On Mon, Feb 27, 2023 at 9:19 AM Davidlohr Bueso [off-list ref] wrote:

On Fri, 24 Feb 2023, freak07 wrote:

quoted

Here are some measurements from a Pixel 7 Pro that´s running a kernel either with the Per-VMA locks patchset or without.
If there´s interest I can provide results of other specific apps as well.

Results are from consecutive cold app launches issued with "am start" command spawning the main activity of Slack Android app.

https://docs.google.com/spreadsheets/d/1ktujfcyDmIJoQMWsoizGIE-0A_jMS9VMw_seehUY9s0/edit?usp=sharing

There´s quite a noticeable improvement, as can be seen in the graph. The results were reproducible.

Thanks for sharing. I am not surprised - after all, per-vma locks narrow some of the performance gaps
between vanilla and speculative pfs, with less complexity (albeit this is now a 35 patch series :).

Yes, depending on the specific app we would expect some launch time
improvement (in this case average improvement is 7%). Thanks for
sharing the numbers!
I'll be posting v4 today, which is a 33 patch series now, so a bit better :)
Thanks,
Suren.

Thanks,
Davidlohr