Fusion 7 is running VERY SLOW on the GM of Yosemite

[Stevamundo]

Basically Fusion 7 is unusable on the GM of Yosemite because fusion 7 is so slow.

When I was still on Mavericks 10.9.5  Fusion 7 was working just great. What happened?

Is anyway that i can fix this? Or do I have to wait for an update?

[mihaval]

For me, it is the same. Burning CPU like mad. It is better after sleep/resume.

[CuriosTiger]

Not the same issue anymore (interrupt counts are sane), but I am still seeing the vmware-esx process take 186% CPU. Windows isn't quite as unusable as it was with the interrupt storm bug, but it's very sluggish.

This is with VMware Fusion 7.0.1 running on Yosemite 10.10.1 on a MacBookPro 9,1 (mid-2012 non-retina), Windows 7 guest OS and boot-args       kext-dev-mode=1 debug=0x10.

[dariusd]

Hi CuriosTiger,

Hmmmm... Fusion 7.0.1 is working just fine on two Macs here that we've updated to 10.10.1.  Could be another hardware-specific issue...

Can you try using the host's Activity Monitor to collect a "sample" of vmware-vmx?  (Note that the earlier interrupt storm bug caused sample to fail... You might need to cancel it or power down your VM if sample gets stuck.)

If that doesn't work, try collecting a Spindump from within Activity Monitor.

In either case, please post the resulting sample or spindump file as an attachment.

Thanks!

--

Darius

[bitslap]

dariusd wrote:

Hi CuriosTiger,

Hmmmm... Fusion 7.0.1 is working just fine on two Macs here that we've updated to 10.10.1.  Could be another hardware-specific issue...

Can you try using the host's Activity Monitor to collect a "sample" of vmware-vmx?  (Note that the earlier interrupt storm bug caused sample to fail... You might need to cancel it or power down your VM if sample gets stuck.)

If that doesn't work, try collecting a Spindump from within Activity Monitor.

In either case, please post the resulting sample or spindump file as an attachment.

Thanks!

--

Darius

Still not usable for us either. Top spec'd MacBook Pro Retina 15"

Had a 30 minute call with Nikunj today and mentioned this issue to him.

[antoniokratos]

I'm having the same issue on my iMac 12,2 (27" mid 2011, i7, 16 GB RAM, HD6970M 2 GB).

I've posted this issue on Apple Community forums, this is the link: https://discussions.apple.com/thread/6683965

[dgross]

Many people are having this problem, not just with VMware but also with other virtualization software and also with UI responsiveness in general. I personally had problems with Virtualbox VMs. Please report this problem to Apple so that they are aware of it so that it can be fixed for a future version of OS X. Here are instructions for doing so. Make sure to include a lot of relevant keywords: Yosemite, 2011, com.apple.driver.AppleACPIPlatform, iMac, slow, Virtual Machine.

To report this problem to Apple:

1. Visit https://www.apple.com/feedback/imac.html
2. Here's a sample for the "Subject" field. Please adapt it to include your specific details: Yosemite slow on 2011 iMac - com.apple.driver.AppleACPIPlatform
3. Here's a sample for the "Message" field. Please adapt it to include your specific details: After upgrading my iMac from Mavericks to Yosemite, there was a significant decrease in performance. The UI became less responsive and animations become much choppier, and Virtual Machines became slow to the point of no longer being usable. With the help of a Parallels article at http://kb.parallels.com/122767, I traced the issue to an excessive number of interrupts being generated by com.apple.driver.AppleACPIPlatform . The solution was to run sudo nvram boot-args="debug=0xd4e". Many people are experiencing the same problem - please see http://forum.parallels.com/threads/parallels-10-running-slow-on-yosemite.326205/page-2 andhttps://discussions.apple.com/thread/6621147 for example.

If you have an Apple Developer account, you can also file a bug report for OS X at https://bugreport.apple.com/

Please also see this issue reported on other forums: https://discussions.apple.com/thread/6621147

Parallels 10 running slow on Yosemite | Page 2 | Parallels Forums

[alaselva]

Same issue on my Mac

iMac 27inch, late 2013

Cpu core i7 3.5 Ghz, 16Gb ram 1600 DD3

VideoCard GeForce GTX775M 2048 MB

Unluckily nvram boot-args="debug=0x10" does not make any difference to me.

Regards

[vmxmr]

DGross means well by encouraging people to report this issue to Apple, and it is perfectly reasonable to express your feelings and experiences in the Apple community forum.

DEVELOPERS:

THIS BUG HAS BEEN REPORTED AND ACKNOWLEDGED BY APPLE. IT DOES NOT HELP TO REPORT IT AGAIN AND AGAIN.

It is counterproductive for developers to file additional bug reports with Apple through the bugreport.apple.com link. If you read this thread thoroughly, you will note that Apple has already acknowledged the bug and closed VMware's bug report because it is a duplicate of a previously reported bug. Filing additional bug reports will divert Apple's engineering resources to review and analyze those reports, only to determine that they are duplicates of a previously reported bug. We don't want that. They know about the problem. Let's give them the time they need to fix this bug.

I can't say whether more reports to the apple.com/feedback/imac link will encourage Apple to give this issue a higher priority. I will leave it to your personal judgement.

Finally, if you read this thread thoroughly, you will learn that sudo nvram boot-args="debug=0xd4e" is not the best solution. Better to use: sudo nvram boot-args="debug=0x10", which has the least impact on your iMac. In Mountain Lion and Mavericks, "debug=0x10" was a "no-op" (a dummy instruction that does nothing) and is likely to be the same in Yosemite.

Or simply put the Mac to sleep first.

Let us hope that Apple and their software partners (such as VMware and Parallels) find a quick, reliable, and robust solution to this bug.

[alaselva]

Yesterday VMware has sent out a Fusion update and the overall performance is better than before. Of course, I have removed the boot args parameters

[vmxmr]

Yesterday VMware has sent out a Fusion update and the overall performance is better than before. Of course, I have removed the boot args parameters

The serious performance problem with mid-2011 iMacs was NOT fixed in VMware Fusion 7.1.0. If you have an affected Mac, it may be premature to remove the boot arg parameters. 

The release notes are worth your time to read, as they include a list known issues (bugs) that were not fixed in this release:

VMware Fusion 7 and VMware Fusion 7 Professional Release Notes

[CuriosTiger]

Hi Darius,

The problem went away for a bit, but returned with a vengeance today (reboot took about half an hour, rather than the sub-30 seconds that are normal.) So I looked this thread back up and saw your response.

I attempted to collect a sample in Activity Monitor, but after about 7 minutes of waiting, I cancelled it. How long is it supposed to take?

I'll try to get a spindump later on.

[dariusd]

A sample is supposed to take about 5-10 seconds to complete, but this problem seems to prevent the sample from completing at all.  A spindump will take a while longer to complete (should still be much less than a minute), but it should successfully complete even when the host is in this state.

Please post the spindump (as an attachment) when you have obtained it.

Thanks!

--

Darius

[CuriosTiger]

Note that I do have the boot-args flag set:

peterbilt:~ stian$ nvram boot-args

boot-args       debug=0x10

In spite of this, and in spite of there not appearing to be crazy numbers of interrupts anymore (see my earlier thread in this post), I was still not able to collect a sample.

I collected a spindump; however, right now, Fusion is behaving normally. I'm not seeing the slowdown at the moment. It returned yesterday after being absent for several months following the boot-args workaround. That makes this spindump of limited value; I guess if nothing else it can serve as a baseline of my machine in case I'm able to catch it during an extreme slowdown again later.

Curiously, an attempt to collect a sample still failed even now that the problem is NOT occurring.

[CuriosTiger]

As an added note, I checked interrupts during the slowdown episode yesterday, and they were not out of whack. A few thousand per second, not hundreds of thousands. So whatever this other slowdown bug is, it does appear that the boot-args workaround is still doing its job.

[Cornflake5603]

@lordboogie YOU ARE A HERO!!! THANK YOU SOOOOOO MUCH!

That problem nearly drove me crazy. Your solution is so simple - one command in the terminal window, a reboot and everything is fine again. I am so relieved that my virtual machines work again like a charm 🙂

[pprem]

I upgraded to Yosemite 10.10.2 and no more bug.

VMWare 6 on iMac mi-2011 works fine. I don't need to sleep the computer after starting it for using my VM.

Hope the problem is solved for you too.

[HPReg]

Can other people on this thread confirm that OS X 10.10.2 fixes the issue?

[lordboogie]

Yes! Finally I can say as well 10.10.2 has fixed the issue with me as well. Tested using VMware as well as Parallels.

[Putschli]

Hi,

yes I can confirm that OS X 10.10.2 Update surprisingly fixes the issue on iMac (27", Middle 2011).

[HPReg]

Somebody (the original poster) might want to change the "correct answer" for this thread. The correct answer is now to update to Yosemite 10.10.2.