Could you capture a trace and attach it to this bug?

In Chrome 61, could you try disabling "Color correct rendering" to see if that affects performance?

Since this is a very noticeable perf regression (it sounds?), I am increasing the priority.

I have this model of iMac as a Home computer.  I will look at this issue from home next week.

When you say "stutter" do you mean like frames are being displayed out of order? If so, try applying this patch locally:
https://chromium-review.googlesource.com/c/chromium/src/+/579690

What kind of trace did you mean? I've attached a Javascript+Rendering trace from about:tracing. Do you need another one?

Enabling or disabling "Color correct rendering" didn't make a difference.

The stuttering is definitely noticeable but I don't know if it's a regression.

The box seems to jump ahead a bit more than it should about 5-10% of the frames. Looking more carefully, it does seem like frames are rendering out of order. There is a "tail" ghost behind the box whenever it jumps, which may be caused by the ooo frame.

I can try the patch but it will take a while to prepare a build :-) OTOH, the Canary at 63.0.3227.0 has the same problem (and already includes that patch.) FWIW the Canary doesn't have the "Color correct rendering" flag.

Deleted: trace_trace.json.gz 4.8 MB

trace_trace.json.gz 4.8 MB Download

As expected, the build from head has the same problems and already includes [1].

[1] https://chromium-review.googlesource.com/c/chromium/src/+/579690

> FWIW the Canary doesn't have the "Color correct rendering" flag.
I know that's why I asked to try it on M61.

Okay, so this is not a regression AFAICT.

I can confirm that I am able to repro on iMac 27" Retina 5K/ AMD Radeon R9 M390.

There is nothing fishy with the trace.  The one I was able to generate locally look pretty much the same as the one in #2.

I tried running with --disable-2d-canvas-image-chromium (hardware overlay mode) and there was still stuttering.  

After observing the problem closely: I am pretty sure we are not dealing with frames being out of order (that type of artifact is easy to recognize because the animation janks backwards when it happens).  What we have is more likely a case of dropped/doubled frames.  This should not be happening when the system is under such light load.

This looks like a problem with vsync alignment IMHO

Re-triaging as a GPU scheduling issue.

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot