Apple's own web-site with GM - possible bug

Hi Tim,

Working fine for me on El Capitan GM, Safari Browser - loads quick and displays well. I suggest clearing the cache for the Apple site and then restarting your Mac.

Any better?

Max.

Hi Max,

I tried that - something I've just tried is disabling the graphics switching - works fine now.

It's just when it's set-up to switch automatically that there is an issue, and looking through the logs it looks like it's switching rapidly between the Intel HD 4000 and Nvidia graphics.

Will try a reboot and clearing the PRAM / NVRAM etc, see if that helps.

Thanks

Tim

It's a graphics intensive site alright. I have a MBP with the same integrated card on the GM too. I'll enable graphics switching and try it there. I doubt that the NVRAM reset will help. It's the SMC that is related to (external) displays.

SMC is what I meant - got confused a little!

I wish Apple would make it so that when you are on power it went straight to the Nvidia graphics, and when on power if would switch graphics where necessary.

I can repeat your experience like that - the initial slide takes a long time to load and has peculiar graphical glitches while it does - it also slows Safari right down and, even when the slide is loaded, it isn't working and Safari is still slowed right down when trying to display it. The later slides all work fine though and Safari is quick again once I've scrolled past the initial slide.

I don't have an MBP with Yosemite still installed to test on unfortunately.

Thanks Max, I'll raise a bug report - better to get it sorted - did you notice the screen flickering? Did you see any evidence in the logs that it was rapidly switching between the graphics cards?

Yes, I can confirm the rapid graphics card switching. Quite a bit of this sort of thing:

12/09/2015 11:46:57.403 a.m. WindowServer[197]: Received display connect changed for display 0x4280382
12/09/2015 11:46:57.404 a.m. WindowServer[197]: Found 1 modes for display 0x04280382 [1, 0]
12/09/2015 11:46:57.404 a.m. WindowServer[197]: Received display connect changed for display 0x3f003f
12/09/2015 11:46:57.404 a.m. WindowServer[197]: Found 1 modes for display 0x003f003f [1, 0]
12/09/2015 11:46:57.404 a.m. WindowServer[197]: Received display connect changed for display 0x3f0040
12/09/2015 11:46:57.404 a.m. WindowServer[197]: Found 1 modes for display 0x003f0040 [1, 0]
12/09/2015 11:46:57.404 a.m. WindowServer[197]: Received display connect changed for display 0x3f0041
12/09/2015 11:46:57.404 a.m. WindowServer[197]: Found 1 modes for display 0x003f0041 [1, 0]

No screen flickering for me though.

That'sa exactly what I'm seeing

9/12/15 11:22:16.168 AM WindowServer[191]: CGXMuxAcknowledge: Posting glitchless acknowledge

9/12/15 11:22:16.312 AM WindowServer[191]: Received display connect changed for display 0x4280382

9/12/15 11:22:16.312 AM WindowServer[191]: Found 1 modes for display 0x04280382 [1, 0]

9/12/15 11:22:16.313 AM WindowServer[191]: Received display connect changed for display 0x3f003f

9/12/15 11:22:16.313 AM WindowServer[191]: Found 1 modes for display 0x003f003f [1, 0]

9/12/15 11:22:16.313 AM WindowServer[191]: Received display connect changed for display 0x3f0040

9/12/15 11:22:16.313 AM WindowServer[191]: Found 1 modes for display 0x003f0040 [1, 0]

9/12/15 11:22:16.313 AM WindowServer[191]: Received display connect changed for display 0x3f0041

9/12/15 11:22:16.313 AM WindowServer[191]: Found 1 modes for display 0x003f0041 [1, 0]

9/12/15 11:22:16.315 AM WindowServer[191]: _CGXRemoveWindowFromWindowMovementGroup: window 0xf16 is not attached to window 0xf17

The screen flickering is only just noticible, more just the menubar at the top that flickers - the same level of flickering you see when it switches between graphics cards.

The whole graphics card switching has been a fly in Apple's ointment for a long while - it's been the cause of many hard freezes on some MBPr's - mine included, right to the point that I had the logic board replaced because both Apple and I thought it was needed as part of the replacement program - was excited to get it back thinking no more crashing, alas not the case.

The only way I stopped it was to roll back to 10.9.4 - the bug seemed to appear in 10.9.1, resolved in 10.9.4 and came back with vengeance in 10.10 - seems to be a lot better in 10.11 but nowhere near as good as 10.9.4.

Bug raised - 22673911

Interesting that the display references (0x04280382, 0x003f003f, 0x003f0040, 0x003f0041) are identical. I'll dupe 22673911.

Could be that we are on the same hardware (MPBr Early 2013 - 16GB 512GB SSD)

Thanks for duping the bug report - wonder if we will get a GM2?

Yes, good hunch - same hardware.

Plenty of time for another GM before the 30th - this one's only a canditate after all.

Indeed, that was a nice leak by Apple - obviously they knew what they were doing! Got to love the marking skills!

BTW, do you get / have you had loads of hard freezes with your MBPr? Is that why you disabled graphics switching?

Actually no. I've never had a hard freeze on it. I disabled graphics switching because it spends almost all its time plugged in (I cycle the battery once a week though). I remember deciding that graphics switching was just one more thing to go wrong, but you are reminding me that I should routinely test with switching enabled - something I've not been doing.