Hi,<br><br><signature></signature>
<div class="gmail_quote nylas-quote nylas-quote-id-1ajeoxn75act1yq12ntx7771f">
<br>
On Nov 10 2016, at 2:09 pm, Pekka Paalanen <ppaalanen@gmail.com> wrote:
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<p>On Wed, 9 Nov 2016 14:57:22 -0600<br>Derek Foreman <derekf@osg.samsung.com> wrote:</p>
<p>> We find the oldest backbuffer we can, on the grounds that clients are<br>> only going to keep a fixed history queue, so this gives them the<br>> greatest chance of being able to use that queue via making sure<br>> the age is ~always less than the depth of the swapchain</p>
<p>right, that is one side of the story.<br></p>
<p>On the other side, damage accumulates frame by frame. When you always<br>pick the oldest free buffer, you also maximize the accumulated damage<br>and need to repaint more than if you picked the youngest free buffer.</p>
<p>Furthermore, if you normally can do with cycling just two buffers, and<br>a hickup causes you to allocate up to four buffers and afterwards you<br>come back to the normal steady flow, you are now always cycling through<br>all the four buffers and repainting damage for age=4 instead of age=2.</p></blockquote><br>This would be a separate problem; the queue depth should really not be any longer than need be. If there is a case where it ends up longer (say when a surface switches from direct scanout to GPU composition), then Mesa should be realising this and trimming the queue. Other stacks do this.<br><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<p>If one was always picking the youngest free buffer, we could even have<br>heuristics to destroy buffers that have not been used for N swaps to<br>release some memory.</p></blockquote><br>That is not strictly dependent on picking the youngest buffer; it could/should be done regardless.<br><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<p>There is a trade-off between repainting a little more than necessary in<br>the normal case to mitigate the impact on hickups and making the normal<br>case optimized while hickups cause a heavy hit (full repaint plus<br>possibly even allocating a new buffer). However, in a proper compositor<br>I fail to see how such hickups would even occur - so you wouldn't need<br>to mitigate hickups but also using the oldest would not be any worse<br>since you never allocate extra buffers.</p></blockquote><br>Repainting more than necessary will only occur when the damage area is changing literally frame-by-frame. This is kind of an odd case, since it means that your eyes will have to be retargeting every 16ms.<br><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<p>It would be nice to see more rationale in the commit message about why<br>picking the oldest is better than picking the youngest buffer. Age<br>greater than the length of the swapchain is only a trigger for full<br>repaint, just like a freshly allocated buffer is, except you skip the<br>cost of allocating.</p>
<p>My counter-proposal is probably worse, but I'd like to see an<br>explanation because it's not obvious.</p></blockquote><br>
</div><div class="gmail_quote nylas-quote nylas-quote-id-1ajeoxn75act1yq12ntx7771f">Picking the youngest means that, as you say, predictability decreases in the case where one buffer gets stuck for a very long time. This arguably shouldn't happen generally, but provably does right now until Mesa gains the smarts to trim the buffer queue. I can see how this would be provoked on a CPU-bound system where we are doing direct scanout; by picking the youngest, we react to this situation by generating more CPU load.</div><div class="gmail_quote nylas-quote nylas-quote-id-1ajeoxn75act1yq12ntx7771f"><br></div><div class="gmail_quote nylas-quote nylas-quote-id-1ajeoxn75act1yq12ntx7771f">Picking the oldest means that the queue remains balanced, which does imply a complexity increase in the case that the surface damage area changes from frame to frame. I do not believe that this is a common case, and even if it were, the downside of this is not as bad as the downside of picking the youngest.</div><div class="gmail_quote nylas-quote nylas-quote-id-1ajeoxn75act1yq12ntx7771f"><br></div><div class="gmail_quote nylas-quote nylas-quote-id-1ajeoxn75act1yq12ntx7771f">Cheers,</div><div class="gmail_quote nylas-quote nylas-quote-id-1ajeoxn75act1yq12ntx7771f">Daniel</div><div id="n1-quoted-text-marker"></div>