hashmap_free(), hashmap_free_entries(), and hashmap_free_() have existed
for a while, but aren't necessarily the clearest names, especially with
hashmap_partial_clear() being added to the mix and lazy-initialization
now being supported. Peff suggested we adopt the following names[1]:
- hashmap_clear() - remove all entries and de-allocate any
hashmap-specific data, but be ready for reuse
- hashmap_clear_and_free() - ditto, but free the entries themselves
- hashmap_partial_clear() - remove all entries but don't deallocate
table
- hashmap_partial_clear_and_free() - ditto, but free the entries
This patch provides the new names and converts all existing callers over
to the new naming scheme.
[1] https://lore.kernel.org/git/20201030125059.GA3277724@coredump.intra.peff.net/
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Introduce a command-line flag to specify the maximum number of new Bloom
filters that a 'git commit-graph write' is willing to compute from
scratch.
Prior to this patch, a commit-graph write with '--changed-paths' would
compute Bloom filters for all selected commits which haven't already
been computed (i.e., by a previous commit-graph write with '--split'
such that a roll-up or replacement is performed).
This behavior can cause prohibitively-long commit-graph writes for a
variety of reasons:
* There may be lots of filters whose diffs take a long time to
generate (for example, they have close to the maximum number of
changes, diffing itself takes a long time, etc).
* Old-style commit-graphs (which encode filters with too many entries
as not having been computed at all) cause us to waste time
recomputing filters that appear to have not been computed only to
discover that they are too-large.
This can make the upper-bound of the time it takes for 'git commit-graph
write --changed-paths' to be rather unpredictable.
To make this command behave more predictably, introduce
'--max-new-filters=<n>' to allow computing at most '<n>' Bloom filters
from scratch. This lets "computing" already-known filters proceed
quickly, while bounding the number of slow tasks that Git is willing to
do.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a changed-path Bloom filter has either zero, or more than a
certain number (commonly 512) of entries, the commit-graph machinery
encodes it as "missing". More specifically, it sets the indices adjacent
in the BIDX chunk as equal to each other to indicate a "length 0"
filter; that is, that the filter occupies zero bytes on disk.
This has heretofore been fine, since the commit-graph machinery has no
need to care about these filters with too few or too many changed paths.
Both cases act like no filter has been generated at all, and so there is
no need to store them.
In a subsequent commit, however, the commit-graph machinery will learn
to only compute Bloom filters for some commits in the current
commit-graph layer. This is a change from the current implementation
which computes Bloom filters for all commits that are in the layer being
written. Critically for this patch, only computing some of the Bloom
filters means adding a third state for length 0 Bloom filters: zero
entries, too many entries, or "hasn't been computed".
It will be important for that future patch to distinguish between "not
representable" (i.e., zero or too-many changed paths), and "hasn't been
computed". In particular, we don't want to waste time recomputing
filters that have already been computed.
To that end, change how we store Bloom filters in the "computed but not
representable" category:
- Bloom filters with no entries are stored as a single byte with all
bits low (i.e., all queries to that Bloom filter will return
"definitely not")
- Bloom filters with too many entries are stored as a single byte with
all bits set high (i.e., all queries to that Bloom filter will
return "maybe").
These rules are sufficient to not incur a behavior change by changing
the on-disk representation of these two classes. Likewise, no
specification changes are necessary for the commit-graph format, either:
- Filters that were previously empty will be recomputed and stored
according to the new rules, and
- old clients reading filters generated by new clients will interpret
the filters correctly and be none the wiser to how they were
generated.
Clients will invoke the Bloom machinery in more cases than before, but
this can be addressed by returning a NULL filter when all bits are set
high. This can be addressed in a future patch.
Note that this does increase the size of on-disk commit-graphs, but far
less than other proposals. In particular, this is generally more
efficient than storing a bitmap for which commits haven't computed their
Bloom filters. Storing a bitmap incurs a penalty of one bit per commit,
whereas storing explicit filters as above incurs a penalty of one byte
per too-large or empty commit.
In practice, these boundary commits likely occupy a small proportion of
the overall number of commits, and so the size penalty is likely smaller
than storing a bitmap for all commits.
See, for example, these relative proportions of such boundary commits
(collected by SZEDER Gábor):
| Percentage of | commit-graph | |
| commits modifying | file size | |
├────────┬──────────────┼───────────────────┤ pct. |
| 0 path | >= 512 paths | before | after | change |
┌────────────────┼────────┼──────────────┼─────────┼─────────┼───────────┤
| android-base | 13.20% | 0.13% | 37.468M | 37.534M | +0.1741 % |
| cmssw | 0.15% | 0.23% | 17.118M | 17.119M | +0.0091 % |
| cpython | 3.07% | 0.01% | 7.967M | 7.971M | +0.0423 % |
| elasticsearch | 0.70% | 1.00% | 8.833M | 8.835M | +0.0128 % |
| gcc | 0.00% | 0.08% | 16.073M | 16.074M | +0.0030 % |
| gecko-dev | 0.14% | 0.64% | 59.868M | 59.874M | +0.0105 % |
| git | 0.11% | 0.02% | 3.895M | 3.895M | +0.0020 % |
| glibc | 0.02% | 0.10% | 3.555M | 3.555M | +0.0021 % |
| go | 0.00% | 0.07% | 3.186M | 3.186M | +0.0018 % |
| homebrew-cask | 0.40% | 0.02% | 7.035M | 7.035M | +0.0065 % |
| homebrew-core | 0.01% | 0.01% | 11.611M | 11.611M | +0.0002 % |
| jdk | 0.26% | 5.64% | 5.537M | 5.540M | +0.0590 % |
| linux | 0.01% | 0.51% | 63.735M | 63.740M | +0.0073 % |
| llvm-project | 0.12% | 0.03% | 25.515M | 25.516M | +0.0050 % |
| rails | 0.10% | 0.10% | 6.252M | 6.252M | +0.0027 % |
| rust | 0.07% | 0.17% | 9.364M | 9.364M | +0.0033 % |
| tensorflow | 0.09% | 1.02% | 7.009M | 7.010M | +0.0158 % |
| webkit | 0.05% | 0.31% | 17.405M | 17.406M | +0.0047 % |
(where the above increase is determined by computing a non-split
commit-graph before and after this patch).
Given that these projects are all "large" by commit count, the storage
cost by writing these filters explicitly is negligible. In the most
extreme example, android-base (which has 494,848 commits at the time of
writing) would have its commit-graph increase by a modest 68.4 KB.
Finally, a test to exercise filters which contain too many changed path
entries will be introduced in a subsequent patch.
Suggested-by: SZEDER Gábor <szeder.dev@gmail.com>
Suggested-by: Jakub Narębski <jnareb@gmail.com>
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Helped-by: SZEDER Gábor <szeder.dev@gmail.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit e3696980 (diff: halt tree-diff early after max_changes,
2020-03-30) intended to create a mechanism to short-circuit a diff
calculation after a certain number of paths were modified. By
incrementing a "num_changes" counter throughout the recursive
ll_diff_tree_paths(), this was supposed to match the number of changes
that would be written into the changed-path Bloom filters.
Unfortunately, this was not implemented correctly and instead misses
simple cases like file modifications. This then does not stop very
large changed-path filters from being written (unless they add or remove
many files).
To start, change the implementation in ll_diff_tree_paths() to instead
use the global diff_queue_diff struct's 'nr' member as the count. This
is a way to simplify the logic instead of making more mistakes in the
complicated diff code.
This has a drawback: the diff_queue_diff struct only lists the paths
corresponding to blob changes, not their leading directories. Thus,
get_or_compute_bloom_filter() needs an additional check to see if the
hashmap with the leading directories becomes too large.
One reason why this was not caught by test cases was that the test in
t4216-log-bloom.sh that was supposed to check this "too many changes"
condition only checked this on the initial commit of a repository. The
old logic counted these values correctly. Update this test in a few
ways:
1. Use GIT_TEST_BLOOM_SETTINGS_MAX_CHANGED_PATHS to reduce the limit,
allowing smaller commits to engage with this logic.
2. Create several interesting cases of edits, adds, removes, and mode
changes (in the second commit). By testing both sides of the
inequality with the *_MAX_CHANGED_PATHS variable, we can see that
the count is exactly correct, so none of these changes are missed
or over-counted.
3. Use the trace2 data value filter_found_large to verify that these
commits are on the correct side of the limit.
Another way to verify the behavior is correct is through performance
tests. By testing on my local copies of the Git repository and the Linux
kernel repository, I could measure the effect of these short-circuits
when computing a fresh commit-graph file with changed-path Bloom filters
using the command
GIT_TEST_BLOOM_SETTINGS_MAX_CHANGED_PATHS=N time \
git commit-graph write --reachable --changed-paths
and reporting the wall time and resulting commit-graph size.
For Git, the results are
| | N=1 | N=10 | N=512 |
|--------|----------------|----------------|----------------|
| HEAD~1 | 10.90s 9.18MB | 11.11s 9.34MB | 11.31s 9.35MB |
| HEAD | 9.21s 8.62MB | 11.11s 9.29MB | 11.29s 9.34MB |
For Linux, the results are
| | N=1 | N=20 | N=512 |
|--------|----------------|---------------|---------------|
| HEAD~1 | 61.28s 64.3MB | 76.9s 72.6MB | 77.6s 72.6MB |
| HEAD | 49.44s 56.3MB | 68.7s 65.9MB | 69.2s 65.9MB |
Naturally, the improvement becomes much less as the limit grows, as
fewer commits satisfy the short-circuit.
Reported-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When 'get_or_compute_bloom_filter()' needs to compute a Bloom filter
from scratch, it looks to the default 'struct bloom_filter_settings' in
order to determine the maximum number of changed paths, number of bits
per entry, and so on.
All of these values have so far been constant, and so there was no need
to pass in a pointer from the caller (eg., the one that is stored in the
'struct write_commit_graph_context').
Start passing in a 'struct bloom_filter_settings *' instead of using the
default values to respect graph-specific settings (eg., in the case of
setting 'GIT_TEST_BLOOM_SETTINGS_MAX_CHANGED_PATHS').
In order to have an initialized value for these settings, move its
initialization to earlier in the commit-graph write.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'get_bloom_filter' takes a flag to control whether it will compute a
Bloom filter if the requested one is missing. In the next patch, we'll
add yet another parameter to this method, which would force all but one
caller to specify an extra 'NULL' parameter at the end.
Instead of doing this, split 'get_bloom_filter' into two functions:
'get_bloom_filter' and 'get_or_compute_bloom_filter'. The former only
looks up a Bloom filter (and does not compute one if it's missing,
thus dropping the 'compute_if_not_present' flag). The latter does
compute missing Bloom filters, with an additional parameter to store
whether or not it needed to do so.
This simplifies many call-sites, since the majority of existing callers
to 'get_bloom_filter' do not want missing Bloom filters to be computed
(so they can drop the parameter entirely and use the simpler version of
the function).
While we're at it, instrument the new 'get_or_compute_bloom_filter()'
with counters in the 'write_commit_graph_context' struct which store
the number of filters that we did and didn't compute, as well as filters
that were truncated.
It would be nice to drop the 'compute_if_not_present' flag entirely,
since all remaining callers of 'get_or_compute_bloom_filter' pass it as
'1', but this will change in a future patch and hence cannot be removed.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Many places in the code often need a pointer to the commit-graph's
'struct bloom_filter_settings', in which case they often take the value
from the top-most commit-graph.
In the non-split case, this works as expected. In the split case,
however, things get a little tricky. Not all layers in a chain of
incremental commit-graphs are required to themselves have Bloom data,
and so whether or not some part of the code uses Bloom filters depends
entirely on whether or not the top-most level of the commit-graph chain
has Bloom filters.
This has been the behavior since Bloom filters were introduced, and has
been codified into the tests since a759bfa9ee (t4216: add end to end
tests for git log with Bloom filters, 2020-04-06). In fact, t4216.130
requires that Bloom filters are not used in exactly the case described
earlier.
There is no reason that this needs to be the case, since it is perfectly
valid for commits in an earlier layer to have Bloom filters when commits
in a newer layer do not.
Since Bloom settings are guaranteed in practice to be the same for any
layer in a chain that has Bloom data, it is sufficient to traverse the
'->base_graph' pointer until either (1) a non-null 'struct
bloom_filter_settings *' is found, or (2) until we are at the root of
the commit-graph chain.
Introduce a 'get_bloom_filter_settings()' function that does just this,
and use it instead of purely dereferencing the top-most graph's
'->bloom_filter_settings' pointer.
While we're at it, add an additional test in t5324 to guard against code
in the commit-graph writing machinery that doesn't correctly handle a
NULL 'struct bloom_filter *'.
Co-authored-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Updates to the changed-paths bloom filter.
* ds/commit-graph-bloom-updates:
commit-graph: check all leading directories in changed path Bloom filters
revision: empty pathspecs should not use Bloom filters
revision.c: fix whitespace
commit-graph: check chunk sizes after writing
commit-graph: simplify chunk writes into loop
commit-graph: unify the signatures of all write_graph_chunk_*() functions
commit-graph: persist existence of changed-paths
bloom: fix logic in get_bloom_filter()
commit-graph: change test to die on parse, not load
commit-graph: place bloom_settings in context
The get_bloom_filter() method is a bit complicated in some parts where
it does not need to be. In particular, it needs to return a NULL filter
only when compute_if_not_present is zero AND the filter data cannot be
loaded from a commit-graph file. This currently happens by accident
because the commit-graph does not load changed-path Bloom filters from
an existing commit-graph when writing a new one. This will change in a
later patch.
Also clean up some style issues while we are here.
One side-effect of returning a NULL filter is that the filters that are
reported as "too large" will now be reported as NULL insead of length
zero. This case was not properly covered before, so add a test. Further,
remote the counting of the zero-length filters from revision.c and the
trace2 logs.
Helped-by: René Scharfe <l.s.r@web.de>
Helped-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In an earlier patch, multiple struct acccesses to `graph_pos` and
`generation` were auto-converted to multiple method calls.
Since the values are fixed and commit-slab access costly, we would be
better off with storing the values as a local variable and reusing it.
Signed-off-by: Abhishek Kumar <abhishekkumar8222@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We remove members `graph_pos` and `generation` from the struct commit.
The default assignments in init_commit_node() are no longer valid,
which is fine as the slab helpers return appropriate default values and
the assignments are removed.
We will replace existing use of commit->generation and commit->graph_pos
by commit_graph_data_slab helpers using
`contrib/coccinelle/commit.cocci'.
Signed-off-by: Abhishek Kumar <abhishekkumar8222@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git log -L..." now takes advantage of the "which paths are touched
by this commit?" info stored in the commit-graph system.
* ds/line-log-on-bloom:
line-log: integrate with changed-path Bloom filters
line-log: try to use generation number-based topo-ordering
line-log: more responsive, incremental 'git log -L'
t4211-line-log: add tests for parent oids
line-log: remove unused fields from 'struct line_log_data'
Code cleanup and typofixes
* ds/bloom-cleanup:
completion: offer '--(no-)patch' among 'git log' options
bloom: use num_changes not nr for limit detection
bloom: de-duplicate directory entries
Documentation: changed-path Bloom filters use byte words
bloom: parse commit before computing filters
test-bloom: fix usage typo
bloom: fix whitespace around tab length
The previous changes to the line-log machinery focused on making the
first result appear faster. This was achieved by no longer walking the
entire commit history before returning the early results. There is still
another way to improve the performance: walk most commits much faster.
Let's use the changed-path Bloom filters to reduce time spent computing
diffs.
Since the line-log computation requires opening blobs and checking the
content-diff, there is still a lot of necessary computation that cannot
be replaced with changed-path Bloom filters. The part that we can reduce
is most effective when checking the history of a file that is deep in
several directories and those directories are modified frequently. In
this case, the computation to check if a commit is TREESAME to its first
parent takes a large fraction of the time. That is ripe for improvement
with changed-path Bloom filters.
We must ensure that prepare_to_use_bloom_filters() is called in
revision.c so that the bloom_filter_settings are loaded into the struct
rev_info from the commit-graph. Of course, some cases are still
forbidden, but in the line-log case the pathspec is provided in a
different way than normal.
Since multiple paths and segments could be requested, we compute the
struct bloom_key data dynamically during the commit walk. This could
likely be improved, but adds code complexity that is not valuable at
this time.
There are two cases to care about: merge commits and "ordinary" commits.
Merge commits have multiple parents, but if we are TREESAME to our first
parent in every range, then pass the blame for all ranges to the first
parent. Ordinary commits have the same condition, but each is done
slightly differently in the process_ranges_[merge|ordinary]_commit()
methods. By checking if the changed-path Bloom filter can guarantee
TREESAME, we can avoid that tree-diff cost. If the filter says "probably
changed", then we need to run the tree-diff and then the blob-diff if
there was a real edit.
The Linux kernel repository is a good testing ground for the performance
improvements claimed here. There are two different cases to test. The
first is the "entire history" case, where we output the entire history
to /dev/null to see how long it would take to compute the full line-log
history. The second is the "first result" case, where we find how long
it takes to show the first value, which is an indicator of how quickly a
user would see responses when waiting at a terminal.
To test, I selected the paths that were changed most frequently in the
top 10,000 commits using this command (stolen from StackOverflow [1]):
git log --pretty=format: --name-only -n 10000 | sort | \
uniq -c | sort -rg | head -10
which results in
121 MAINTAINERS
63 fs/namei.c
60 arch/x86/kvm/cpuid.c
59 fs/io_uring.c
58 arch/x86/kvm/vmx/vmx.c
51 arch/x86/kvm/x86.c
45 arch/x86/kvm/svm.c
42 fs/btrfs/disk-io.c
42 Documentation/scsi/index.rst
(along with a bogus first result). It appears that the path
arch/x86/kvm/svm.c was renamed, so we ignore that entry. This leaves the
following results for the real command time:
| | Entire History | First Result |
| Path | Before | After | Before | After |
|------------------------------|--------|--------|--------|--------|
| MAINTAINERS | 4.26 s | 3.87 s | 0.41 s | 0.39 s |
| fs/namei.c | 1.99 s | 0.99 s | 0.42 s | 0.21 s |
| arch/x86/kvm/cpuid.c | 5.28 s | 1.12 s | 0.16 s | 0.09 s |
| fs/io_uring.c | 4.34 s | 0.99 s | 0.94 s | 0.27 s |
| arch/x86/kvm/vmx/vmx.c | 5.01 s | 1.34 s | 0.21 s | 0.12 s |
| arch/x86/kvm/x86.c | 2.24 s | 1.18 s | 0.21 s | 0.14 s |
| fs/btrfs/disk-io.c | 1.82 s | 1.01 s | 0.06 s | 0.05 s |
| Documentation/scsi/index.rst | 3.30 s | 0.89 s | 1.46 s | 0.03 s |
It is worth noting that the least speedup comes for the MAINTAINERS file
which is
* edited frequently,
* low in the directory heirarchy, and
* quite a large file.
All of those points lead to spending more time doing the blob diff and
less time doing the tree diff. Still, we see some improvement in that
case and significant improvement in other cases. A 2-4x speedup is
likely the more typical case as opposed to the small 5% change for that
file.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As diff_tree_oid() computes a diff, it will terminate early if the
total number of changed paths is strictly larger than max_changes.
This includes the directories that changed, not just the file paths.
However, only the file paths are reflected in the resulting diff
queue's "nr" value.
Use the "num_changes" from diffopt to check if the diff terminated
early. This is incredibly important, as it can result in incorrect
filters! For example, the first commit in the Linux kernel repo
reports only 471 changes, but since these are nested inside several
directories they expand to 513 "real" changes, and in fact the
total list of changes is not reported. Thus, the computed filter
for this commit is incorrect.
Demonstrate the subtle difference by using one fewer file change
in the 'get bloom filter for commit with 513 changes' test. Before,
this edited 513 files inside "bigDir" which hit this inequality.
However, dropping the file count by one demonstrates how the
previous inequality was incorrect but the new one is correct.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When computing a changed-path Bloom filter, we need to take the
files that changed from the diff computation and extract the parent
directories. That way, a directory pathspec such as "Documentation"
could match commits that change "Documentation/git.txt".
However, the current code does a poor job of this process. The paths
are added to a hashmap, but we do not check if an entry already
exists with that path. This can create many duplicate entries and
cause the filter to have a much larger length than it should. This
means that the filter is more sparse than intended, which helps the
false positive rate, but wastes a lot of space.
Properly use hashmap_get() before hashmap_add(). Also be sure to
include a comparison function so these can be matched correctly.
This has an effect on a test in t0095-bloom.sh. This makes sense,
there are ten changes inside "smallDir" so the total number of
paths in the filter should be 11. This would result in 11 * 10 bits
required, and with 8 bits per byte, this results in 14 bytes.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When computing changed-path Bloom filters for a commit, we need to
know if the commit has a parent or not. If the commit is not parsed,
then its parent pointer will be NULL.
As far as I can tell, the only opportunity to reach this code
without parsing the commit is inside "test-tool bloom
get_filter_for_commit" but it is best to be safe.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* We need a `final_new_line` to make our source code as text file, per
POSIX and C specification.
* `bloom_filters` should be limited to interal linkage only
Signed-off-by: Đoàn Trần Công Danh <congdanhqx@gmail.com>
Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix alignment issues that were likely introduced due to an editor
using tab lengths of 4 instead of 8.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The changed-path Bloom filters record an entry in the filter for
every path that was changed. This includes every add and delete,
regardless of whether a rename was detected. Detecting renames
causes significant performance issues, but also will trigger
downloading missing blobs in partial clone.
The simple fix is to disable rename detection when computing a
changed-path Bloom filter. This should already be disabled by
default, but it is good to explicitly enforce the intended
behavior.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Revision walk will now use Bloom filters for commits to speed up
revision walks for a particular path (for computing history for
that path), if they are present in the commit-graph file.
We load the Bloom filters during the prepare_revision_walk step,
currently only when dealing with a single pathspec. Extending
it to work with multiple pathspecs can be explored and built on
top of this series in the future.
While comparing trees in rev_compare_trees(), if the Bloom filter
says that the file is not different between the two trees, we don't
need to compute the expensive diff. This is where we get our
performance gains. The other response of the Bloom filter is '`:maybe',
in which case we fall back to the full diff calculation to determine
if the path was changed in the commit.
We do not try to use Bloom filters when the '--walk-reflogs' option
is specified. The '--walk-reflogs' option does not walk the commit
ancestry chain like the rest of the options. Incorporating the
performance gains when walking reflog entries would add more
complexity, and can be explored in a later series.
Performance Gains:
We tested the performance of `git log -- <path>` on the git repo, the linux
and some internal large repos, with a variety of paths of varying depths.
On the git and linux repos:
- we observed a 2x to 5x speed up.
On a large internal repo with files seated 6-10 levels deep in the tree:
- we observed 10x to 20x speed ups, with some paths going up to 28 times
faster.
Helped-by: Derrick Stolee <dstolee@microsoft.com
Helped-by: SZEDER Gábor <szeder.dev@gmail.com>
Helped-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add logic to
a) parse Bloom filter information from the commit graph file and,
b) re-use existing Bloom filters.
See Documentation/technical/commit-graph-format for the format in which
the Bloom filter information is written to the commit graph file.
To read Bloom filter for a given commit with lexicographic position
'i' we need to:
1. Read BIDX[i] which essentially gives us the starting index in BDAT for
filter of commit i+1. It is essentially the index past the end
of the filter of commit i. It is called end_index in the code.
2. For i>0, read BIDX[i-1] which will give us the starting index in BDAT
for filter of commit i. It is called the start_index in the code.
For the first commit, where i = 0, Bloom filter data starts at the
beginning, just past the header in the BDAT chunk. Hence, start_index
will be 0.
3. The length of the filter will be end_index - start_index, because
BIDX[i] gives the cumulative 8-byte words including the ith
commit's filter.
We toggle whether Bloom filters should be recomputed based on the
compute_if_not_present flag.
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When computing the changed-paths bloom filters for the commit-graph,
we limit the size of the filter by restricting the number of paths
in the diff. Instead of computing a large diff and then ignoring the
result, it is better to halt the diff computation early.
Create a new "max_changes" option in struct diff_options. If non-zero,
then halt the diff computation after discovering strictly more changed
paths. This includes paths corresponding to trees that change.
Use this max_changes option in the bloom filter calculations. This
reduces the time taken to compute the filters for the Linux kernel
repo from 2m50s to 2m35s. On a large internal repository with ~500
commits that perform tree-wide changes, the time reduced from
6m15s to 3m48s.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add the core implementation for computing Bloom filters for
the paths changed between a commit and it's first parent.
We fill the Bloom filters as (const char *data, int len) pairs
as `struct bloom_filters" within a commit slab.
Filters for commits with no changes and more than 512 changes,
is represented with a filter of length zero. There is no gain
in distinguishing between a computed filter of length zero for
a commit with no changes, and an uncomputed filter for new commits
or for commits with more than 512 changes. The effect on
`git log -- path` is the same in both cases. We will fall back to
the normal diffing algorithm when we can't benefit from the
existence of Bloom filters.
Helped-by: Jeff King <peff@peff.net>
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Jakub Narębski <jnareb@gmail.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Introduce the constructs for Bloom filters, Bloom filter keys
and Bloom filter settings.
For details on what Bloom filters are and how they work, refer
to Dr. Derrick Stolee's blog post [1]. It provides a concise
explanation of the adoption of Bloom filters as described in
[2] and [3].
Implementation specifics:
1. We currently use 7 and 10 for the number of hashes and the
size of each entry respectively. They served as great starting
values, the mathematical details behind this choice are
described in [1] and [4]. The implementation, while not
completely open to it at the moment, is flexible enough to allow
for tweaking these settings in the future.
Note: The performance gains we have observed with these values
are significant enough that we did not need to tweak these
settings. The performance numbers are included in the cover letter
of this series and in the commit message of the subsequent commit
where we use Bloom filters to speed up `git log -- path`.
2. As described in [1] and [3], we do not need 7 independent hashing
functions. We use the Murmur3 hashing scheme, seed it twice and
then combine those to procure an arbitrary number of hash values.
3. The filters will be sized according to the number of changes in
each commit, in multiples of 8 bit words.
[1] Derrick Stolee
"Supercharging the Git Commit Graph IV: Bloom Filters"
https://devblogs.microsoft.com/devops/super-charging-the-git-commit-graph-iv-Bloom-filters/
[2] Flavio Bonomi, Michael Mitzenmacher, Rina Panigrahy, Sushil Singh, George Varghese
"An Improved Construction for Counting Bloom Filters"
http://theory.stanford.edu/~rinap/papers/esa2006b.pdfhttps://doi.org/10.1007/11841036_61
[3] Peter C. Dillinger and Panagiotis Manolios
"Bloom Filters in Probabilistic Verification"
http://www.ccs.neu.edu/home/pete/pub/Bloom-filters-verification.pdfhttps://doi.org/10.1007/978-3-540-30494-4_26
[4] Thomas Mueller Graf, Daniel Lemire
"Xor Filters: Faster and Smaller Than Bloom and Cuckoo Filters"
https://arxiv.org/abs/1912.08258
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Jakub Narębski <jnareb@gmail.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In preparation for computing changed paths Bloom filters,
implement the Murmur3 hash algorithm as described in [1].
It hashes the given data using the given seed and produces
a uniformly distributed hash value.
[1] https://en.wikipedia.org/wiki/MurmurHash#Algorithm
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Helped-by: Szeder Gábor <szeder.dev@gmail.com>
Reviewed-by: Jakub Narębski <jnareb@gmail.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Elijah Newren
Taylor Blau
Derrick Stolee
Junio C Hamano
Abhishek Kumar
Đoàn Trần Công Danh
Garima Singh