Performance issues...no, not those kind :)

F

freeloader1969

2[H]4U
Joined
Dec 13, 2003
Messages
2,238
I just built a dual G34 system (Asus mobo, dual 6128's) and have Ubuntu 10.10 (kraken installed) running in Virtualbox.

Just finished a project 6996 and got the following results after the run. These results seem awfully low for some reason. Can anyone with more experience in multi proc setups tell me if this looks right?

]
Capture-1.jpg
 
Vbox may be slowing things down, but the chips are not the world's fastes. what PPD are you currently getting as my only linux rig is chewing through 6901 & 6903 Wu at the moment?
 
Nathan_P said:
Vbox may be slowing things down, but the chips are not the world's fastes. what PPD are you currently getting as my only linux rig is chewing through 6901 & 6903 Wu at the moment?
Click to expand...

I'm doing about 20,000 PPD. I'm not running any bigadv units as I don't think the current CPUS would be fast enough to complete a unit on time. Is that incorrect? I'm planning on moving to two 12 core chips within a few weeks.
 
i think i remember reading the -bigbeta flag will now give you the old big adv work units which should run fine.
 
Just received a 6903 unit. I will report back with times for comparison. :) Thanks for the bigbeta advice.
 
Edit: Just re-read it's a dual not quad.
Useless info from me..... :eek:
 
Last edited:
I think I get like 25-26k PPD on SMP on my dual 6128 but that's with the 2nd chip with no memory on it.
 
Here's a 6903 after a few hours....

Capture-2.jpg



Do those times seem right? If not, I'll delete the Windows install and try Ubuntu (with Kraken) directly.
 
My old quad 8389 16c combo got better times. I would definitely try punting the windows install and try the Ubuntu setup. The setup guide for it is pretty bulletproof at this point.
 
Deleted the Windows install and installed latest version of Ubuntu. Now I can't get a work unit. It's been hanging there for the last five minutes.

Code: 
14:22:22] Work directory not found. Creating...
[14:22:22] Loaded queue successfully.
[14:22:22] - Preparing to get new work unit...
[14:22:22] - Autosending finished units... [June 27 14:22:22 UTC]
[14:22:22] Cleaning up work directory
[14:22:22] Trying to send all finished work units
[14:22:22] + No unsent completed units remaining.
[14:22:22] + Attempting to get work packet
[14:22:22] - Autosend completed
[14:22:22] Passkey found
[14:22:22] - Will indicate memory of 16048 MB
[14:22:22] - Connecting to assignment server
[14:22:22] Connecting to http://assign.stanford.edu:8080/
[14:22:22] Posted data.
[14:22:22] Initial: ED82; - Successful: assigned to (130.237.232.237).
[14:22:22] + News From Folding@Home: Welcome to Folding@Home
[14:22:23] Loaded queue successfully.
[14:22:23] Sent data
[14:22:23] Connecting to http://130.237.232.237:8080/
[14:22:36] Posted data.
[14:22:36] Initial: 0000; - Receiving payload (expected size: 57242762)

*** Never mind, it just started up now *** Hopefully I'll see tpf in 1 hour or so. :)
 
Last edited:
You're getting 57MB unit. How long does it take to pull 57MB w/your link?

Here's my last one -- almost 8 minutes (looks like servers are a big hogged;
typically, it's much faster):
Code: 
[14:05:51] Initial: 0000; - Receiving payload (expected size: 57209289)
[14:13:35] - Downloaded at ~120 kB/s
 
tear said:
You're getting 57MB unit. How long does it take to pull 57MB w/your link?

Here's my last one -- almost 8 minutes (looks like servers are a big hogged;
typically, it's much faster):
Code: 
[14:05:51] Initial: 0000; - Receiving payload (expected size: 57209289)
[14:13:35] - Downloaded at ~120 kB/s
Click to expand...


Thanks for the heads up. It started now. Tear, what tpf should I expect with this setup? Dual 6128's/16GB Ram/quad channel CL 8-8-8/NUMA ON/Kraken installed.
 
1 hour and 5 minutes TPF P6904 for a PPD of ~83k.
47.5 minute TPF P6903 for a PPD of ~81k.
 
You must log in or register to reply here.
Back
Top