Was unable to reproduce this on Chrome 52.0.2743.116 (stable) or 54.0.2832.2 (canary). The image just goes to the next one as expected. Have you tried this with all extensions disabled?

Either way this should be a blink issue - marking it as such so hopefully someone who knows about this can help with triage.

Able to reproduce this on the latest canary(54.0.2836.0) and the latest beta(53.0.2785.70) on Mac OS 10.11.6 on Graphics driver: Intel HD Graphics 3000 384 MB. The same works fine on Mac with Intel HD Graphics 5000 OpenGL Engine driver on the same chrome versions.

Bisected on Mac OS 10.11.6 having Intel HD Graphics 3000 384 MB driver , Regressed in M-53.

Last good build: 53.0.2769.0
First bad build: 53.0.2770.0

Changelog: https://chromium.googlesource.com/chromium/src/+log/e973b43fd81bb5066ec3c95034e8290cde5068eb..2b6875bfd954d336432702512d1ffe4c83343d2b

Not sure though but could this be related to https://codereview.chromium.org/2060043002

jbauman@: Please take a look and help in finding an appropriate owner if the change is unrelated.

Cc'ing  ccameron@ as well I could repro this on Mac with Intel HD Graphics 3000 384 MB driver and there is https://codereview.chromium.org/2076443002 related to this driver. 


Note: Issue is seen on Windows-10, Linux Ubuntu 14.04 on the latest stable(52.0.2743.116) and the latest canary(54.0.2836.0) as well. Attached is the gpu details of the Windows-10 and Linux Ubuntu 14.04 where issue is seen.

Deleted: GPU_Windows10.txt 8.4 KB

GPU_Windows10.txt 8.4 KB View Download

Deleted: Gpu_LinuxUbuntu14 13.2 KB

Gpu_LinuxUbuntu14 13.2 KB View Download

A friendly reminder that M53 Stable is launching VERY soon! Your bug is labelled as Stable ReleaseBlock, pls make sure to land the fix and get it merged into the release branch ASAP (before 5:00 PM PT, Tuesday) so we can take it for this week LAST Beta release for Desktop. Thank you!

Note: Merge has to happen by Friday, August 26th, 5:00 PM PST in order to make into the desktop Stable final build cut. 

I saw this on 52 (see my user agent in #1) on a macbook pro

My change is windows-only, so not the cause of this. Maybe we're running out of memory for tiles? That might make sense with the software renderer, because the limit's pretty low, but I'm not sure why it would happen on the windows 10 or ubuntu machines that ajha@ tested on.

My MacBookPro has Geforce GT 750M + Intel Iris Pro graphics chipsets btw.

OOO

 jbauman@ as ccameron@ OOO, who else will be a right owner for this? 

In M52 as well, so not ReleaseBlock-Stable.

This seems fixed.  Please re-open if you are still experiencing problems.

My bad, scrolling triggers the issue and it looks really bad.

vmpstr@ could you please take a look?  I'm not 100% sure what's wrong but may be constant invalidation.

This page produces a layer for each of the image in the carousel, which at this point is 195 layers, plus there's a bunch of layers off to the left of the page which are within the prepaint region. All in all, we run out of 512mb tile memory quite easily, especially on retina.

One thing to note is that the pictures have a transform3d(100%, 0, 0), which when replaced with translateX(100%) doesn't cause the layerization and behaves normally.

I'm not sure if this is something we can fix on our end. 

Deleted: screen.jpg 320 KB

	screen.jpg 320 KB View Download

(reassigning to GPU team now that the cause has been identified, as per comment 13)

I filed crbug.com/663831 to track rasterization of layers in order of depth. This bug is WontFix, since I think the page needs to change.