On Mon, Jun 15, 2020 at 10:46:39AM -0600, Taylor Blau wrote: > Hi Patrick, > > Sorry for the slow response. I was out of the office last week and am > only just now getting a chance to catch up on the backlog of emails that > I missed. No worries! > On Thu, Jun 04, 2020 at 09:36:32AM +0200, Patrick Steinhardt wrote: > > On Wed, Jun 03, 2020 at 10:51:42AM -0600, Taylor Blau wrote: > > > Hi Patrick, > > > > > > On Wed, Jun 03, 2020 at 02:27:50PM +0200, Patrick Steinhardt wrote: > > > > In order to test the impact on the case where we don't have any > > > > "reference-transaction" hook installed in the repository, this commit > > > > introduces a new performance test for git-update-refs(1). Run against an > > > > empty repository, it produces the following results: > > > > > > > > Test HEAD~ HEAD > > > > ------------------------------------------------------------------------------ > > > > 1400.2: update existing reference 2.05(1.58+0.54) 2.08(1.58+0.57) +1.5% > > > > 1400.3: create and destroy reference 1.79(1.38+0.49) 1.82(1.39+0.51) +1.7% > > > > > > > > So the overhead is around ~1.5%. Given that git-update-refs(1) is a > > > > near-direct wrapper around reference transactions, there likely is no > > > > other command that is impacted much worse than this. > > > > > > This is a serious performance regression that is worth considering. For > > > example, a 1.5% slow-down on average in reference transactions would > > > cause be enough for me to bisect the regression down to see what > > > changed. > > > > > > What are ways that this could be avoided? I bet that this would work > > > quite well with Emily's work on hooks, where we could check in the > > > configuration first whether a hook is even configured. > > > > > > Could this be integrated with that? If not, could you cache the result > > > of whether or not the hook exists, and/or implement some mechanism to > > > say something like, for eg., "only run reference transaction hooks > > > core.blah = 1" is true? > > > > I wasn't aware of her work until now, so I'll take a look. > > > > Meanwhile, I toyed around with performance and tried out two different > > caching mechanisms: > > > > - map-cache: `find_hook()` uses a map of hook names mapped to their > > resolved hook path (or `NULL` if none was found). This is a > > generic mechanism working across all hooks, but has some overhead > > in maintaining the map. > > > > - reftx-cache: `run_transaction_hook()` caches the path returned by > > `find_hook()`. It's got less overhead as it only caches the path, > > but obviously only applies to the reference-transaction hook. > > > > In theory, we could go even further and cache the hook's file > > descriptor, executing it via fexecve(3P) whenever it's required, but I > > didn't go down that route yet. This would also solve the atomicity > > issue, though, if the administrator replaces the reference-transactions > > hook halfway through the transaction. > > > > Benchmarking results are mixed, mostly in the sense that I can choose > > which run of the benchmarks I take in order to have my own work look > > better or worse, as timings vary quite a lot between runs. Which > > probably hints at the fact that the benchmarks themselves aren't really > > reliable. The issue is that a single git-update-ref(1) run finishes so > > quick that it's hard to measure with our benchmarks, but spawning > > thousands of them to get something different than 0.0s very much depends > > on the operating system and thus fluctuates. On the other hand, if we > > only spawn a single git-update-refs and have it perform a few thousand > > ref updates, overhead of the hook will not show at all as it is only > > executed once per prepare/commit of the transaction. > > > > The following timings are for the case where we execute > > > > git update-ref refs/heads/update-branch PRE POST && > > git update-ref refs/heads/update-branch POST PRE > > > > respectively > > > > git update-ref refs/heads/new POST && > > git update-ref -d refs/heads/new > > > > a thousand times: > > > > Test master ref-hook-no-cache ref-hook-map-cache ref-hook-reftx-cache > > ------------------------------------------------------------------------------------------------------------------------------ > > 1400.2: update existing reference 1.96(1.50+0.53) 2.00(1.54+0.53) +2.0% 2.02(1.54+0.55) +3.1% 1.98(1.52+0.52) +1.0% > > 1400.3: create and destroy reference 1.74(1.33+0.49) 1.77(1.38+0.47) +1.7% 1.77(1.36+0.48) +1.7% 1.76(1.35+0.49) +1.1% > > Huh. It is super interesting (to me, at least) that caching the set of > hooks that are on disk and should be run makes this *slower* than > without the cache. What's going on there? In p1400.2, I'd expect to see > 'ref-hook-map-cache' attain at most a 2.0% slow-down, plus a little bit > to process the hook when it is there, but not much more than that. I think the issue is that a single git-update-ref(1) invocation does so little work that allocating the hashmap and inserting the hook already has noticeable impact on the program's runtime. E.g. in above benchmark, a single call to git-update-ref in p1400.2 takes roughly 0.002s on my machine. You also see this by the fact that doing a single stat(3P) call as introduced by my patch adds a 1% performance penalty already, and with the map cache we still have to do this single stat(3P) call in addition to the dynamic memory allocations for the map and insertion of the hook. > I think that this just seems a little contrived to me. I can understand > why server administrators may want this feature, but the general > user-base of Git doesn't seem to stand to benefit much from this change > (in my own mind, at least). That's true for several hooks we have, though. > So... I'm not sure where this leaves us. Maybe a 2.0% slow-down on an > already fast 'git update-ref' invocation for the average user won't be > noticeable. It certainly *will* be noticeable to administrators who > processes a much higher volume of such transactions. I think we should keep in mind that it's git-update-ref(1) we're talking about, which is a nearly direct wrapper around reference transactions. The 1% perfomance hit is thus the worst case that can happen, as there is no other tool that does as little work around the reftx as this one. For any other tool, I imagine the performance hit to be at worst the same (e.g. git-branch(1)) or not noticeable at all because a single stat(3P) call will be drowned out by other things (e.g. git-clone(1)). That's not to say that nobody will be impacted by this change, I bet there are setups that make heavy use of git-update-ref(1). > > For such a short-lived program like git-update-refs(1), one can see that > > the overhead of using a map negatively impacts performance compared to > > the no-cache case. But using the reftx-cache roughly cuts the overhead > > in half as expected, as we only need to look up the hook once instead of > > twice. > > > > And here's the results if we use a single `git update-ref --stdin` with a > > thousand reference updates at once: > > > > Test master ref-hook-no-cache ref-hook-map-cache ref-hook-reftx-cache > > ------------------------------------------------------------------------------------------------------------------------ > > 1400.2: git update-ref --stdin 0.21(0.09+0.12) 0.21(0.07+0.14) +0.0% 0.21(0.07+0.13) +0.0% 0.21(0.07+0.13) +0.0% > > > > As expected, there's nothing much to be seen here because there's only a > > single transaction for all ref updates and, as a result, at most two > > calls to `access(refhook_path, X_OK)`. > > Better, but I have to imagine that real-world usage will look much more > like a thousand tiny transactions than one large one. Most likely, yes. Doesn't happen too often that one updates multiple references at once. Patrick
Attachment:
signature.asc
Description: PGP signature
