D6430: rust-discovery: using from Python code

Closed by commit rHG4d20b1fe8a72: rust-discovery: using from Python code (authored by gracinet). This revision was automatically updated to reflect the committed changes. This revision was not accepted when it landed; it landed in state "Needs Review". REPOSITORY rHG Mercurial CHANGES SINCE

D6430: rust-discovery: using from Python code

durin42 added a comment. I'm not thrilled with this - it's a lot of code, and I think there are some simpler options that haven't been explored. For example, clients could use the logexchange infrastructure to prime the sampling process and do a lot better. That said, it seems "fine"

D6430: rust-discovery: using from Python code

pulkit added a comment. I tested this series on our repository and found the following speedup in a case when there are 6k heads and 15k csets missing: Before: ! wall 3.281062 comb 3.28 user 3.20 sys 0.08 (best of 25) ! wall 4.941540 comb 4.94 user 4.86 sys

D6430: rust-discovery: using from Python code

marmoute added a comment. In the couple of last version, we already saved minutes in real life use case simply by improving pure CPU processing time client side. Can you elaborate on what other types of evidences you need to convince you there exist CPU bound case for discovery? REPOSITORY

D6430: rust-discovery: using from Python code

martinvonz added a comment. In D6430#96225 , @marmoute wrote: > The ×2.5 speedup is the kind of things that motivate this series. Even if the mozilla-unified/mozilla-try is just and example it triggers the kind of pathological case we

D6430: rust-discovery: using from Python code

marmoute added a comment. The ×2.5 speedup is the kind of things that motivate this series. Even if the mozilla-unified/mozilla-try is just and example it triggers the kind of pathological case we encounter in real life: large undecided set. We keep finding such pathological case from time

D6430: rust-discovery: using from Python code

gracinet added a comment. Nobody has reacted yet to @martinvonz call for other opinions, can someone look at it? Let me restate that the series makes a good difference when the sampling part of the discovery process actually kicks in, and it does that roughly three times faster. In

D6430: rust-discovery: using from Python code

martinvonz added a comment. > This example is not a goal in itself, Hmm, I would actually consider that the strongest selling point of this series. The other case (the 6.8% improvement) doesn't seem enough to justify all this code. > but it showcases several different > areas in

D6430: rust-discovery: using from Python code

gracinet added a comment. Instability fixed, and I took the opportunity to rebase again to leverage `policy.importrust`, that's been queued yesterday (rHGf7385ed775a8 ) I added new performance measurements for

D6430: rust-discovery: using from Python code

gracinet edited the summary of this revision. gracinet updated this revision to Diff 15487. REPOSITORY rHG Mercurial CHANGES SINCE LAST UPDATE https://phab.mercurial-scm.org/D6430?vs=15473=15487 CHANGES SINCE LAST ACTION https://phab.mercurial-scm.org/D6430/new/ REVISION DETAIL

D6430: rust-discovery: using from Python code

gracinet added a comment. Update of the whole series done but, ended up in an inconsistent state, and I have to go. Please don't act on it until I give the signal it's ready (sorry for inconvenience) REPOSITORY rHG Mercurial CHANGES SINCE LAST ACTION

D6430: rust-discovery: using from Python code

gracinet edited the summary of this revision. gracinet updated this revision to Diff 15473. REPOSITORY rHG Mercurial CHANGES SINCE LAST UPDATE https://phab.mercurial-scm.org/D6430?vs=15230=15473 CHANGES SINCE LAST ACTION https://phab.mercurial-scm.org/D6430/new/ REVISION DETAIL

D6430: rust-discovery: using from Python code

martinvonz added a comment. I think this series needs to be updated for https://phab.mercurial-scm.org/D2647. I'm curious to see the timing numbers after that patch (I can get those myself once the series is updated if you're tired of running the perf commands). REPOSITORY rHG Mercurial

D6430: rust-discovery: using from Python code

gracinet created this revision. Herald added a subscriber: mercurial-devel. Herald added a reviewer: hg-reviewers. REVISION SUMMARY As previously done in other topics, the Rust version is used if it's been built. The version fully in Rust of the partialdiscovery class has the performance