"git pack-objects" learns to find delta bases from blobs at the
same path, using the --path-walk API.
* ds/path-walk-2:
pack-objects: allow --shallow and --path-walk
path-walk: add new 'edge_aggressive' option
pack-objects: thread the path-based compression
pack-objects: refactor path-walk delta phase
scalar: enable path-walk during push via config
pack-objects: enable --path-walk via config
repack: add --path-walk option
t5538: add tests to confirm deltas in shallow pushes
pack-objects: introduce GIT_TEST_PACK_PATH_WALK
p5313: add performance tests for --path-walk
pack-objects: update usage to match docs
pack-objects: add --path-walk option
pack-objects: extract should_attempt_deltas()
Remove the leftover hints to the test framework to mark tests that
do not pass the leak checker tests, as they should no longer be
needed.
* kn/passing-leak-tests:
t: remove unexpected SANITIZE_LEAK variables
"git apply" and "git add -i/-p" code paths no longer unnecessarily
expand sparse-index while working.
* ds/sparse-apply-add-p:
p2000: add performance test for patch-mode commands
reset: integrate sparse index with --patch
git add: make -p/-i aware of sparse index
apply: integrate with the sparse index
As of 1fc7ddf35b (test-lib: unconditionally enable leak checking,
2024-11-20), both the `GIT_TEST_PASSING_SANITIZE_LEAK` and
`TEST_PASSES_SANITIZE_LEAK` variables no longer have any meaning, the
leak checks are enabled by default. However, some newly added tests
include them by mistake. Let's clean this up.
Signed-off-by: Karthik Nayak <karthik.188@gmail.com>
Acked-by: Justin Tobler <jltobler@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 'git pack-objects' supports a --path-walk option, allow passing it
through in 'git repack'. This presents interesting testing opportunities for
comparing the different repacking strategies against each other.
Add the --path-walk option to the performance tests in p5313.
For the microsoft/fluentui repo [1] checked out at a specific commit [2],
the --path-walk tests in p5313 look like this:
Test this tree
-------------------------------------------------------------------------
5313.18: thin pack with --path-walk 0.08(0.06+0.02)
5313.19: thin pack size with --path-walk 18.4K
5313.20: big pack with --path-walk 2.10(7.80+0.26)
5313.21: big pack size with --path-walk 19.8M
5313.22: shallow fetch pack with --path-walk 1.62(3.38+0.17)
5313.23: shallow pack size with --path-walk 33.6M
5313.24: repack with --path-walk 81.29(96.08+0.71)
5313.25: repack size with --path-walk 142.5M
[1] https://github.com/microsoft/fluentui
[2] e70848ebac1cd720875bccaa3026f4a9ed700e08
Along with the earlier tests in p5313, I'll instead reformat the
comparison as follows:
Repack Method Pack Size Time
---------------------------------------
Hash v1 439.4M 87.24s
Hash v2 161.7M 21.51s
Path Walk 142.5M 81.29s
There are a few things to notice here:
1. The benefits of --name-hash-version=2 over --name-hash-version=1 are
significant, but --path-walk still compresses better than that
option.
2. The --path-walk command is still using --name-hash-version=1 for the
second pass of delta computation, using the increased name hash
collisions as a potential method for opportunistic compression on
top of the path-focused compression.
3. The --path-walk algorithm is currently sequential and does not use
multiple threads for delta compression. Threading will be
implemented in a future change so the computation time will improve
to better compete in this metric.
There are small benefits in size for my copy of the Git repository:
Repack Method Pack Size Time
---------------------------------------
Hash v1 248.8M 30.44s
Hash v2 249.0M 30.15s
Path Walk 213.2M 142.50s
As well as in the nodejs/node repository [3]:
Repack Method Pack Size Time
---------------------------------------
Hash v1 739.9M 71.18s
Hash v2 764.6M 67.82s
Path Walk 698.1M 208.10s
[3] https://github.com/nodejs/node
This benefit also repeats in my copy of the Linux kernel repository:
Repack Method Pack Size Time
---------------------------------------
Hash v1 2.5G 554.41s
Hash v2 2.5G 549.62s
Path Walk 2.2G 1562.36s
It is important to see that even when the repository shape does not have
many name-hash collisions, there is a slight space boost to be found
using this method.
As this repacking strategy was released in Git for Windows 2.47.0, some
users have reported cases where the --path-walk compression is slightly
worse than the --name-hash-version=2 option. In those cases, it may be
beneficial to combine the two options. However, there has not been a
released version of Git that has both options and I don't have access to
these repos for testing.
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous change added a --path-walk option to 'git pack-objects'.
Create a performance test that demonstrates the time and space benefits
of the feature.
In order to get an appropriate comparison, we need to avoid reusing
deltas and recompute them from scratch.
Compare the creation of a thin pack representing a small push and the
creation of a relatively large non-thin pack.
Running on my copy of the Git repository results in this data (removing
the repack tests for --name-hash-version):
Test this tree
------------------------------------------------------------------------
5313.2: thin pack with --name-hash-version=1 0.02(0.01+0.01)
5313.3: thin pack size with --name-hash-version=1 1.6K
5313.4: big pack with --name-hash-version=1 2.55(4.20+0.26)
5313.5: big pack size with --name-hash-version=1 16.4M
5313.6: shallow fetch pack with --name-hash-version=1 1.24(2.03+0.08)
5313.7: shallow pack size with --name-hash-version=1 12.2M
5313.10: thin pack with --name-hash-version=2 0.03(0.01+0.01)
5313.11: thin pack size with --name-hash-version=2 1.6K
5313.12: big pack with --name-hash-version=2 1.91(3.23+0.20)
5313.13: big pack size with --name-hash-version=2 16.4M
5313.14: shallow fetch pack with --name-hash-version=2 1.06(1.57+0.10)
5313.15: shallow pack size with --name-hash-version=2 12.5M
5313.18: thin pack with --path-walk 0.03(0.01+0.01)
5313.19: thin pack size with --path-walk 1.6K
5313.20: big pack with --path-walk 2.05(3.24+0.27)
5313.21: big pack size with --path-walk 16.3M
5313.22: shallow fetch pack with --path-walk 1.08(1.66+0.07)
5313.23: shallow pack size with --path-walk 12.4M
This can be reformatted as follows:
Pack Type Hash v1 Hash v2 Path Walk
---------------------------------------------------
thin pack (time) 0.02s 0.03s 0.03s
(size) 1.6K 1.6K 1.6K
big pack (time) 2.55s 1.91s 2.05s
(size) 16.4M 16.4M 16.3M
shallow pack (time) 1.24s 1.06s 1.08s
(size) 12.2M 12.5M 12.4M
Note that the timing is slower because there is no threading in the
--path-walk case (yet). Also, the shallow pack cases are really not
using the --path-walk logic right now because it is disabled until some
additions are made to the path walk API.
The cases where the --path-walk option really shines is when the default
name-hash is overwhelmed with unhelpful collisions. An open source
example can be found in the microsoft/fluentui repo [1] at a certain
commit [2].
[1] https://github.com/microsoft/fluentui
[2] e70848ebac1cd720875bccaa3026f4a9ed700e08
Running the tests on this repo results in the following comparison table:
Pack Type Hash v1 Hash v2 Path Walk
---------------------------------------------------
thin pack (time) 0.36s 0.12s 0.08s
(size) 1.2M 22.0K 18.4K
big pack (time) 2.00s 2.90s 2.21s
(size) 20.4M 25.9M 19.5M
shallow pack (time) 1.41s 1.80s 1.65s
(size) 34.4M 33.7M 33.6M
Notice in particular that in the small thin pack, the time performance
has improved from 0.36s for --name-hash-version=1 to 0.08s and this is
likely due to the improved size of the resulting pack: 18.4K instead of
1.2M. The relatively new --name-hash-version=2 is competitive with
--path-walk (0.12s and 22.0K) but not quite as successful.
Finally, running this on a copy of the Linux kernel repository results
in these data points:
Pack Type Hash v1 Hash v2 Path Walk
---------------------------------------------------
thin pack (time) 0.03s 0.13s 0.03s
(size) 4.6K 4.6K 4.6K
big pack (time) 15.29s 12.32s 13.92s
(size) 201.1M 159.1M 158.5M
shallow pack (time) 10.88s 22.93s 22.74s
(size) 269.2M 273.8M 267.7M
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous three changes contributed performance improvements to 'git
apply', 'git add -p', and 'git reset -p' when using a sparse index. The
improvement to 'git apply' also improved 'git checkout -p'. Add
performance tests to demonstrate this (and to help validate that
performance remains good in the future).
In the truncated test output below, we see that the full checkout
performance changes within noise expectations, but the sparse index
cases improve 33% and then 96% for 'git add -p' and 41% and then 95% for
'git reset -p'. 'git checkout -p' improves immediatley by 91% because it
does not need any change to its builtin.
Test HEAD~4 HEAD~3 HEAD~2 HEAD~1
-------------------------------------------------------------------------------------
2000.118: ... git add -p (full-v3) 0.79 0.79 +0.0% 0.82 +3.8% 0.82 +3.8%
2000.119: ... git add -p (full-v4) 0.74 0.76 +2.7% 0.74 +0.0% 0.76 +2.7%
2000.120: ... git add -p (sparse-v3) 1.94 1.28 -34.0% 0.07 -96.4% 0.07 -96.4%
2000.121: ... git add -p (sparse-v4) 1.93 1.28 -33.7% 0.06 -96.9% 0.06 -96.9%
2000.122: ... git checkout -p (full-v3) 1.18 1.18 +0.0% 1.18 +0.0% 1.19 +0.8%
2000.123: ... git checkout -p (full-v4) 1.10 1.12 +1.8% 1.11 +0.9% 1.11 +0.9%
2000.124: ... git checkout -p (sparse-v3) 1.31 0.11 -91.6% 0.11 -91.6% 0.11 -91.6%
2000.125: ... git checkout -p (sparse-v4) 1.29 0.11 -91.5% 0.11 -91.5% 0.11 -91.5%
2000.126: ... git reset -p (full-v3) 0.81 0.80 -1.2% 0.83 +2.5% 0.83 +2.5%
2000.127: ... git reset -p (full-v4) 0.78 0.77 -1.3% 0.77 -1.3% 0.78 +0.0%
2000.128: ... git reset -p (sparse-v3) 1.58 0.92 -41.8% 0.91 -42.4% 0.07 -95.6%
2000.129: ... git reset -p (sparse-v4) 1.58 0.92 -41.8% 0.92 -41.8% 0.07 -95.6%
It is worth noting that if our test was more involved and had multiple
hunks to evaluate, then the time spent in 'git apply' would dominate due
to multiple index loads and writes. As it stands, we need the sparse
index improvement in 'git add -p' itself to confirm this performance
improvement.
Since the change for 'git add -i' is identical, we avoid a second test
case for that similar operation.
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The build procedure based on Meson learned to drive the
benchmarking tests.
* ps/meson-build-perf-bench:
meson: wire up benchmarking options
meson: wire up benchmarks
t/perf: fix benchmarks with out-of-tree builds
t/perf: use configured PERL_PATH
t/perf: fix benchmarks with alternate repo formats
The "perf-lib.sh" script is sourced by all of our benchmarking suites to
make available common infrastructure. The script assumes that build and
source directory are the same, which works for our Makefile. But the
assumption breaks with both CMake and Meson, where the build directory
can be located in an arbitrary place.
Adapt the script so that it works with out-of-tree builds. Most
importantly, this requires us to figure out the location of the build
directory:
- When running benchmarks via our Makefile the build directory is the
same as the source directory. We already know to derive the test
directory ("t/") via `$(pwd)/..`, which works because we chdir into
"t/perf" before executing benchmarks. We can thus derive the build
directory by appending another "/.." to that path.
- When running benchmarks via Meson the build directory is located at
an arbitrary location. The build system thus has to make the path
known by exporting the `GIT_BUILD_DIR` environment variable.
This change prepares us for wiring up benchmarks in Meson.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our benchmarks use a couple of Perl scripts to compute results. These
Perl scripts get executed directly, and as the shebang is hardcoded to
"/usr/bin/perl" this will fail on any system where the Perl interpreter
is located in a different path.
Our build infrastructure already lets users configure the location of
Perl, which ultimately gets written into the GIT-BUILD-OPTIONS file.
This file is being sourced by "test-lib.sh", and consequently we already
have the "PERL_PATH" variable available that contains its configured
location.
Use "PERL_PATH" to execute Perl scripts, which makes them work on more
esoteric systems like NixOS. Furthermore, adapt the shebang to use
env(1) to execute Perl so that users who have Perl in PATH, but in a
non-standard location can execute the script directly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Many of our benchmarks operate on a user-defined repository that we copy
over before running the benchmarked logic. To keep unintentional side
effects caused by on-disk state at bay we skip copying some files. This
includes for example hooks, but also the repo's configuration.
It is quite sensible to not copy over the configuration, as it is quite
easy to inadvertently carry over configuration that may significantly
impact the performance measurements. But we cannot fully ignore the
configuration either, as it may contain information about the repository
format. This will cause failures when for example using a repository
with SHA256 object format or the reftable ref format.
Fix the issue by parsing the reference and object formats from the
source repository and passing them to git-init(1).
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This perf script creates a midx by running "git multi-pack-index write"
with the "--stdin-packs" option. We feed that stdin by running "find" on
.git/objects/pack, using sed to strip off everything but the basename.
But that sed invocation also does something peculiar: it adds a "+" to
the start of each pack name. This causes the multi-pack-index command to
barf. The modified name does not match any pack it knows about, so it
ends up with an empty list of packs to put in the midx. And thus nothing
matches the --preferred-pack option we pass, which causes it die().
The fix is to remove the extra "+" (which also lets us simplify the sed
invocation a bit, as it is now just stripping the leading directories).
But that leaves the mystery of why it was ever there in the first place.
The answer is that an earlier iteration of the patch series had a
concept of "disjoint" packs in the midx. And one of its patches here:
https://lore.kernel.org/git/c52d7e7b27a9add4f58b8334db4fe4498af1c90f.1701198172.git.me@ttaylorr.com/
taught read_packs_from_stdin() to treat a leading "+" as marking a
disjoint pack. But in the second version of the series, which was
ultimately merged, that disjoint concept went away, and the code to
parse "+" did likewise. The regular regression tests were adjusted to
match, but this case in t/perf was forgotten.
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A common way to run Git's performance benchmarks on repositories other
than Git's own repository (which is not exactly large when compared to
actually large repositories) is to run them like this:
GIT_PERF_LARGE_REPO=/path/to/my/large/repo \
./p1234-*.sh -ivx
Contrary to developers' common expectations, this failed to work when
Git was built with a different `GIT_PERF_LARGE_REPO` value specified at
build time: That build-time option would have been written to the
`GIT-BUILD-OPTIONS` file, which in turn would have been sourced by
`test-lib.sh`, which in turn would have been sourced by `perf-lib.sh`,
which in turn would have been sourced by the perf test script,
_overriding_ the environment variable specified in the way illustrated
above.
Since perf tests are not run as part of the build, this most likely
unintended behavior was not caught and certainly not fixed, as the
`GIT_PERF_*` values would have been empty at build-time.
However, in 4638e8806e (Makefile: use common template for
GIT-BUILD-OPTIONS, 2024-12-06), a subtle change of behavior was
introduced: Whereas before, a couple of build-time options (the
`GIT_PERF_*` ones included) were written to `GIT-BUILD-OPTIONS` only
when their values were non-empty. With this commit, they are also
written when they are empty.
The consequence is that above-mentioned way to run the perf tests will
not only fail to pick up the desired `GIT_PERF_*` settings when they
were specified differently while building Git, instead the desired
settings will be only respected when specified _while building_ Git.
Let's work around the original issue, i.e. let `GIT_PERF_*` environment
variables override what is recorded in `GIT-BUILD-OPTIONS`.
Note that this is just the tip of the iceberg, there are a couple of
`GIT_TEST_*` options that may want a similar fix in `test-lib.sh`. Due
to time constraints on my side, this here patch focuses exclusively on
the `GIT_PERF_*` settings.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 7b31b55db1 (perf: amend the grep tests to test grep.threads,
2017-12-29), p7821 was tweaked to test the performance of 'git grep'
under different number of threads. These tests are run if
GIT_PERF_GREP_THREADS is set to a list of thread numbers, but the
comment at the top of the file instead mentions GIT_PERF_7821_THREADS.
Fix the comment.
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In p9210-scalar-clone.sh, we test using 'scalar clone' to clone
$GIT_PERF_LARGE_REPO (copied locally as 'to-clone'), which defaults to
the git.git checkout we are running the test from.
When --branch is not specified (as in this test), 'scalar clone' tries
to get the default branch of the remote repository by parsing the output
of 'git ls-remote --symref $URL HEAD', as implemented in
scalar.c:remote_default_branch. When the git.git checkout we are running
the test from is in detached HEAD, this fails and we fall back to using
the name of the currently checked out branch in the newly initialized
repository, which in this case is the value returned earlier in
cmd_clone by repo_default_branch_name.
We then invoke 'git checkout -t origin/$branch', with $branch being the
name we got from remote_default_branch. This invocation fails if
'$branch' does not exist as a branch in the current git.git checkout.
Fix this by creating a local branch in 'to-clone' in the setup test
"enable server-side partial clone", making sure to use '-B' in case a
branch named 'test-branch' already exists.
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 5dccd9155f (t/perf: add iteration setup mechanism to perf-lib,
2022-04-04), perf tests need to declare their prerequisites with
'--prereq', after the test title. p7821 was forgotten in that commit,
such that running that test on a machine where the PCRE prereq is not
satisfied aborts the test with:
error: bug in the test script: test_wrapper_ needs 2 positional parameters
Fix this by correcting the two 'test_perf' invocations in that test
suite.
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Correct the default target in Documentation/Makefile, and
future-proof all Makefiles from similar breakages by declaring the
default target (which happens to be "all") upfront.
* ad/set-default-target-in-makefiles:
Makefile: set default goals in makefiles
Explicitly set the default goal at the very top of various makefiles.
This is already present in some makefiles, but not all of them.
In particular, this corrects a regression introduced in a38edab7c8
(Makefile: generate doc versions via GIT-VERSION-GEN, 2024-12-06). That
commit added some config files as build targets for the Documentation
directory, and put the target configuration in a sensible place.
Unfortunately, that sensible place was above any other build target
definitions, meaning the default goal changed to being those
configuration files only, rather than the HTML and man page
documentation.
Signed-off-by: Adam Dinwoodie <adam@dinwoodie.org>
Helped-by: Junio C Hamano <gitster@pobox.com>
Acked-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git pack-objects" and its wrapper "git repack" learned an option
to use an alternative path-hash function to improve delta-base
selection to produce a packfile with deeper history than window
size.
* ds/name-hash-tweaks:
pack-objects: prevent name hash version change
test-tool: add helper for name-hash values
p5313: add size comparison test
pack-objects: add GIT_TEST_NAME_HASH_VERSION
repack: add --name-hash-version option
pack-objects: add --name-hash-version option
pack-objects: create new name-hash function version
Add a new test-tool helper, name-hash, to output the value of the
name-hash algorithms for the input list of strings, one per line.
Since the name-hash values can be stored in the .bitmap files, it is
important that these hash functions do not change across Git versions.
Add a simple test to t5310-pack-bitmaps.sh to provide some testing of
the current values. Due to how these functions are implemented, it would
be difficult to change them without disturbing these values. The paths
used for this test are carefully selected to demonstrate some of the
behavior differences of the two current name hash versions, including
which conditions will cause them to collide.
Create a performance test that uses test_size to demonstrate how
collisions occur for these hash algorithms. This test helps inform
someone as to the behavior of the name-hash algorithms for their repo
based on the paths at HEAD.
My copy of the Git repository shows modest statistics around the
collisions of the default name-hash algorithm:
Test this tree
--------------------------------------------------
5314.1: paths at head 4.5K
5314.2: distinct hash value: v1 4.1K
5314.3: maximum multiplicity: v1 13
5314.4: distinct hash value: v2 4.2K
5314.5: maximum multiplicity: v2 9
Here, the maximum collision multiplicity is 13, but around 10% of paths
have a collision with another path.
In a more interesting example, the microsoft/fluentui [1] repo had these
statistics at time of committing:
Test this tree
--------------------------------------------------
5314.1: paths at head 19.5K
5314.2: distinct hash value: v1 8.2K
5314.3: maximum multiplicity: v1 279
5314.4: distinct hash value: v2 17.8K
5314.5: maximum multiplicity: v2 44
[1] https://github.com/microsoft/fluentui
That demonstrates that of the nearly twenty thousand path names, they
are assigned around eight thousand distinct values. 279 paths are
assigned to a single value, leading the packing algorithm to sort
objects from those paths together, by size.
With the v2 name hash function, the maximum multiplicity lowers to 44,
leaving some room for further improvement.
In a more extreme example, an internal monorepo had a much worse
collision rate:
Test this tree
--------------------------------------------------
5314.1: paths at head 227.3K
5314.2: distinct hash value: v1 72.3K
5314.3: maximum multiplicity: v1 14.4K
5314.4: distinct hash value: v2 166.5K
5314.5: maximum multiplicity: v2 138
Here, we can see that the v2 name hash function provides somem
improvements, but there are still a number of collisions that could lead
to repacking problems at this scale.
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As custom options are added to 'git pack-objects' and 'git repack' to
adjust how compression is done, use this new performance test script to
demonstrate their effectiveness in performance and size.
The recently-added --name-hash-version option allows for testing
different name hash functions. Version 2 intends to preserve some of the
locality of version 1 while more often breaking collisions due to long
filenames.
Distinguishing objects by more of the path is critical when there are
many name hash collisions and several versions of the same path in the
full history, giving a significant boost to the full repack case. The
locality of the hash function is critical to compressing something like
a shallow clone or a thin pack representing a push of a single commit.
This can be seen by running pt5313 on the open source fluentui
repository [1]. Most commits will have this kind of output for the thin
and big pack cases, though certain commits (such as [2]) will have
problematic thin pack size for other reasons.
[1] https://github.com/microsoft/fluentui
[2] a637a06df05360ce5ff21420803f64608226a875
Checked out at the parent of [2], I see the following statistics:
Test HEAD
---------------------------------------------------------------
5313.2: thin pack with version 1 0.37(0.44+0.02)
5313.3: thin pack size with version 1 1.2M
5313.4: big pack with version 1 2.04(7.77+0.23)
5313.5: big pack size with version 1 20.4M
5313.6: shallow fetch pack with version 1 1.41(2.94+0.11)
5313.7: shallow pack size with version 1 34.4M
5313.8: repack with version 1 95.70(676.41+2.87)
5313.9: repack size with version 1 439.3M
5313.10: thin pack with version 2 0.12(0.12+0.06)
5313.11: thin pack size with version 2 22.0K
5313.12: big pack with version 2 2.80(5.43+0.34)
5313.13: big pack size with version 2 25.9M
5313.14: shallow fetch pack with version 2 1.77(2.80+0.19)
5313.15: shallow pack size with version 2 33.7M
5313.16: repack with version 2 33.68(139.52+2.58)
5313.17: repack size with version 2 160.5M
To make comparisons easier, I will reformat this output into a different
table style:
| Test | V1 Time | V2 Time | V1 Size | V2 Size |
|--------------|---------|---------|---------|---------|
| Thin Pack | 0.37 s | 0.12 s | 1.2 M | 22.0 K |
| Big Pack | 2.04 s | 2.80 s | 20.4 M | 25.9 M |
| Shallow Pack | 1.41 s | 1.77 s | 34.4 M | 33.7 M |
| Repack | 95.70 s | 33.68 s | 439.3 M | 160.5 M |
The v2 hash function successfully differentiates the CHANGELOG.md files
from each other, which leads to significant improvements in the thin
pack (simulating a push of this commit) and the full repack. There is
some bloat in the "big pack" scenario and essentially the same results
for the shallow pack.
In the case of the Git repository, these numbers show some of the issues
with this approach:
| Test | V1 Time | V2 Time | V1 Size | V2 Size |
|--------------|---------|---------|---------|---------|
| Thin Pack | 0.02 s | 0.02 s | 1.1 K | 1.1 K |
| Big Pack | 1.69 s | 1.95 s | 13.5 M | 14.5 M |
| Shallow Pack | 1.26 s | 1.29 s | 12.0 M | 12.2 M |
| Repack | 29.51 s | 29.01 s | 237.7 M | 238.2 M |
Here, the attempts to remove conflicts in the v2 function seem to cause
slight bloat to these sizes. This shows that the Git repository benefits
a lot from cross-path delta pairs.
The results are similar with the nodejs/node repo:
| Test | V1 Time | V2 Time | V1 Size | V2 Size |
|--------------|---------|---------|---------|---------|
| Thin Pack | 0.02 s | 0.02 s | 1.6 K | 1.6 K |
| Big Pack | 4.61 s | 3.26 s | 56.0 M | 52.8 M |
| Shallow Pack | 7.82 s | 7.51 s | 104.6 M | 107.0 M |
| Repack | 88.90 s | 73.75 s | 740.1 M | 764.5 M |
Here, the v2 name-hash causes some size bloat more often than it reduces
the size, but it also universally improves performance time, which is an
interesting reversal. This must mean that it is helping to short-circuit
some delta computations even if it is not finding the most efficient
ones. The performance improvement cannot be explained only due to the
I/O cost of writing the resulting packfile.
The Linux kernel repository was the initial target of the default name
hash value, and its naming conventions are practically build to take the
most advantage of the default name hash values:
| Test | V1 Time | V2 Time | V1 Size | V2 Size |
|--------------|----------|----------|---------|---------|
| Thin Pack | 0.17 s | 0.07 s | 4.6 K | 4.6 K |
| Big Pack | 17.88 s | 12.35 s | 201.1 M | 159.1 M |
| Shallow Pack | 11.05 s | 22.94 s | 269.2 M | 273.8 M |
| Repack | 727.39 s | 566.95 s | 2.5 G | 2.5 G |
Here, the thin and big packs gain some performance boosts in time, with
a modest gain in the size of the big pack. The shallow pack, however, is
more expensive to compute, likely because similarly-named files across
different directories are farther apart in the name hash ordering in v2.
The repack also gains benefits in computation time but no meaningful
change to the full size.
Finally, an internal Javascript repo of moderate size shows significant
gains when repacking with --name-hash-version=2 due to it having many name
hash collisions. However, it's worth noting that only the full repack
case has significant differences from the v1 name hash:
| Test | V1 Time | V2 Time | V1 Size | V2 Size |
|-----------|-----------|----------|---------|---------|
| Thin Pack | 8.28 s | 7.28 s | 16.8 K | 16.8 K |
| Big Pack | 12.81 s | 11.66 s | 29.1 M | 29.1 M |
| Shallow | 4.86 s | 4.06 s | 42.5 M | 44.1 M |
| Repack | 3126.50 s | 496.33 s | 6.2 G | 855.6 M |
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git describe" optimization.
* jk/describe-perf:
describe: split "found all tags" and max_candidates logic
describe: stop traversing when we run out of names
describe: stop digging for max_candidates+1
t/perf: add tests for git-describe
t6120: demonstrate weakness in disjoint-root handling
The perf test suite prefers to use test_file_size over 'wc -c' when
inside of a test_size block. One advantage is that accidentally writign
"wc -c file" (instead of "wc -c <file") does not inadvertently break the
tests (since the former will include the filename in the output of wc).
Both of the two uses of test_size use "wc -c", but let's convert those
to the more conventional test_file_size helper instead.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We don't have a perf script for git-describe, despite it often being
accused of slowness. Let's add a few simple tests to start with.
Rather than use the existing tags from our test repo, we'll make our own
so that we have a known quantity and position. We'll add a "new" tag
near the tip of HEAD, and an "old" one that is at the very bottom. And
then our tests are:
1. Describing HEAD naively requires walking all the way down to the
old tag as we collect candidates. This gives us a baseline for what
"slow" looks like.
2. Doing the same with --candidates=1 can potentially be fast, because
we can quie after finding "new". But we don't, and it's also slow.
3. Likewise we should be able to quit when there are no more tags to
find. This can happen naturally if a repo has few tags, but also if
you restrict the set of tags with --match.
Here are the results running against linux.git. Note that I have a
commit-graph built for the repo, so "slow" here is ~700ms. Without a
commit graph it's more like 9s!
Test HEAD
--------------------------------------------------------------
6100.2: describe HEAD 0.70(0.66+0.04)
6100.3: describe HEAD with one max candidate 0.70(0.66+0.04)
6100.4: describe HEAD with one tag 0.70(0.64+0.06)
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous two changes introduced a commit walking heuristic for finding
the most likely base branch for a given source. This algorithm walks
first-parent histories until reaching a collision.
This walk _should_ be very fast. Exceptions include cases where a
commit-graph file does not exist, leading to a full walk of all reachable
commits to compute generation numbers, or a case where no collision in the
first-parent history exists, leading to a walk of all first-parent history
to the root commits.
The p1500 test script guarantees a complete commit-graph file during its
setup, so we will not test that scenario. Do create a new root commit in an
effort to test the scenario of parallel first-parent histories.
Even with the extra root commit, these tests take no longer than 0.02
seconds on my machine for the Git repository. However, the results are
slightly more interesting in a copy of the Linux kernel repository:
Test
---------------------------------------------------------------
1500.2: ahead-behind counts: git for-each-ref 0.12
1500.3: ahead-behind counts: git branch 0.12
1500.4: ahead-behind counts: git tag 0.12
1500.5: contains: git for-each-ref --merged 0.04
1500.6: contains: git branch --merged 0.04
1500.7: contains: git tag --merged 0.04
1500.8: is-base check: test-tool reach (refs) 0.03
1500.9: is-base check: test-tool reach (tags) 0.03
1500.10: is-base check: git for-each-ref 0.03
1500.11: is-base check: git for-each-ref (disjoint-base) 0.07
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Implement a straightforward performance test demonstrating the benefit
of pseudo-merge bitmaps by measuring how long it takes to count
reachable objects in a few different scenarios:
- without bitmaps, to demonstrate a reasonable baseline
- with bitmaps, but without pseudo-merges
- with bitmaps and pseudo-merges
Results from running this test on git.git are as follows:
Test this tree
-----------------------------------------------------------------------------------
5333.2: git rev-list --count --all --objects (no bitmaps) 3.54(3.45+0.08)
5333.3: git rev-list --count --all --objects (no pseudo-merges) 0.43(0.40+0.03)
5333.4: git rev-list --count --all --objects (with pseudo-merges) 0.12(0.11+0.01)
On a private repository which is much larger, and has many spikey parts
of history that aren't merged into the 'master' branch, the results are
as follows:
Test this tree
---------------------------------------------------------------------------------------
5333.1: git rev-list --count --all --objects (no bitmaps) 122.29(121.31+0.97)
5333.2: git rev-list --count --all --objects (no pseudo-merges) 21.88(21.30+0.58)
5333.3: git rev-list --count --all --objects (with pseudo-merges) 5.05(4.77+0.28)
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Take care to redirect stdin, otherwise the output of wc would also contain
the file name.
Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Streaming spans of packfile data used to be done only from a
single, primary, pack in a repository with multiple packfiles. It
has been extended to allow reuse from other packfiles, too.
* tb/multi-pack-verbatim-reuse: (26 commits)
t/perf: add performance tests for multi-pack reuse
pack-bitmap: enable reuse from all bitmapped packs
pack-objects: allow setting `pack.allowPackReuse` to "single"
t/test-lib-functions.sh: implement `test_trace2_data` helper
pack-objects: add tracing for various packfile metrics
pack-bitmap: prepare to mark objects from multiple packs for reuse
pack-revindex: implement `midx_pair_to_pack_pos()`
pack-revindex: factor out `midx_key_to_pack_pos()` helper
midx: implement `midx_preferred_pack()`
git-compat-util.h: implement checked size_t to uint32_t conversion
pack-objects: include number of packs reused in output
pack-objects: prepare `write_reused_pack_verbatim()` for multi-pack reuse
pack-objects: prepare `write_reused_pack()` for multi-pack reuse
pack-objects: pass `bitmapped_pack`'s to pack-reuse functions
pack-objects: keep track of `pack_start` for each reuse pack
pack-objects: parameterize pack-reuse routines over a single pack
pack-bitmap: return multiple packs via `reuse_partial_packfile_from_bitmap()`
pack-bitmap: simplify `reuse_partial_packfile_from_bitmap()` signature
ewah: implement `bitmap_is_empty()`
pack-bitmap: pass `bitmapped_pack` struct to pack-reuse functions
...
Stale URLs have been updated to their current counterparts (or
archive.org) and HTTP links are replaced with working HTTPS links.
* js/update-urls-in-doc-and-comment:
doc: refer to internet archive
doc: update links for andre-simon.de
doc: switch links to https
doc: update links to current pages
To ensure that we don't regress either the size or runtime performance
of multi-pack reuse, add a performance test to measure both of these.
The test partitions the objects in GIT_TEST_PERF_LARGE_REPO into 1, 10,
and 100 packs, and then tries to perform a "clone" at each stage with
both single- and multi-pack reuse enabled.
Note that the `repack_into_n_chunks()` function in this new test script
differs from the existing `repack_into_n()`. The former partitions the
repository into N equal-sized chunks, while the latter produces N packs
of five commits each (plus their objects), and then another pack with
the remainder.
On git.git, I can produce the following results on my machine:
Test this tree
--------------------------------------------------------------------------------
5332.3: clone for 1-pack scenario (single-pack reuse) 1.57(2.99+0.15)
5332.4: clone size for 1-pack scenario (single-pack reuse) 231.8M
5332.5: clone for 1-pack scenario (multi-pack reuse) 1.79(2.96+0.21)
5332.6: clone size for 1-pack scenario (multi-pack reuse) 231.7M
5332.9: clone for 10-pack scenario (single-pack reuse) 3.89(16.75+0.35)
5332.10: clone size for 10-pack scenario (single-pack reuse) 209.9M
5332.11: clone for 10-pack scenario (multi-pack reuse) 1.56(2.99+0.17)
5332.12: clone size for 10-pack scenario (multi-pack reuse) 224.4M
5332.15: clone for 100-pack scenario (single-pack reuse) 8.24(54.31+0.59)
5332.16: clone size for 100-pack scenario (single-pack reuse) 278.3M
5332.17: clone for 100-pack scenario (multi-pack reuse) 2.13(2.44+0.33)
5332.18: clone size for 100-pack scenario (multi-pack reuse) 357.9M
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git for-each-ref --no-sort" still sorted the refs alphabetically
which paid non-trivial cost. It has been redefined to show output
in an unspecified order, to allow certain optimizations to take
advantage of.
* vd/for-each-ref-unsorted-optimization:
t/perf: add perf tests for for-each-ref
ref-filter.c: use peeled tag for '*' format fields
for-each-ref: clean up documentation of --format
ref-filter.c: filter & format refs in the same callback
ref-filter.c: refactor to create common helper functions
ref-filter.c: rename 'ref_filter_handler()' to 'filter_one()'
ref-filter.h: add functions for filter/format & format-only
ref-filter.h: move contains caches into filter
ref-filter.h: add max_count and omit_empty to ref_format
ref-filter.c: really don't sort when using --no-sort
These sites offer https versions of their content.
Using the https versions provides some protection for users.
Signed-off-by: Josh Soref <jsoref@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add performance tests for 'for-each-ref'. The tests exercise different
combinations of filters/formats/options, as well as the overall performance
of 'git for-each-ref | git cat-file --batch-check' to demonstrate the
performance difference vs. 'git for-each-ref' with "%(*fieldname)" format
specifiers.
All tests are run against a repository with 40k loose refs - 10k commits,
each having a unique:
- branch
- custom ref (refs/custom/special_*)
- annotated tag pointing at the commit
- annotated tag pointing at the other annotated tag (i.e., a nested tag)
After those tests are finished, the refs are packed with 'pack-refs --all'
and the same tests are rerun.
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our coding guidelines say to not use `test` with `-a` and `-o` because
it can easily lead to bugs. Convert trivial cases where we still use
these to instead instead concatenate multiple invocations of `test` via
`&&` and `||`, respectively.
While not all of the converted instances can cause ambiguity, it is
worth getting rid of all of them regardless:
- It becomes easier to reason about the code as we do not have to
argue why one use of `-a`/`-o` is okay while another one isn't.
- We don't encourage people to use these expressions.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Set the requires-full-index to false for "check-attr".
Add a test to ensure that the index is not expanded whether the files
are outside or inside the sparse-checkout cone when the sparse index is
enabled.
The `p2000` tests demonstrate a ~63% execution time reduction for
'git check-attr' using a sparse index.
Test before after
-----------------------------------------------------------------------
2000.106: git check-attr -a f2/f4/a (full-v3) 0.05 0.05 +0.0%
2000.107: git check-attr -a f2/f4/a (full-v4) 0.05 0.05 +0.0%
2000.108: git check-attr -a f2/f4/a (sparse-v3) 0.04 0.02 -50.0%
2000.109: git check-attr -a f2/f4/a (sparse-v4) 0.04 0.01 -75.0%
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shuqi Liang <cheskaqiqi@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The index is read in 'worktree.c' at two points:
1.The 'validate_no_submodules' function, which checks if there are any
submodules present in the worktree.
2.The 'check_clean_worktree' function, which verifies if a worktree is
'clean', i.e., there are no untracked or modified but uncommitted files.
This is done by running the 'git status' command, and an error message
is thrown if the worktree is not clean. Given that 'git status' is
already sparse-aware, the function is also sparse-aware.
Hence we can just set the requires-full-index to false for
"git worktree".
Add tests that verify that 'git worktree' behaves correctly when the
sparse index is enabled and test to ensure the index is not expanded.
The `p2000` tests demonstrate a ~20% execution time reduction for
'git worktree' using a sparse index:
(Note:the p2000 test results didn't reflect the huge speedup because of
the index reading time is minuscule comparing to the filesystem
operations.)
Test before after
-----------------------------------------------------------------------
2000.102: git worktree add....(full-v3) 3.15 2.82 -10.5%
2000.103: git worktree add....(full-v4) 3.14 2.84 -9.6%
2000.104: git worktree add....(sparse-v3) 2.59 2.14 -16.4%
2000.105: git worktree add....(sparse-v4) 2.10 1.57 -25.2%
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shuqi Liang <cheskaqiqi@gmail.com>
Acked-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The index is read in 'cmd_diff_tree' at two points:
1. The first index read was added in fd66bcc31f (diff-tree: read the
index so attribute checks work in bare repositories, 2017-12-06) to deal
with reading '.gitattributes' content. 77efbb366a (attr: be careful
about sparse directories, 2021-09-08) established that, in a sparse
index, we do _not_ try to load a '.gitattributes' file from within a
sparse directory.
2. The second index access point is involved in rename detection,
specifically when reading from stdin.This was initially added in
f0c6b2a2fd ([PATCH] Optimize diff-tree -[CM]--stdin, 2005-05-27), where
'setup' was set to 'DIFF_SETUP_USE_SIZE_CACHE |DIFF_SETUP_USE_CACHE'.
That assignment was later modified to drop the'DIFF_SETUP_USE_CACHE' in
ff7fe37b05 (diff.c: move read_index() code back to the caller,
2018-08-13).However, 'DIFF_SETUP_USE_SIZE_CACHE' seems to be unused as
of 6e0b8ed6d3 (diff.c: do not use a separate "size cache"., 2007-05-07)
and nothing about 'detect_rename' otherwise indicates index usage.
Hence we can just set the requires-full-index to false for "diff-tree".
Add tests that verify that 'git diff-tree' behaves correctly when the
sparse index is enabled and test to ensure the index is not expanded.
The `p2000` tests demonstrate a ~98% execution time reduction for
'git diff-tree' using a sparse index:
Test before after
-----------------------------------------------------------------------
2000.94: git diff-tree HEAD (full-v3) 0.05 0.04 -20.0%
2000.95: git diff-tree HEAD (full-v4) 0.06 0.05 -16.7%
2000.96: git diff-tree HEAD (sparse-v3) 0.59 0.01 -98.3%
2000.97: git diff-tree HEAD (sparse-v4) 0.61 0.01 -98.4%
2000.98: git diff-tree HEAD -- f2/f4/a (full-v3) 0.05 0.05 +0.0%
2000.99: git diff-tree HEAD -- f2/f4/a (full-v4) 0.05 0.04 -20.0%
2000.100: git diff-tree HEAD -- f2/f4/a (sparse-v3) 0.58 0.01 -98.3%
2000.101: git diff-tree HEAD -- f2/f4/a (sparse-v4) 0.55 0.01 -98.2%
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shuqi Liang <cheskaqiqi@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach "diff-files" not to expand sparse-index unless needed.
* sl/diff-files-sparse:
diff-files: integrate with sparse index
t1092: add tests for `git diff-files`
Remove full index requirement for `git diff-files`. Refactor the
ensure_expanded and ensure_not_expanded functions by introducing a
common helper function, ensure_index_state. Add test to ensure the index
is no expanded in `git diff-files`.
The `p2000` tests demonstrate a ~96% execution time reduction for 'git
diff-files' and a ~97% execution time reduction for 'git diff-files'
for a file using a sparse index:
Test before after
-----------------------------------------------------------------------
2000.94: git diff-files (full-v3) 0.09 0.08 -11.1%
2000.95: git diff-files (full-v4) 0.09 0.09 +0.0%
2000.96: git diff-files (sparse-v3) 0.52 0.02 -96.2%
2000.97: git diff-files (sparse-v4) 0.51 0.02 -96.1%
2000.98: git diff-files -- f2/f4/a (full-v3) 0.06 0.07 +16.7%
2000.99: git diff-files -- f2/f4/a (full-v4) 0.08 0.08 +0.0%
2000.100: git diff-files -- f2/f4/a (sparse-v3) 0.46 0.01 -97.8%
2000.101: git diff-files -- f2/f4/a (sparse-v4) 0.51 0.02 -96.1%
Signed-off-by: Shuqi Liang <cheskaqiqi@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The on-disk reverse index that allows mapping from the pack offset
to the object name for the object stored at the offset has been
enabled by default.
* tb/pack-revindex-on-disk:
t: invert `GIT_TEST_WRITE_REV_INDEX`
config: enable `pack.writeReverseIndex` by default
pack-revindex: introduce `pack.readReverseIndex`
pack-revindex: introduce GIT_TEST_REV_INDEX_DIE_ON_DISK
pack-revindex: make `load_pack_revindex` take a repository
t5325: mark as leak-free
pack-write.c: plug a leak in stage_tmp_packfiles()
Back in e37d0b8730 (builtin/index-pack.c: write reverse indexes,
2021-01-25), Git learned how to read and write a pack's reverse index
from a file instead of in-memory.
A pack's reverse index is a mapping from pack position (that is, the
order that objects appear together in a ".pack") to their position in
lexical order (that is, the order that objects are listed in an ".idx"
file).
Reverse indexes are consulted often during pack-objects, as well as
during auxiliary operations that require mapping between pack offsets,
pack order, and index index.
They are useful in GitHub's infrastructure, where we have seen a
dramatic increase in performance when writing ".rev" files[1]. In
particular:
- an ~80% reduction in the time it takes to serve fetches on a popular
repository, Homebrew/homebrew-core.
- a ~60% reduction in the peak memory usage to serve fetches on that
same repository.
- a collective savings of ~35% in CPU time across all pack-objects
invocations serving fetches across all repositories in a single
datacenter.
Reverse indexes are also beneficial to end-users as well as forges. For
example, the time it takes to generate a pack containing the objects for
the 10 most recent commits in linux.git (representing a typical push) is
significantly faster when on-disk reverse indexes are available:
$ { git rev-parse HEAD && printf '^' && git rev-parse HEAD~10 } >in
$ hyperfine -L v false,true 'git.compile -c pack.readReverseIndex={v} pack-objects --delta-base-offset --revs --stdout <in >/dev/null'
Benchmark 1: git.compile -c pack.readReverseIndex=false pack-objects --delta-base-offset --revs --stdout <in >/dev/null
Time (mean ± σ): 543.0 ms ± 20.3 ms [User: 616.2 ms, System: 58.8 ms]
Range (min … max): 521.0 ms … 577.9 ms 10 runs
Benchmark 2: git.compile -c pack.readReverseIndex=true pack-objects --delta-base-offset --revs --stdout <in >/dev/null
Time (mean ± σ): 245.0 ms ± 11.4 ms [User: 335.6 ms, System: 31.3 ms]
Range (min … max): 226.0 ms … 259.6 ms 13 runs
Summary
'git.compile -c pack.readReverseIndex=true pack-objects --delta-base-offset --revs --stdout <in >/dev/null' ran
2.22 ± 0.13 times faster than 'git.compile -c pack.readReverseIndex=false pack-objects --delta-base-offset --revs --stdout <in >/dev/null'
The same is true of writing a pack containing the objects for the 30
most-recent commits:
$ { git rev-parse HEAD && printf '^' && git rev-parse HEAD~30 } >in
$ hyperfine -L v false,true 'git.compile -c pack.readReverseIndex={v} pack-objects --delta-base-offset --revs --stdout <in >/dev/null'
Benchmark 1: git.compile -c pack.readReverseIndex=false pack-objects --delta-base-offset --revs --stdout <in >/dev/null
Time (mean ± σ): 866.5 ms ± 16.2 ms [User: 1414.5 ms, System: 97.0 ms]
Range (min … max): 839.3 ms … 886.9 ms 10 runs
Benchmark 2: git.compile -c pack.readReverseIndex=true pack-objects --delta-base-offset --revs --stdout <in >/dev/null
Time (mean ± σ): 581.6 ms ± 10.2 ms [User: 1181.7 ms, System: 62.6 ms]
Range (min … max): 567.5 ms … 599.3 ms 10 runs
Summary
'git.compile -c pack.readReverseIndex=true pack-objects --delta-base-offset --revs --stdout <in >/dev/null' ran
1.49 ± 0.04 times faster than 'git.compile -c pack.readReverseIndex=false pack-objects --delta-base-offset --revs --stdout <in >/dev/null'
...and savings on trivial operations like computing the on-disk size of
a single (packed) object are even more dramatic:
$ git rev-parse HEAD >in
$ hyperfine -L v false,true 'git.compile -c pack.readReverseIndex={v} cat-file --batch-check="%(objectsize:disk)" <in'
Benchmark 1: git.compile -c pack.readReverseIndex=false cat-file --batch-check="%(objectsize:disk)" <in
Time (mean ± σ): 305.8 ms ± 11.4 ms [User: 264.2 ms, System: 41.4 ms]
Range (min … max): 290.3 ms … 331.1 ms 10 runs
Benchmark 2: git.compile -c pack.readReverseIndex=true cat-file --batch-check="%(objectsize:disk)" <in
Time (mean ± σ): 4.0 ms ± 0.3 ms [User: 1.7 ms, System: 2.3 ms]
Range (min … max): 1.6 ms … 4.6 ms 1155 runs
Summary
'git.compile -c pack.readReverseIndex=true cat-file --batch-check="%(objectsize:disk)" <in' ran
76.96 ± 6.25 times faster than 'git.compile -c pack.readReverseIndex=false cat-file --batch-check="%(objectsize:disk)" <in'
In the more than two years since e37d0b8730 was merged, Git's
implementation of on-disk reverse indexes has been thoroughly tested,
both from users enabling `pack.writeReverseIndexes`, and from GitHub's
deployment of the feature. The latter has been running without incident
for more than two years.
This patch changes Git's behavior to write on-disk reverse indexes by
default when indexing a pack, which should make the above operations
faster for everybody's Git installation after a repack.
(The previous commit explains some potential drawbacks of using on-disk
reverse indexes in certain limited circumstances, that essentially boil
down to a trade-off between time to generate, and time to access. For
those limited cases, the `pack.readReverseIndex` escape hatch can be
used).
[1]: https://github.blog/2021-04-29-scaling-monorepo-maintenance/#reverse-indexes
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git for-each-ref" learns '%(ahead-behind:<base>)' that computes the
distances from a single reference point in the history with bunch
of commits in bulk.
* ds/ahead-behind:
commit-reach: add tips_reachable_from_bases()
for-each-ref: add ahead-behind format atom
commit-reach: implement ahead_behind() logic
commit-graph: introduce `ensure_generations_valid()`
commit-graph: return generation from memory
commit-graph: simplify compute_generation_numbers()
commit-graph: refactor compute_topological_levels()
for-each-ref: explicitly test no matches
for-each-ref: add --stdin option
Update 'git write-tree' to allow using the sparse-index in memory
without expanding to a full one.
The recursive algorithm for update_one() was already updated in 2de37c5
(cache-tree: integrate with sparse directory entries, 2021-03-03) to
handle sparse directory entries in the index. Hence we can just set the
requires-full-index to false for "write-tree".
The `p2000` tests demonstrate a ~96% execution time reduction for 'git
write-tree' using a sparse index:
Test before after
-----------------------------------------------------------------
2000.78: git write-tree (full-v3) 0.34 0.33 -2.9%
2000.79: git write-tree (full-v4) 0.32 0.30 -6.3%
2000.80: git write-tree (sparse-v3) 0.47 0.02 -95.8%
2000.81: git write-tree (sparse-v4) 0.45 0.02 -95.6%
Signed-off-by: Shuqi Liang <cheskaqiqi@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano