VMware Communities
Stevamundo
Enthusiast
Enthusiast
Jump to solution

Fusion 7 is running VERY SLOW on the GM of Yosemite

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?

Steve Harper-Webmaster: http://madentec.com/intro/ Get the latest and greatest in assistive technology.
138 Replies
Polymorph
Contributor
Contributor
Jump to solution

Spoke too soon, VM's just get progressively slower over the hours to the point where they become unusable. 😕

Reply
0 Kudos
HPReg
VMware Employee
VMware Employee
Jump to solution

So you are saying that the boot-args workaround only works after boot, but the interrupt storm progressively returns after that?

What about manually making the Mac sleep? Does that workaround stick?

Reply
0 Kudos
Polymorph
Contributor
Contributor
Jump to solution

Not sure it ever helped the interrupts actually, I really shouldn't have upgraded to Fusion7 and Yosemite on the same day x_x

Here's a 5 second interrupt sample:

mac:~ user$ sudo nvram boot-args

boot-args debug=0x10

*** Sampled system activity (Wed Oct 22 14:50:40 2014 +0100) (5000.87ms elapsed) ***

****  Interrupt distribution ****

CPU 0:

  Vector 0x20(): 4.40 interrupts/sec

  Vector 0x46(SMC): 1.80 interrupts/sec

  Vector 0x49(MacBookPro11,3): 16.20 interrupts/sec

  Vector 0x74(XHC1): 30.39 interrupts/sec

  Vector 0x77(GFX0): 7.80 interrupts/sec

  Vector 0x79(ARPT): 983.43 interrupts/sec

  Vector 0xdd(TMR): 520.71 interrupts/sec

  Vector 0xde(IPI): 4339.04 interrupts/sec

CPU 1:

  Vector 0x20(): 2.40 interrupts/sec

  Vector 0xdd(TMR): 154.57 interrupts/sec

  Vector 0xde(IPI): 3088.66 interrupts/sec

CPU 2:

  Vector 0x20(): 3.20 interrupts/sec

  Vector 0xdd(TMR): 521.31 interrupts/sec

  Vector 0xde(IPI): 4540.41 interrupts/sec

CPU 3:

  Vector 0x20(): 1.80 interrupts/sec

  Vector 0xdd(TMR): 175.17 interrupts/sec

  Vector 0xde(IPI): 3038.67 interrupts/sec

CPU 4:

  Vector 0x20(): 2.80 interrupts/sec

  Vector 0xdd(TMR): 457.12 interrupts/sec

  Vector 0xde(IPI): 4689.78 interrupts/sec

CPU 5:

  Vector 0x20(): 3.40 interrupts/sec

  Vector 0xdd(TMR): 155.77 interrupts/sec

  Vector 0xde(IPI): 3025.87 interrupts/sec

CPU 6:

  Vector 0x20(): 4.00 interrupts/sec

  Vector 0xdd(TMR): 548.30 interrupts/sec

  Vector 0xde(IPI): 4782.17 interrupts/sec

CPU 7:

  Vector 0x20(): 3.40 interrupts/sec

  Vector 0xdd(TMR): 191.77 interrupts/sec

  Vector 0xde(IPI): 3375.01 interrupts/sec


MacBook Pro (Retina, 15-inch, Late 2013)


btw, I run Hyper-v stuff on Win8.1 within my Fusion VMs too so have hypervisor support enabled for this VM but not currently using it. Other than my VM feels about 10 times slower than it did yesterday there's not much else I can put my finger on 😕

Reply
0 Kudos
Polymorph
Contributor
Contributor
Jump to solution

(Just tried the sleeping thing, nope sadly not for me! Just tried, flicked between some mails in Outlook and I can physically see it drawing the mails it's that slow Smiley Wink

Reply
0 Kudos
renegad3
Contributor
Contributor
Jump to solution

So, after doing some light investigating (not had as much time as I had hoped).

Using the original terminal command from the thread on the other forum (0x4de) the CPU usage issue was either actually fixed, or just suppressed. However using the machine over time, my fans would eventually ramp up to the point of them being very audible, which almost never happened under Mavericks (unless I was actually pushing the machine).

I then switched to the command recommended here (0x10) and basically the same results, CPU usage seemed a bit more stable, but over time, again, the machine would warm and the fans would ramp up.

I do not normally sleep my machine, only my monitors. Overnight, the machine while not being used, would cool and fans would ramp down, but upon waking monitors and using fans/temps would again rise.

I used the sleep method posted earlier, and it now seems to be much better with regards to fans/CPU, more so than with terminal command.

As a note, I have not ever had the Vector 0x49 reported when using sudo powermetrics -s interrupts

for example, run just now:

^CJimmys-iMac-2:~ jimmy$ sudo powermetrics -s interrupts

Password:

Machine model: iMac12,2

SMC version: 1.72f1

EFI version: IM121.0047.B1F

OS version: 14A389

Boot arguments: debug=0x10

Boot time: Wed Oct 22 07:28:27 2014

*** Sampled system activity (Wed Oct 22 10:16:17 2014 -0400) (5001.11ms elapsed) ***

****  Interrupt distribution ****

CPU 0:

  Vector 0x51(HDAU/RP02/ARPT/pci12d8,400f/pci-bridge/ethernet/pci-bridge/DSB3/UPS2/pci-bridge/pci12d8,400e/pci-bridge): 23.59 interrupts/sec

  Vector 0x52(RP03/FRWR/pci-bridge/pci11c1,5901/pci12d8,400f/pci-bridge/ethernet/pci-bridge/DSB4/SBUS): 1.20 interrupts/sec

  Vector 0x56(HDEF/EHC1): 1.40 interrupts/sec

  Vector 0x57(EHC2): 10.80 interrupts/sec

  Vector 0x73(GFX0): 292.13 interrupts/sec

  Vector 0x75(SATA): 1.80 interrupts/sec

  Vector 0x78(GIGE): 16.20 interrupts/sec

  Vector 0xdd(TMR): 672.85 interrupts/sec

  Vector 0xde(IPI): 115.97 interrupts/sec

CPU 1:

  Vector 0xdd(TMR): 10.80 interrupts/sec

  Vector 0xde(IPI): 67.58 interrupts/sec

CPU 2:

  Vector 0xdd(TMR): 105.58 interrupts/sec

  Vector 0xde(IPI): 117.37 interrupts/sec

CPU 3:

  Vector 0xdd(TMR): 8.20 interrupts/sec

  Vector 0xde(IPI): 63.39 interrupts/sec

CPU 4:

  Vector 0xdd(TMR): 236.15 interrupts/sec

  Vector 0xde(IPI): 125.57 interrupts/sec

CPU 5:

  Vector 0xdd(TMR): 8.20 interrupts/sec

  Vector 0xde(IPI): 83.38 interrupts/sec

CPU 6:

  Vector 0xdd(TMR): 200.76 interrupts/sec

  Vector 0xde(IPI): 108.98 interrupts/sec

CPU 7:

  Vector 0xdd(TMR): 8.00 interrupts/sec

  Vector 0xde(IPI): 76.18 interrupts/sec

When I ran using 0x4de:

Jimmys-iMac-2:~ jimmy$ sudo powermetrics -s interrupts

Machine model: iMac12,2

SMC version: 1.72f1

EFI version: IM121.0047.B1F

OS version: 14A389

Boot arguments: debug=0xd4e

Boot time: Sun Oct 19 21:31:21 2014

*** Sampled system activity (Tue Oct 21 15:59:38 2014 -0400) (5000.20ms elapsed) ***

****  Interrupt distribution ****

CPU 0:

  Vector 0x51(HDAU/RP02/ARPT/pci12d8,400f/pci-bridge/ethernet/pci-bridge/DSB3/UPS2/pci-bridge/pci12d8,400e/pci-bridge): 21.00 interrupts/sec

  Vector 0x52(RP03/FRWR/pci-bridge/pci11c1,5901/pci12d8,400f/pci-bridge/ethernet/pci-bridge/DSB4/SBUS): 1.40 interrupts/sec

  Vector 0x56(HDEF/EHC1): 1.40 interrupts/sec

  Vector 0x57(EHC2): 69.60 interrupts/sec

  Vector 0x73(GFX0): 780.97 interrupts/sec

  Vector 0x75(SATA): 20.80 interrupts/sec

  Vector 0x76(GIGE): 2854.28 interrupts/sec

  Vector 0xdd(TMR): 2523.70 interrupts/sec

  Vector 0xde(IPI): 1031.56 interrupts/sec

CPU 1:

  Vector 0xdd(TMR): 48.20 interrupts/sec

  Vector 0xde(IPI): 321.19 interrupts/sec

CPU 2:

  Vector 0xdd(TMR): 1445.54 interrupts/sec

  Vector 0xde(IPI): 869.76 interrupts/sec

CPU 3:

  Vector 0xdd(TMR): 42.20 interrupts/sec

  Vector 0xde(IPI): 376.78 interrupts/sec

CPU 4:

  Vector 0xdd(TMR): 1507.14 interrupts/sec

  Vector 0xde(IPI): 839.17 interrupts/sec

CPU 5:

  Vector 0xdd(TMR): 42.80 interrupts/sec

  Vector 0xde(IPI): 295.19 interrupts/sec

CPU 6:

  Vector 0xdd(TMR): 1766.73 interrupts/sec

  Vector 0xde(IPI): 840.37 interrupts/sec

CPU 7:

  Vector 0xdd(TMR): 42.40 interrupts/sec

  Vector 0xde(IPI): 261.79 interrupts/sec

Not sure what to make of all of it so far, but in the shorter amount of time, it appears the sleep method is the 'solution' to this point, at least for me.

I will also add, that my VM guest was not degraded use wise to the point many have described here (unusable), but it was noticeably sluggish.

Thank to all those who have added to this, and hopefully we can get Apple to fix things properly quickly. I will also be filing with Apple as well.

Reply
0 Kudos
HPReg
VMware Employee
VMware Employee
Jump to solution

Interesting. You are not seeing a vector 0x49 storm. You are seeing high IPI activity, but not out of this world (like the vector 0x49 storm was). It could be a separate issue.

Can you try disabling hypervisor support inside the VM (VM > Settings > Processors & Memory > Advanced > Enable hypervisor applications...) and see if it helps?

How many vCPUs do you have in that VM? Can you try reducing that number to see if that helps as well?

Reply
0 Kudos
renegad3
Contributor
Contributor
Jump to solution

my settings, as they have been...

Also, it seems fans are still ramping, though it has taken longer than it did previously, with same basic usage, so perhaps the sleep method doesn't work any better/worse for me.

Reply
0 Kudos
HPReg
VMware Employee
VMware Employee
Jump to solution

mono-processor VM? Then Fusion is not directly generating the high IPI counts you are seeing (it might just be regular Mac OS TLB shootdowns).


Can you try sampling the interrupts both when Fusion is running your Win8.1 VM and when you quit Fusion, to see if that makes any difference?

Reply
0 Kudos
dmes
Contributor
Contributor
Jump to solution

This works for me doing the following:

(didn't bother with shutdown, etc)

- Pause guest

- manually sleep iMac waiting for internal HD to stop

- Wake Mac

- start guest vm

The vm ran just like it used to before Yosemite.

and btw, interrupts went from 127k/sec to 10/sec

Reply
0 Kudos
renegad3
Contributor
Contributor
Jump to solution

yes, I dont need any real-horsepower in the VM, its just for a few minor legacy apps.

For the record, its Windows 7 enterprise.

Here are the files....

Reply
0 Kudos
AbortRetryFail
Contributor
Contributor
Jump to solution

hw.model: iMac14,2

I did not apply the nvram boot-args workaround.  The sleep workaround seems to have halted the behavior, at least for now.

Steven

Reply
0 Kudos
mikenr1
Contributor
Contributor
Jump to solution

Same Problem with Fusion 7 and Yosemite

iMac12,2

Reply
0 Kudos
vansrus
Contributor
Contributor
Jump to solution

Same Problem with Fusion 7 and Yosemite

iMac12,2

Reply
0 Kudos
BillT1901
Contributor
Contributor
Jump to solution

Same here. iMac 12,2

sudo powermetrics -s interrupts shows heavy interrupts per sec.

CPU 0:

    Vector 0x49(iMac12,2): 116223.48 interrupts/sec

This occurs whether VMWare Fusion is running or not. So Serious Apple regression on this hardware (this is NOT occurring on my 2012 macbook). So again, not an issue in particular to Fusion 7.

Put the iMac to sleep and wait at least 30 seconds. Bring it back up.

Powermetrics command now shows interrupts/sec on CPU 0 back to normal

CPU 0:

    Vector 0x51(HDAU/RP02/ARPT/DSB3): 10.20 interrupts/sec

At that point everything, including VMWare Fusion is working fine. Problem comes back after a reboot, of course...

Reply
0 Kudos
Nemo10904
Contributor
Contributor
Jump to solution

I found this elsewhere and it worked for me - thanks shanebickley

HOW TO FIX THE SLUGGISHNESS OF Fusion 5,6 & 7 AFTER YOSEMITE UPDATE

https://communities.vmware.com/message/2440024?tstart=0

hw.model: iMac12,2

i7 3.4GHZ w/16G - 1 processor and 4G assigned to W7 vm

interrupts down to 10-12/sec

WM is performing well

iMac is also performing normally

Reply
0 Kudos
caoslord
Enthusiast
Enthusiast
Jump to solution

Hello this is the real fix! (corrected, with a better suggestion)

open terminal.

Type: sudo nvram boot-args="debug=0x10"

press enter.

insert your password.

Reboot


Enjoy.


I fount in another post.

Here

fusion 6.04 with Yosemite GM very slow

Reply
0 Kudos
HPReg
VMware Employee
VMware Employee
Jump to solution

It is not the real fix. Don't use

   sudo nvram boot-args="debug=0xd4e"

as it is way too intrusive. Instead, use

   sudo nvram boot-args="debug=0x10"

as recommended above.

caoslord
Enthusiast
Enthusiast
Jump to solution

Thank you,

please, what is the difference between the two option?

Best Regards,

Luca

Reply
0 Kudos
HPReg
VMware Employee
VMware Employee
Jump to solution

See debug.h for the meaning of the various debug= bits. Search for DB_HALT on this page. Some bits are not documented, in that case download the xnu source code and grep for the bit.

Reply
0 Kudos
caoslord
Enthusiast
Enthusiast
Jump to solution

Thank You!

Reply
0 Kudos