• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Giving VMware Player another go

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

ChasR

Senior Member
Joined
Apr 12, 2004
Location
Atlanta
With Linux SMP performance getting closer to Windows SMP performance with each new core release and now being much better than Windows on p6701, I'm going to crank up some VMs and see how they do. Let you know in a few days. I can say preliminary indications are looking really good for those of us with mere quads.
 
Preliminary somewhat unscientific results from a Q6600 @ 3.4 running with GTX 260:
Code:
 Project ID: 6701
 Core: GRO-A3
 Credit: 921
 Frames: 100


 Name: ChasR VM (Ubuntu 10.10)
 Path: \\UBUNTU\fah\
 Number of Frames Observed: 12

 Min. Time / Frame : 00:11:33 - 6,359 PPD
 Avg. Time / Frame : 00:11:47 - 6,171 PPD
 Cur. Time / Frame : 00:11:49 - 6,131 PPD
 R3F. Time / Frame : 00:11:47 - 6,149 PPD
 All  Time / Frame : 00:11:47 - 6,149 PPD
 Eff. Time / Frame : 00:12:15 - 5,915 PPD


 Name: ChasR Win SMP (Win 7)
 Path: C:\Users\ChasR\Win FAH SMP\
 Number of Frames Observed: 300

 Min. Time / Frame : 00:11:41 - 6,251 PPD
 Avg. Time / Frame : 00:12:51 - 5,419 PPD

Minimum frame time on the Win 7 client is probably when the GTX260 was out of work or off. Comparing the average production the VM gets 752 ppd more. It's about 8% faster. Now we'll have to see how it does on other WUs and without the GPU. My two native Linux Q6600s (3.2 GHz) without a GPU do about 7000 ppd, so the VM looks to be close to that.
 
Here's the VM without the GPU running:

Project ID: 6701
Core: GRO-A3
Credit: 921
Frames: 100


Name: ChasR VM
Path: \\UBUNTU\fah\
Number of Frames Observed: 15

Min. Time / Frame : 00:10:48 - 7,033 PPD
Avg. Time / Frame : 00:11:51 - 6,119 PPD
Cur. Time / Frame : 00:10:48 - 6,980 PPD
R3F. Time / Frame : 00:12:09 - 6,204 PPD
All Time / Frame : 00:11:51 - 6,361 PPD
Eff. Time / Frame : 00:11:54 - 6,334 PPD

Up to 7000 ppd. About 6% slower than native Linux.
 
Brought up an old Ubuntu 8.04 VM on an XP32 rig with Q9450 at 3.44GHz and 2x9600GSO. The VM was ~18% better on p6701 and ~12% on p6075. The GPU were running and may have impacted results. Looks promising. Have brought up two additional VM (U910, U1004) on XP32 hosts, Q9450 with 2x9600GSO, and Q8400 with 9800GTX+9600GSO.

Project ID: 6701
Core: GRO-A3
Credit: 921
Frames: 100
Name: Pam-FAH1-P5Kpro-Q9450-XP32
Path: \\P5K-PRO-PAM\folding\FAH-1\
Number of Frames Observed: 300
Min. Time / Frame : 00:10:49 - 7,017 PPD
Avg. Time / Frame : 00:11:26 - 6,457 PPD

Name: Pam-VM1-P5Kpro-Q9450-U804
Path: \\PAM-VM1-U64-804\folding\FAH-SMP1\
Number of Frames Observed: 98
Min. Time / Frame : 00:10:07 - 7,757 PPD
Avg. Time / Frame : 00:10:12 - 7,663 PPD

Project ID: 6075
Core: GRO-A3
Credit: 481
Frames: 100
Name: Pam-FAH1-P5Kpro-Q9450-XP32
Path: \\P5K-PRO-PAM\folding\FAH-1\
Number of Frames Observed: 300
Min. Time / Frame : 00:04:49 - 8,826 PPD
Avg. Time / Frame : 00:05:05 - 8,141 PPD

Name: Pam-VM1-P5Kpro-Q9450-U804
Path: \\PAM-VM1-U64-804\folding\FAH-SMP1\
Number of Frames Observed: 36
Min. Time / Frame : 00:04:41 - 9,205 PPD
Avg. Time / Frame : 00:04:43 - 9,108 PPD
Cur. Time / Frame : 00:04:44 - 9,019 PPD
R3F. Time / Frame : 00:04:44 - 9,019 PPD
All Time / Frame : 00:04:43 - 9,051 PPD
Eff. Time / Frame : 00:04:51 - 8,802 PPD
 
In terms of time, your numbers are 11% faster for the VM on p6701 (VM finished in 89% of the time it took Win SMP) and 7% faster on p6075. I'm using the machine I'm testing on so my numbers are probably a little bit lower than yours as a result. I also have the VM running at idle priority when ungrabbed, so it does get squeezed for cpu cycles. It is promising indeed. I've got one more instance (VM/Ubuntu 8.04) at the office and plan on implementing another VM/Ubuntu 10.10 tomorrow.
 
plan on implementing another VM/Ubuntu 10.10 tomorrow.

Just an fyi.........I don't remember what the issue was/is, but i do remember there being an issue with Ubuntu 10.10 and folding.........

I'm sure your already aware. Thought i should throw it out there anyway.
 
You have to install nscd to make 10.10 work. For those interested:

#open a terminal
sudo apt-get install nscd
cd /etc
sudo chmod 666 nscd.conf
# open the nscd.conf file with a text editor
# find following line:
# enable-cache hosts no
# change "no" to "yes"
# save & exit the editor
sudo service nscd restart
 
Last edited:
Another data point on the VM compared to Win SMP, this time on a Q9450 @ 3.6 GHz running a 8800GTS (g92):

Code:
Project ID: 6076
 Core: GRO-A3
 Credit: 481
 Frames: 100


  Name: Val Win SMP
 Path: \\Val2\fah\FAH SMP\
 Number of Frames Observed: 300

 Min. Time / Frame : 00:04:32 - 9,666 PPD
 Avg. Time / Frame : 00:04:41 - 9,205 PPD


 Name: Val-VM 10.10
 Path: \\VAL-UBUNTU\fah\
 Number of Frames Observed: 96

 Min. Time / Frame : 00:04:21 - 10,283 PPD
 Avg. Time / Frame : 00:04:22 - 10,225 PPD
 Cur. Time / Frame : 00:04:23 - 10,136 PPD
 R3F. Time / Frame : 00:04:23 - 10,136 PPD
 All  Time / Frame : 00:04:22 - 10,174 PPD
 Eff. Time / Frame : 00:04:24 - 10,097 PPD

1000 ppd more on the Linux VM.
 
And another on a Q6600 @ 3.4:

Code:
 Project ID: 6072
 Core: GRO-A3
 Credit: 481
 Frames: 100


 Name: ChasR Win SMP
 Path: C:\Users\ChasR\Win FAH SMP\
 Number of Frames Observed: 300

 Min. Time / Frame : 00:05:08 - 8,022 PPD
 Avg. Time / Frame : 00:05:26 - 7,367 PPD


 Name: ChasR-VM
 Path: \\chasr-UBUNTU\fah\
 Number of Frames Observed: 94

 Min. Time / Frame : 00:05:17 - 7,683 PPD
 Avg. Time / Frame : 00:05:26 - 7,367 PPD
 Cur. Time / Frame : 00:05:22 - 7,432 PPD
 R3F. Time / Frame : 00:05:30 - 7,252 PPD
 All  Time / Frame : 00:05:26 - 7,341 PPD
 Eff. Time / Frame : 00:05:28 - 7,296 PPD

Nearly identical performance.
 
More data points for SMP on Ubuntu VM and XP32.
All the XP32 hosts are also folding on two G92 cards.

Project ID: 6050
Core: GRO-A3
Credit: 481
Frames: 100
Code:
 Name: Hayes-FAH1-P5Kpro-Q9450-XP32
 Path: \\P5K-PRO-HAYES\folding\FAH-1\
 Number of Frames Observed: 300
 Min. Time / Frame : 00:04:34 - 9,560 PPD
 Avg. Time / Frame : 00:04:46 - 8,965 PPD

 Name: Hayes-VM-P5Kpro-Q9450-U1004
 Path: \\HAYES-VM-U1004\folding\FAH-1\
 Number of Frames Observed: 100
 Min. Time / Frame : 00:04:26 - 9,995 PPD
 Avg. Time / Frame : 00:04:27 - 9,939 PPD

Project ID: 6058
Core: GRO-A3
Credit: 481
Frames: 100
Code:
 Name: F51-FAH1-P45Neo2-Q8400-XP32
 Path: \\F51-P45NEO-2\folding\FAH-1\
 Number of Frames Observed: 300
 Min. Time / Frame : 00:05:51 - 6,594 PPD
 Avg. Time / Frame : 00:06:07 - 6,167 PPD

 Name: F51-VM-P45neo2-Q8400-U910
 Path: \\F51-VM-U64-910\folding\FAH-1\
 Number of Frames Observed: 56
 Min. Time / Frame : 00:05:43 - 6,826 PPD
 Avg. Time / Frame : 00:05:45 - 6,767 PPD
 Cur. Time / Frame : 00:05:45 - 6,744 PPD
 R3F. Time / Frame : 00:05:44 - 6,764 PPD
 All  Time / Frame : 00:05:45 - 6,744 PPD
 Eff. Time / Frame : 00:05:49 - 6,667 PPD

Project ID: 6060
Core: GRO-A3
Credit: 481
Frames: 100
Code:
 Name: Hayes-FAH1-P5Kpro-Q9450-XP32
 Path: \\P5K-PRO-HAYES\folding\FAH-1\
 Number of Frames Observed: 300
 Min. Time / Frame : 00:04:31 - 9,720 PPD
 Avg. Time / Frame : 00:04:48 - 8,872 PPD

 Name: Hayes-VM-P5Kpro-Q9450-U1004
 Path: \\HAYES-VM-U1004\folding\FAH-1\
 Number of Frames Observed: 7
 Min. Time / Frame : 00:04:33 - 9,613 PPD
 Avg. Time / Frame : 00:04:33 - 9,613 PPD
 Cur. Time / Frame : 00:04:33 - 9,588 PPD
 R3F. Time / Frame : 00:04:33 - 9,588 PPD
 All  Time / Frame : 00:04:33 - 9,588 PPD
 Eff. Time / Frame : 00:04:53 - 8,934 PPD

Project ID: 6073
Core: GRO-A3
Credit: 481
Frames: 100
Code:
 Name: Pam-FAH1-P5Kpro-Q9450-XP32
 Path: \\P5K-PRO-PAM\folding\FAH-1\
 Number of Frames Observed: 300
 Min. Time / Frame : 00:04:46 - 8,965 PPD
 Avg. Time / Frame : 00:05:04 - 8,181 PPD

 Name: Pam-VM1-P5Kpro-Q9450-U804
 Path: \\PAM-VM1-U64-804\folding\FAH-SMP1\
 Number of Frames Observed: 100
 Min. Time / Frame : 00:04:45 - 9,012 PPD
 Avg. Time / Frame : 00:04:50 - 8,780 PPD

Project ID: 6075
Core: GRO-A3
Credit: 481
Frames: 100
Code:
 Name: Pam-FAH1-P5Kpro-Q9450-XP32
 Path: \\P5K-PRO-PAM\folding\FAH-1\
 Number of Frames Observed: 300
 Min. Time / Frame : 00:04:49 - 8,826 PPD
 Avg. Time / Frame : 00:05:05 - 8,141 PPD

 Name: Pam-VM1-P5Kpro-Q9450-U804
 Path: \\PAM-VM1-U64-804\folding\FAH-SMP1\
 Number of Frames Observed: 100
 Min. Time / Frame : 00:04:41 - 9,205 PPD
 Avg. Time / Frame : 00:04:46 - 8,965 PPD

Project ID: 6077
Core: GRO-A3
Credit: 481
Frames: 100
Code:
 Name: Pam-FAH1-P5Kpro-Q9450-XP32
 Path: \\P5K-PRO-PAM\folding\FAH-1\
 Number of Frames Observed: 300
 Min. Time / Frame : 00:04:46 - 8,965 PPD
 Avg. Time / Frame : 00:05:00 - 8,345 PPD

 Name: Pam-VM1-P5Kpro-Q9450-U804
 Path: \\PAM-VM1-U64-804\folding\FAH-SMP1\
 Number of Frames Observed: 47
 Min. Time / Frame : 00:04:44 - 9,060 PPD
 Avg. Time / Frame : 00:05:04 - 8,181 PPD
 Cur. Time / Frame : 00:04:46 - 8,834 PPD
 R3F. Time / Frame : 00:04:45 - 8,865 PPD
 All  Time / Frame : 00:05:04 - 8,311 PPD
 Eff. Time / Frame : 00:05:05 - 8,284 PPD
 
OK, so now VMWare Player 3/Linux is better than the WinSMP client again?

Seesaws can make you dizzy LOL :p

Good thing I didn't uninstall VMWare Player ;)

I'll have to read up on how to run the latest Linux clients.

... and also figure out how to OC my Q6600 to >3.2GHz stably under FAH 24/7. I can't seem to get it past 3.0GHz on my P965 chipset Commando.
 
Last edited:
If the swing is only ~1k PPD that won't get me off WinSMP... besides, no bigadv on Linux either. It's just to darn convenient not to have to "fool" with VMs.
 
I have the same feelings. I've changed over 4 machines, but got lazy when I found on some WUs there was no gain at all. It makes a bit more difference on Q9XXXs than on Q6600s.
 
If I ran in native Linux, would that make it more worth changing back to the new Linux client?
 
Definately, as long as you aren't gpu folding. What you gain on Linux SMP, you lose back by not being able to overclock the gpu without a bios flash. You still can't fold on the GPU on kernels above 2.6.28 (Ubuntu 9.04) without killing smp production.

Native Linux production on a Q6600 @ 3.24 GHz on one of the worst and best WUs:
Code:
Project ID: 6701
 Core: GRO-A3
 Credit: 921
 Frames: 100

 
 Name: CR L 1 
 Path: \\Cr-l1\fah\SMP\
 Number of Frames Observed: 300

 Min. Time / Frame : 00:10:35 - 7,250 PPD
 Avg. Time / Frame : 00:10:41 - 7,148 PPD

Project ID: 6050
 Core: GRO-A3
 Credit: 481
 Frames: 100


 Name: CR L 1
 Path: \\Cr-l1\fah\SMP\
 Number of Frames Observed: 200

 Min. Time / Frame : 00:04:45 - 9,012 PPD
 Avg. Time / Frame : 00:04:48 - 8,872 PPD
 
Last edited:
Since I'm not a -bigadv player, I suppose the best bet for bonus points on a3 work is to OC my Q6600 and not burden it with too many GPU clients.

I'd like to eventually add a GTX460 Fermi to my 9800GTX+ GPU2/3 folder, but I don't know how much I'd lose from a3 for the extra 10kPPD I'd gain. With an i7 -bigadv system it could be self defeating. With my Q6600 quad, I don't think I'd lose too much bonus as I'm only getting ~4-5kPPD from it now.
 
With the nVidia cards, there is always a substantial net gain from adding a 96 sp or better card. The net gain from adding the 460 would be well worth it, around 9000 ppd. Make sure you get one with good cooling. A lot of folks are having trouble with temps on new WUs. THe EVGA that I'm running stays below 50 though.
 
Back