Re: regression/bisected/6.9 commit 587d67fd929ad89801bcc429675bda90d53f6592 decrease 30% of gaming performance

[Takashi Iwai <tiwai@xxxxxxx>]

On Mon, 01 Apr 2024 13:07:26 +0200,
Mikhail Gavrilov wrote:
> 
> On Mon, Apr 1, 2024 at 12:07 PM Takashi Iwai <tiwai@xxxxxxx> wrote:
> >
> > Is it a regression against the performance on 6.8?
> >
> > If so, what happens when you revert both this (587d67fd929a) and
> > beb45974dd49?  That should make the timer code back to 6.8 code.
> >
> >
> > Takashi
> 
> The commit beb45974dd49 boosted performance to 30%, but commit
> 587d67fd929a canceled this boost.
> 
> I want to say that reverting commit beb45974dd49 returns performance
> to 83FPS [1] and this is sad.
> 
> So I ask again is it possible to find a better approach than 587d67fd929a?
> 
> [1] https://postimg.cc/xJMjkcHP

The only code of the commit 587d67fd929a is the close of ALSA timer
device.  If it really matters, it implies that the user-space seems
trying to open/close the ALSA timer device so frequently.

Are you using ALSA dmix/dsnoop?  It would explain.


Takashi