Overclock.net › Forums › Overclockers Care › Overclock.net Folding@Home Team › Bad CPU usage in bigbigadv
New Posts  All Forums:Forum Nav:

Bad CPU usage in bigbigadv

post #1 of 16
Thread Starter 
I got a Virtualbox VM setup yesterday running bigbigadv with help from Lutro0 (thank you) but I'm having some problems with it not using all of my CPU. It's taking lots of little dips in my CPU usage and I don't know what could be causing it. I've tried leaving the computer completely alone as well as changing priorities around and nothing seems to be getting it to fully utilize the CPU more consistently. They don't appear to be constant huge dips but they are happening all the time and sometimes there are huge dips. This happened to a much smaller degree in my previous VMware bigbigadv VM, but it still kept my temperatures much more consistent. With this my temps are between 62C and 73C on the same core, whereas previous full folding loads kept it pegged pretty consistently at 73-75C under load. I know the HWMonitor power draw isn't all that accurate but same is happening with that as well, used to be really consistent at 105 watts and now it's between 80 and 95 watts.

Here is what my CPU utilization in task manager looks like. It has looked worse than this too.



This is the site I got the image and virtualbox from, and I don't know too much about the settings I used while setting it up so hopefully Lutro0 can help out there as he made all the decisions of what options/settings to use. It is receiving bigbigadv successfully though because I'm on a 6903 right now, and without using the machine at all for a full frame, it's taking 72 minutes per frame when I was expecting more like 56 minutes. This is at 4.7Ghz.

I'm not too sure what it could be but I'm hoping some of the people running bigbigadv on a 2600k will be able to chime in and let me know if they can think of anything that could be causing this, or at least let me know their setup and how it's working for them so if I have to start again from scratch I can copy their settings.

Thanks for any help.
Edited by juano - 10/14/11 at 1:54am
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
post #2 of 16
Hmm, I did forget to ask. Are you gpu folding as well?
post #3 of 16
A side note:

I've seen a lot of people recommending only 4gb for a folding machine. In the case of a vm folder, it's pretty clear from your screen shot that 8gb would be more wise.

On topic:

But, yeah. GPU folding would be the big question. If you have the vm assigned to 8 threads and are gpu folding, that would explain the problem.
post #4 of 16
Thread Starter 
I am GPU folding the advmethod WUs which take very little CPU cycles but the problem remains even when the GPU folding is stopped. I also tried priority one less than real time which also had no effect. Not using the computer at all since I posted this did not bring my usage up or my TPFs down. I also increased the VMs RAM allowed to 5gb from 4gb to no avail.

Lutro0 does the packet size setting in the web configuration have to do with the actual folding or is that a network and WU downloading thing? Because on the VMplayer bigadv instal that was set to small instead of medium as it is now. I also tried updating the virtual box to the latest 4.1.4 which didn't help. I almost think that it's something wrong with what I installed/selected I should be running the VM in Linux 2.6 64bit for the 1.4.1 image from this site right? Should I try the 1.4.0 image or a different OS for the VM?

Mach1 I do agree that 8gb is the right recommendation for a VM folder. I do still think that for a dedicated folder that 4gb is fine don't you?
Edited by juano - 10/14/11 at 8:16am
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
post #5 of 16
I really think that even if the gpu3 units are the less cpu usage ones, that it is still a gpu3 unit and no mater which gpu3 unit ive done it takes a chunk of cpu and if you run full tilt the cpu always takes a hit. Try that client at smp7 (set it at smp 7 in the client config) and see how that works.
post #6 of 16
Thread Starter 
The problem is still continuing when GPU folding is stooped and closed though. You want me to try SMP7 without GPU folding?
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
post #7 of 16
ahhh, no. If your not running gpu folding then you dont have to do smp7.

The only other setting you could try is smp12 in the client setup (config page)

If that doesnt work, then perhaps the way of getting the core hack to work like that doesnt work well with the new setups, I will look to see which versions I was using. (I have them on a usb somewhere) I do however know they never got close to a native client, but did better then a windows one.
post #8 of 16
Thread Starter 
Alright I will try the SMP 12, that would be in the web configuration right? Because the VM settings are already at 12 cores, the only thing I can find that isn't is the web config page.

Again thanks for the help Lutro0, I was thinking of troubleshooting something along those lines because of that error message in the settings that says the VM is set for more cores than the host but didn't really know if that was what it was supposed to be saying for the core trick or not.
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
Main Rig
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 2700k ASUS P8P67 WS Revolution EVGA 980 Ti SC+ Samsung 4x4GB DDR3 1866MHz 
Hard DriveHard DriveOptical DriveCooling
Samsung 850 Evo 1TB Samsung Spinpoint F4 2TB Samsung BD Combo Noctua NH-D14 
OSMonitorPowerCase
Windows 10 64 bit Asus PG279Q Kingwin Lazer Platinum 1000W Silverstone Raven RV03 
  hide details  
Reply
post #9 of 16
Yea no problem, it might come down to its not the best method for running it, I only ever ran it as smp7 with a gpu going, so I guess we will find out lol.
post #10 of 16
Quote:
Originally Posted by juano View Post
Alright I will try the SMP 12, that would be in the web configuration right? Because the VM settings are already at 12 cores, the only thing I can find that isn't is the web config page.

Again thanks for the help Lutro0, I was thinking of troubleshooting something along those lines because of that error message in the settings that says the VM is set for more cores than the host but didn't really know if that was what it was supposed to be saying for the core trick or not.
Shouldn't that be smp 8 as you have 8 threads or do I need more coffee?
HAF932
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k ASUS P8P67 Deluxe EVGA gtx 450 G.SKILL Ripjaws X 4GB 
Hard DriveOptical DriveOSMonitor
Intel 510 Series 128gb LG Bluray Win 7 64bit S23A950 
PowerCase
Antec TruePower Quattro TPQ-1000 HAF 932 
  hide details  
Reply
HAF932
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k ASUS P8P67 Deluxe EVGA gtx 450 G.SKILL Ripjaws X 4GB 
Hard DriveOptical DriveOSMonitor
Intel 510 Series 128gb LG Bluray Win 7 64bit S23A950 
PowerCase
Antec TruePower Quattro TPQ-1000 HAF 932 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Overclock.net Folding@Home Team
Overclock.net › Forums › Overclockers Care › Overclock.net Folding@Home Team › Bad CPU usage in bigbigadv