Development of a performance tracking tool for Mercurial

[Philippe Pepiot]

Hi,

I rewrote the benchmarks suite (https://hg.logilab.org/review/hgperf) 
using contrib/perf.py from tested version of mercurial while keeping old 
helper functions to write plain benchmarks easily.

Here is a summary of the current status of the project:

All my patches are merged upstream into AirSpeed Velocity, especially 
ordering commit by graph revision instead of date which was the big one.

I'm running a demo https://jenkins.philpep.org/hgperf/ that benchmark 
all revisions added to the repository since one month, with sparse 
earlier commits too (since 3.4). One interesting result is that I'm 
benchmarking mercurial against the same mercurial repository, i.e. the 
reference repository is pulled, this explain all the small detected 
regressions. So benchmarking a moving repository should be avoided 
(seems trivial to say but I wasn't sure how this will be visible on the 
graphs).

ASV is actively maintained some new features are incoming:

https://github.com/spacetelescope/asv/pull/447 Regressions notifications 
using atom feed, this one is part of the plan. If people prefer to 
receive notifications via email, there are tools like rss2email or 
online services to transform new feed items into an email.

https://github.com/spacetelescope/asv/pull/437 This one (WIP) is quite 
interesting because it track improvement too and provide a useful 
summary page.

In the plan, 
https://www.mercurial-scm.org/wiki/PerformanceTrackingSuitePlan , the 
next topic is using unit test execution time as benchmark result and 
related topics (Annotation system, handle changed tests, scenario based 
benchmarks). Most of them are shell style (.t) tests that spawn a lot of 
hg subprocesses and I wonder if it's relevant to use them as benchmark 
result because they work on very small repositories (generated during 
tests) and I think most of the time is spent on hg startup. Maybe we 
could detect startup time regression here, but I think we should miss 
others regressions that will be insignificant comparing to the whole 
test duration. I wonder if it's worth to put efforts on this topic, but 
I maybe missing something, do you have already used unit test execution 
time to compare performance changes introduced by a commit ?

Some other topics might require more work, like improving the web 
interface, for instance having a "per commit" view that summarize 
regressions and improvements introduced by a single commit, integrate 
the benchmark suite in mercurial, write more benchmarks, I could start 
by cover all combination of benchmarks/options offered in 
contrib/perf.py. Another topic could be extensions benchmarking. I put 
the development in pause while waiting your feedback, hoping (or not ;)) 
they will be some real regressions detected soon to validate the tool, 
I'll add the atom feed once it get merged in ASV.

Cheers,

On 05/26/2016 11:57 AM, Philippe Pepiot wrote:
>
>
> On 05/23/2016 11:38 PM, Kevin Bullock wrote:
>> (resurrecting this thread now that I've had a closer look at the plan)
>>
>
> Thanks !
>
>>> >On Apr 12, 2016, at 05:03, Philippe 
>>> Pepiot<philippe.pepiot at logilab.fr>  wrote:
>>> >
>>> >[...]
>>> >Now I've a question about writing and maintaining benchmark code as 
>>> we have multiple choices here:
>>> >
>>> >1) Use mercurial internal API (benefits: unlimited possibilities 
>>> without modifying mercurial and we can write backward compatible 
>>> benchmark with some 'if' statements and benchmarks older versions, 
>>> profits all ASV features (profiling, memory benchmarks etc). 
>>> Drawbacks: duplicate code with contrib/perf.py, will break on 
>>> internal API changes, need more maintenance and more code to 
>>> write/keep backward compatible benchmarks).
>>> >
>>> >2) Use contrib/perf.py extension from the benchmarked version of 
>>> mercurial (benefits: de facto backward compatible, drawbacks: 
>>> limited to what the tool can do in previous versions)
>>> >
>>> >3) Use contrib/perf.py extension from the latest version of 
>>> mercurial (benefits: no duplicate code, easier maintenance, new 
>>> benchmarks profits to both tools. Drawbacks: not backward compatible 
>>> for now (it works only for >= 3.7 versions)). We could also 
>>> implement some glue code, either in the tracking tool or in 
>>> contrib/perf.py, to list available benchmarks and theirs parameters.
>>> >
>>> >At this stage of the project my advice it to use 1), but we could 
>>> also have a mix of 1) and 3). It depend on how fast are internal api 
>>> changes and on your short/mid/long term objectives on the level of 
>>> integration of the tracking tool.
>> My first instinct tells me we should use approach #2, and teach the 
>> benchmark suite what Mercurial release a given perf* command first 
>> appeared in. This can be automated by reading the output of `hg help 
>> -e perf`.
>>
>> #1 sounds like a particularly bad idea given the deliberately high 
>> level of churn in our internal APIs. If we try to maintain an 
>> external tool that also tries to remember the details of that churn, 
>> I'm pretty sure it will rot and fall out of use in short order.
>>
>> #3 has a similar (though lesser) disadvantage. I also don't see a 
>> strong need to run new perf checks against old versions, with the 
>> exception of perhaps the current stable release.
>>
>> All that said, please convince me I'm wrong.:)
>>
>> Regarding ASV: is there any way we can use it to instrument our code 
>> (including profiling and memory tracking) without having it call into 
>> the internals? Specifically, is there a way that we can either feed 
>> it metrics that we output from the perf extension, or integrate it 
>> into perf.py such that it gets optionally imported (similar to how 
>> you can optionally use ipdb via the --debugger flag, wired up in 
>> dispatch.py)?
>
> Ok if the internal API used in perf.py often change we can exclude #1 
> (at least until the benchmark suite is included into mercurial 
> repository).
>
> Nevertheless #1 can still be useful if you want to bisect a old 
> regression (and have no relevant benchmark code on the bisect range), 
> this can be done by just transforming the perf.py benchmark into a 
> regular asv benchmark (quite easy step), handle potential backward 
> compatibility then run the bisect.
>
> With #2 we cannot fix potential "bugs" in perf.py when benchmarking 
> old revisions. For instance if an internal change alter a benchmark 
> that is no more computing the same thing than above and it's fixed 
> later on, you will keep having wrong values even if you re-run the 
> benchmarks (but this is not a big issue).
>
> About #3, currently perf.py is not backward compatible with < 3.7 
> versions but it seems related to the extension code (not benchmark 
> code). After embedding commands.formatteropts and 
> commands.debugrevlogopts and remove the norepo parameter from 
> perflrucache, I can run some benchmarks on 3.0 versions.
>
> For now, #2 (and #3) can be achieved with ASV by writing a "track" 
> benchmark that run the given perf.py command in a subprocess and 
> return the value.
>
> To get a closer integration (and enable profiling features of ASV) we 
> could write a module that declare benchmarks (setup code and body 
> code) and let both asv and perf.py use it. Actually asv benchmark 
> themselves could be a candidate because this is just declarative code 
> (functions and class with naming convention and "setup" attribute on 
> the function), there is no dependency on asv here.
>
> Another idea could be to call perf.py commands programmatically in ASV 
> and eventually monkeypatch the gettimer() function to transform the 
> command into a regular asv "time benchmark".
>
> As a first step, I think we can go for the simplest solution (ie. #2) 
> and keep in mind a future inclusion (or dedicated api) in mercurial.
>

-- 
Philippe Pepiot
https://www.logilab.fr

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mercurial-scm.org/pipermail/mercurial-devel/attachments/20160627/fb147c3f/attachment.html>