(0.13+0.04) +6.3% 7519.4: status -uall (fsmonitor=.git/hooks/fsmonitor-watchman) 1.36(0.77+0.58) 1.37(0.72+0.63) +0.7% 7519.5: diff (fsmonitor=.git/hooks/fsmonitor-watchman) 0.84(0.21+0.63) 0.14(0.11+0.03) -83.3% 7519.6: diff -- 0_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.07+0.05) 0.13(0.09+0.04) +8.3% 7519.7: diff -- 10_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.09+0.04) 0.13(0.07+0.06) +8.3% 7519.8: diff -- 100_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.08+0.05) 0.12(0.08+0.05) +0.0% 7519.9: diff -- 1000_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.08+0.05) 0.13(0.09+0.04) +8.3% 7519.10: diff -- 10000_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.14(0.08+0.06) 0.13(0.07+0.06) -7.1% 7519.11: add (fsmonitor=.git/hooks/fsmonitor-watchman) 2.75(1.41+1.27) 2.03(1.26+0.70) -26.2% 7519.13: status (fsmonitor=) 1.38(1.03+1.04) 1.37(1.04+1.04) -0.7% 7519.14: status -uno (fsmonitor=) 1.11(0.83+0.98) 1.10(0.89+0.90) -0.9% 7519.15: status -uall (fsmonitor=) 2.30(1.57+1.42) 2.31(1.49+1.50) +0.4% 7519.16: diff (fsmonitor=) 1.43(1.13+1.76) 1.46(1.19+1.72) +2.1% 7519.17: diff -- 0_files (fsmonitor=) 0.10(0.08+0.04) 0.11(0.08+0.04) +10.0% 7519.18: diff -- 10_files (fsmonitor=) 0.10(0.07+0.05) 0.11(0.08+0.04) +10.0% 7519.19: diff -- 100_files (fsmonitor=) 0.10(0.07+0.04) 0.11(0.07+0.05) +10.0% 7519.20: diff -- 1000_files (fsmonitor=) 0.10(0.08+0.03) 0.11(0.08+0.04) +10.0% 7519.21: diff -- 10000_files (fsmonitor=) 0.11(0.08+0.05) 0.12(0.07+0.06) +9.1% 7519.22: add (fsmonitor=) 2.26(1.46+1.49) 2.27(1.42+1.55) +0.4% Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-20p7519-fsmonitor: refactor to avoid code duplicationNipunn Koorapati
Much of the benchmark code is redundant. This is easier to understand and edit. Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-20t/perf: add fsmonitor perf test for git diffNipunn Koorapati
Results for the git-diff fsmonitor optimization in patch in the parent-rev (using a 400k file repo to test) As you can see here - git diff with fsmonitor running is significantly better with this patch series (80% faster on my workload)! GIT_PERF_LARGE_REPO=~/src/server ./run v2.29.0-rc1 . -- p7519-fsmonitor.sh Test v2.29.0-rc1 this tree ----------------------------------------------------------------------------------------------------------------- 7519.2: status (fsmonitor=.git/hooks/fsmonitor-watchman) 1.46(0.82+0.64) 1.47(0.83+0.62) +0.7% 7519.3: status -uno (fsmonitor=.git/hooks/fsmonitor-watchman) 0.16(0.12+0.04) 0.17(0.12+0.05) +6.3% 7519.4: status -uall (fsmonitor=.git/hooks/fsmonitor-watchman) 1.36(0.73+0.62) 1.37(0.76+0.60) +0.7% 7519.5: diff (fsmonitor=.git/hooks/fsmonitor-watchman) 0.85(0.22+0.63) 0.14(0.10+0.05) -83.5% 7519.6: diff -- 0_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.08+0.05) 0.13(0.11+0.02) +8.3% 7519.7: diff -- 10_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.08+0.04) 0.13(0.09+0.04) +8.3% 7519.8: diff -- 100_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.07+0.05) 0.13(0.07+0.06) +8.3% 7519.9: diff -- 1000_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.12(0.09+0.04) 0.13(0.08+0.05) +8.3% 7519.10: diff -- 10000_files (fsmonitor=.git/hooks/fsmonitor-watchman) 0.14(0.09+0.05) 0.13(0.10+0.03) -7.1% 7519.12: status (fsmonitor=) 1.67(0.93+1.49) 1.67(0.99+1.42) +0.0% 7519.13: status -uno (fsmonitor=) 0.37(0.30+0.82) 0.37(0.33+0.79) +0.0% 7519.14: status -uall (fsmonitor=) 1.58(0.97+1.35) 1.57(0.86+1.45) -0.6% 7519.15: diff (fsmonitor=) 0.34(0.28+0.83) 0.34(0.27+0.83) +0.0% 7519.16: diff -- 0_files (fsmonitor=) 0.09(0.06+0.04) 0.09(0.08+0.02) +0.0% 7519.17: diff -- 10_files (fsmonitor=) 0.09(0.07+0.03) 0.09(0.06+0.05) +0.0% 7519.18: diff -- 100_files (fsmonitor=) 0.09(0.06+0.04) 0.09(0.06+0.04) +0.0% 7519.19: diff -- 1000_files (fsmonitor=) 0.09(0.06+0.04) 0.09(0.05+0.05) +0.0% 7519.20: diff -- 10000_files (fsmonitor=) 0.10(0.08+0.04) 0.10(0.06+0.05) +0.0% I also added a benchmark for a tiny git diff workload w/ a pathspec. I see an approximately .02 second overhead added w/ and w/o fsmonitor From looking at these results, I suspected that refresh_fsmonitor is already happening during git diff - independent of this patch series' optimization. Confirmed that suspicion by breaking on refresh_fsmonitor. (gdb) bt [simplified] 0 refresh_fsmonitor at fsmonitor.c:176 1 ie_match_stat at read-cache.c:375 2 match_stat_with_submodule at diff-lib.c:237 4 builtin_diff_files at builtin/diff.c:260 5 cmd_diff at builtin/diff.c:541 6 run_builtin at git.c:450 7 handle_builtin at git.c:700 8 run_argv at git.c:767 9 cmd_main at git.c:898 10 main at common-main.c:52 Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-20t/perf/p7519-fsmonitor.sh: warm cache on first git statusNipunn Koorapati
The first git status would be inflated due to warming of filesystem cache. This makes the results comparable. Before Test this tree -------------------------------------------------------------------------------- 7519.2: status (fsmonitor=.git/hooks/fsmonitor-watchman) 2.52(1.59+1.56) 7519.3: status -uno (fsmonitor=.git/hooks/fsmonitor-watchman) 0.18(0.12+0.06) 7519.4: status -uall (fsmonitor=.git/hooks/fsmonitor-watchman) 1.36(0.73+0.62) 7519.7: status (fsmonitor=) 0.69(0.52+0.90) 7519.8: status -uno (fsmonitor=) 0.37(0.28+0.81) 7519.9: status -uall (fsmonitor=) 1.53(0.93+1.32) After Test this tree -------------------------------------------------------------------------------- 7519.2: status (fsmonitor=.git/hooks/fsmonitor-watchman) 0.39(0.33+0.06) 7519.3: status -uno (fsmonitor=.git/hooks/fsmonitor-watchman) 0.17(0.13+0.05) 7519.4: status -uall (fsmonitor=.git/hooks/fsmonitor-watchman) 1.34(0.77+0.56) 7519.7: status (fsmonitor=) 0.70(0.53+0.90) 7519.8: status -uno (fsmonitor=) 0.37(0.32+0.78) 7519.9: status -uall (fsmonitor=) 1.55(1.01+1.25) Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-03-27t/helper: merge test-drop-caches into test-toolNguyễn Thái Ngọc Duy
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-12-18p7519: improve check for prerequisite WATCHMANRené Scharfe
The return code of command -v with a non-existing command is 1 in bash and 127 in dash. Use that return code directly to allow the script to work with dash and without watchman (e.g. on Debian). While at it stop redirecting the output. stderr is redirected to /dev/null by test_lazy_prereq already, and stdout can actually be useful -- the path of the found watchman executable is sent there, but it's shown only if the script was run with --verbose. Signed-off-by: Rene Scharfe <l.s.r@web.de> Acked-by: Ben Peart <benpeart@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-10-01fsmonitor: add a performance testBen Peart
Add a test utility (test-drop-caches) that flushes all changes to disk then drops file system cache on Windows, Linux, and OSX. Add a perf test (p7519-fsmonitor.sh) for fsmonitor. By default, the performance test will utilize the Watchman file system monitor if it is installed. If Watchman is not installed, it will use a dummy integration script that does not report any new or modified files. The dummy script has very little overhead which provides optimistic results. The performance test will also use the untracked cache feature if it is available as fsmonitor uses it to speed up scanning for untracked files. There are 4 environment variables that can be used to alter the default behavior of the performance test: GIT_PERF_7519_UNTRACKED_CACHE: used to configure core.untrackedCache GIT_PERF_7519_SPLIT_INDEX: used to configure core.splitIndex GIT_PERF_7519_FSMONITOR: used to configure core.fsmonitor GIT_PERF_7519_DROP_CACHE: if set, the OS caches are dropped between tests The big win for using fsmonitor is the elimination of the need to scan the working directory looking for changed and untracked files. If the file information is all cached in RAM, the benefits are reduced. Signed-off-by: Ben Peart <benpeart@microsoft.com> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>