> On Sat, Oct 19, 2013 at 10:49:04AM +0200, Axel Davy wrote: >> I've tried benchmarking AsyncSwap with the phoronix-test-suite, >> and I was surprised to see a regression with Openarena and Xonotic. >> According to dri devs, it is because, since I do an exchange, the >> application >> is fullscreen and then Weston uses the buffer as scanout buffer, when >> the buffer is >> released and we render again in the buffer, L3 caching is disabled. > > It is simple to demonstrate that hypothesis incorrect by marking all > colour buffers as uncached. > -Chirs > > -- > Chris Wilson, Intel Open Source Technology Centre >
Unfortunately, I can't test that with wlglamor, since there is too many abstraction: I don't have access to the caching of the buffers. I did many tests, and the only conclusion I've came to is that a buffer we used for the framebuffer is slower to render to after. I've also compared Bypass compositing + no AsyncSwap vs No Bypass compositing + AsyncSwap, and as expected, it performs the same (and in the latter case, we don't get tearings). (And of course, it performs better than No Bypass compositing + no AsyncSwap, or Bypass compositing + AsyncSwap + useless copy) Axel Davy _______________________________________________ wayland-devel mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/wayland-devel
