2022年1月14日(金) 2:32 Scott Violet <s...@google.com>: > > > On Thu, Jan 13, 2022 at 1:41 AM Camillo Bruni <cbr...@google.com> wrote: > >> Yes ExecuteMicroSeconds is per call. >> >> On Thu, 13 Jan 2022 at 01:36, Scott Violet <s...@google.com> wrote: >> >>> If I'm reading it right, ExecuteMicroSeconds is per call, not an >>> aggregate. >>> >> Correct, this is per individual call, aggregation would have to happen >> separately >> >> >>> Additionally as a renderer be running multiple pages trying to track >>> this at the v8 level wouldn't be that accurate. I'm wondering if the better >>> place to track this is in blink's bindings? >>> >> Yes, from what I see this is correct, the histograms are aggregated per >> process. >> There is also the LongTask API where we accumulate the current v8 execute >> <https://source.chromium.org/chromium/chromium/src/+/main:v8/src/logging/counters-scopes.h;drc=11b5b115b432e67d5c2a572d11cebe6a81ced05e;bpv=1;bpt=1;l=176> >> times (if the histogram is enabled) per isolate (not per process). >> Maybe that is something to extend as well without having to add another >> execution-time tracker with additional overhead? >> > > That seems like exactly what I want! I'll investigate it a bit more. > > >> >> >> Specifically in the code generation it could presumably track the times. >>> Yuki, how realistic is this? >>> >> What do you mean by "in the code generation"? >> > > Sorry for not being clear. I'm referring to the IDL compiler in blink that > generates the c++ bindings code: > https://chromium.googlesource.com/chromium/src/+/refs/heads/main/third_party/blink/renderer/bindings/IDLCompiler.md > . >
Generated bindings code usually doesn't call heavy (long running) V8 APIs. Plus, the typical caller-callee relationship is as follows. Webpage's JavaScript code (run on V8) => Web APIs (stub in V8) => Generated bindings => Blink implementation So, I'm not sure if the bindings code generator could help the measurement anyway. If you want to measure the overhead of "Generated bindings" above, then it's the best place though. Long-running V8 APIs (such as v8::Context::FromSnapshot and v8::Context::New) are called from non-generated Blink code (and you have already tried measuring it). As another random idea, you might be interested in measuring the time of the "onload" event handler, which usually runs a lot of (web author) JS code (in V8). Hope V8.ExecuteMicroSeconds fits your purpose. Cheers, Yuki Shiino > -Scott > > >> >> >>> >>> Thanks, >>> >>> -Scott >>> >>> On Wed, Jan 12, 2022 at 4:16 PM Scott Violet <s...@google.com> wrote: >>> >>>> Thanks for the pointer! >>>> >>>> I noticed the histogram has the caveat 'excluding callbacks.' What does >>>> that mean in this context? >>>> >>>> Thanks, >>>> >>>> -Scott >>>> >>>> On Mon, Jan 10, 2022 at 1:32 AM Camillo Bruni <cbr...@google.com> >>>> wrote: >>>> >>>>> Getting the detailed V8 times comes at a measurable overhead, this >>>>> piggybacking on the existing V8ExecuteMicroseconds data from the >>>>> persistent >>>>> V8SlowHistograms trial is most likely your best option here to limit >>>>> negative side effects from multiple trials slowing down the user. >>>>> >>>>> On Fri, 7 Jan 2022, 09:57 Leszek Swirski, <lesz...@chromium.org> >>>>> wrote: >>>>> >>>>>> +Camillo Bruni <cbr...@google.com> >>>>>> We have a V8.ExecuteMicroSeconds histogram which is enabled for 50% >>>>>> of canary and dev (via the V8SlowHistograms >>>>>> <https://source.chromium.org/chromium/chromium/src/+/main:gin/gin_features.cc;l=108;drc=26479386abe0c05c9661f6d8c1412dde636d8ecb> >>>>>> feature >>>>>> flag). Maybe you could somehow intercept writes to this histogram before >>>>>> FCP? >>>>>> >>>>>> On Thu, Jan 6, 2022 at 2:35 PM Yuki Shiino <yukishi...@chromium.org> >>>>>> wrote: >>>>>> >>>>>>> +v8-dev@googlegroups.com <v8-dev@googlegroups.com> >>>>>>> >>>>>>> >>>>>>> 2022年1月5日(水) 9:05 Scott Violet <s...@google.com>: >>>>>>> >>>>>>>> Ping. Any suggestions? >>>>>>>> >>>>>>>> Thanks, >>>>>>>> >>>>>>>> -Scott >>>>>>>> >>>>>>>> On Tue, Dec 21, 2021 at 5:15 AM Yuki Shiino < >>>>>>>> yukishi...@chromium.org> wrote: >>>>>>>> >>>>>>>>> +platform-architecture-dev >>>>>>>>> <platform-architecture-...@chromium.org> >>>>>>>>> >>>>>>>>> 2021年12月21日(火) 9:45 Scott Violet <s...@google.com>: >>>>>>>>> >>>>>>>>>> Hello, >>>>>>>>>> >>>>>>>>>> I'm interested in tracking how much time we spend in v8 prior to >>>>>>>>>> fcp. Any idea if we have a metric for this? Presumably such a metric >>>>>>>>>> would >>>>>>>>>> involve instrumenting bindings, but I'm not positive there. >>>>>>>>>> >>>>>>>>> >>>>>>>>> I know only about RuntimeCallStats >>>>>>>>> <https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/platform/bindings/RuntimeCallStats.md> >>>>>>>>> (V8 >>>>>>>>> looks having the same concept maybe around here >>>>>>>>> <https://source.chromium.org/chromium/chromium/src/+/main:v8/src/logging/runtime-call-stats.h;bpv=1;bpt=1;l=570?q=RuntimeCallStats%20case:yes&ss=chromium>). >>>>>>>>> However, if you're interested in measuring it in the wild, >>>>>>>>> RuntimeCallStats >>>>>>>>> wouldn't be a good fit. >>>>>>>>> >>>>>>>>> Anyone else would know a better idea? >>>>>>>>> >>>>>>>>> Cheers, >>>>>>>>> Yuki Shiino >>>>>>>>> >>>>>>>>> Thanks, >>>>>>>>>> >>>>>>>>>> -Scott >>>>>>>>>> >>>>>>>>> -- >>>>>>> -- >>>>>>> v8-dev mailing list >>>>>>> v8-dev@googlegroups.com >>>>>>> http://groups.google.com/group/v8-dev >>>>>>> --- >>>>>>> You received this message because you are subscribed to the Google >>>>>>> Groups "v8-dev" group. >>>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>>> send an email to v8-dev+unsubscr...@googlegroups.com. >>>>>>> To view this discussion on the web visit >>>>>>> https://groups.google.com/d/msgid/v8-dev/CAN0uC_TCyPxXdVGbAmh7JBR3e7UHxXYXtuW-yUpM9Q1hSbXBKQ%40mail.gmail.com >>>>>>> <https://groups.google.com/d/msgid/v8-dev/CAN0uC_TCyPxXdVGbAmh7JBR3e7UHxXYXtuW-yUpM9Q1hSbXBKQ%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>>>> . >>>>>>> >>>>>> Camillo Bruni | Software Engineer, V8 | Google Germany GmbH | >>>>>> Erika-Mann >> Str. 33, 80636 München >> >> Registergericht und -nummer: Hamburg, HRB 86891 | Sitz der Gesellschaft: >> Hamburg | Geschäftsführer: Paul Manicle, Halimah DeLaine Prado >> >> Diese E-Mail ist vertraulich. Falls Ssie diese fälschlicherweise erhalten >> haben sollten, leiten Sie diese bitte nicht an jemand anderes weiter, >> löschen Sie alle Kopien und Anhänge davon und lassen Sie mich bitte wissen, >> dass die E-Mail an die falsche Person gesendet wurde. This e-mail is >> confidential. If you received this communication by mistake, please don't >> forward it to anyone else, please erase all copies and attachments, and >> please let me know that it has gone to the wrong person. >> > -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to v8-dev+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/v8-dev/CAN0uC_RgqetOS_ptbkrBkEQf22M5SpjffT3N3nthjqBV%3Dn-iBA%40mail.gmail.com.
