• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Hordemember and Time per frame question

S

st4rk

Gawd
Joined
Sep 19, 2003
Messages
1,013
Just wondering if these times per frame sound right.

On my q6600 (3.0 oc), it does 9 minutes per frame running the windows smp client.
On my X2 4400+ (stock), it does 14 minutes per frame in xubuntu 64bit.
On a xeon 2.8 in windows 2000 server, it does ~50 minutes per frame with the console client? Seems long....

As for joining the horde on hardfolding.com, how do you do it? I can log in and everything, but I'm not showing up on the memberlist.
 
First of all, welcome to the team :)

I can just comment on the Q6600. 8-9 mins per frame is right for your speed under Windows. However, for the console client, 50 mins per frame isn't unusual since it has a deadline of 2-3 months anyway and you should complete it within 4 days.

As for the stats, if you just joined, just wait till the first work unit show up on the stats then you will see it. Also, make sure you configured the clients to run with team 33 in addition to your username.

 
Welcome to the Horde!

Are you running the SMP client on the 4400, If so that is a pretty good frame time.

 
Welcome to the [H]ardest Team in the World! Fold on for the Cure! :D

After you turn in some points... come back and post your badge. There is a badge thread where you can test whether your badge works.
 
If you mean logging into the hardfolding.com front page....if you're already registered, throw your login info in the area, then after it refreshes, it should say Welcome <username> above the login boxes. For whatever reason, they leave the login boxes there even when you are logged into the site.

You can also verify this by trying to edit your account on the right side of the page near the top.

 
Hmmm, looks like the X2 has surpassed the Q6600 in minutes per frame! It's now doing 6-7 minutes per frame.

Waiting for the next WU to see if it still goes insanely fast.
 
Can you tell the Project Clone Run Gen number for this "fast" one ?

I heard that under linux, there is a Project which doesn't give lots of points but is incredibly fast. 0-100% in under 8 hours for a Q6600.

 
Here is the log from the fast WU:

23:16:11]
[23:16:11] *------------------------------*
[23:16:11] Folding@Home Gromacs SMP Core
[23:16:11] Version 1.73 (November 27, 2006)
[23:16:11]
[23:16:11] Preparing to commence simulation
[23:16:11] - Ensuring status. Please wait.
[23:16:11] - Starting from initial work packet
[23:16:11]
[23:16:11] Project: 2605 (Run 16, Clone 134, Gen 12)
[23:16:11]
[23:16:12] Assembly optimizations on if available.
[23:16:12] Entering M.D.
[23:16:28] percent)
[23:16:28] - Starting from initial work packet
[23:16:28]
[23:16:28] Project: 2605 (Run 16, Clone 134, Gen 12)
[23:16:28]
[23:16:28] Entering M.D.
NNODES=4, MYRANK=1, HOSTNAME=xubuntu2
NNODES=4, MYRANK=3, HOSTNAME=xubuntu2
NNODES=4, MYRANK=0, HOSTNAME=xubuntu2
NNODES=4, MYRANK=2, HOSTNAME=xubuntu2
NODEID=3 argc=15
NODEID=2 argc=15
NODEID=0 argc=15
NODEID=1 argc=15
Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2004, The GROMACS development team,
check out http://www.gromacs.org for more information.

This inclusion of Gromacs code in the Folding@Home Core is under
a special license (see http://folding.stanford.edu/gromacs.html)
specially granted to Stanford by the copyright holders. If you
are interested in using Gromacs, visit www.gromacs.org where
you can download a free version of Gromacs under
the terms of the GNU General Public License (GPL) as published
by the Free Software Foundation; either version 2 of the License,
or (at your option) any later version.

starting mdrun 'Protein in POPC'
500000 steps, 1000.0 ps.

[23:16:35] g local files
[23:16:35] Extra SSE boost OK.
[23:16:35] ocal files
[23:16:35] t)
[23:16:35] ra SSE boost OK.
[23:22:34] es
[23:22:34] Completed 5000 out of 500000 steps (1 percent)
[23:28:37] Writing local files
[23:28:37] Completed 10000 out of 500000 steps (2 percent)
[23:34:39] Writing local files
[23:34:39] Completed 15000 out of 500000 steps (3 percent)
[23:40:38] Writing local files
[23:40:38] Completed 20000 out of 500000 steps (4 percent)
[23:46:31] Writing local files
[23:46:31] Completed 25000 out of 500000 steps (5 percent)
[23:52:23] Writing local files
[23:52:24] Completed 30000 out of 500000 steps (6 percent)
 
wow....i'm running about 27/frame on windows with a 4800, but i've only been getting stuck with 2653s.....since the 2605 is the same amount of pts, it might be worth throwing a linux partition in just to run smp when that pc isn't in use if the times are that much faster :D

 
I did another test to see how much faster linux is on other machines. One borg is an old p4 2.8ghz HT. On the windows console client it cranked frames out at about 50 minutes a frame! lolslow

So I removed windows and put on 32bit native xubuntu. Same exact results :( Still slow.

I have a coreduo 1.8 borg which was running the windows smp client and doing 22min/frame. I took that off and loaded xubuntu 64bit in vmware. Same result with that one as well, no speed increase.

The non HT 2.8 P4's do ~22min a frame with the windows console, yet the HT p4 2.8 does 50 minutes? The HT p4 machine even has more ram. Weird! I'll have to load cpuz and see what the differences between the procs are, maybe cache? I'd expect the HT p4 to be working alot faster.

At home I will see exactly how much faster/slower vmware xubuntu 64bit instances are than smp windows client on my q6600.

*edit* Aha! Looks like even on windows some WU's go really fast. I just checked up on a non HT p4 2.8 and it's doing 10 minutes per frame on this one WU:

[02:09:16] Preparing to commence simulation
[02:09:16] - Files status OK
[02:09:16] - Expanded 151003 -> 542261 (decompressed 359.1 percent)
[02:09:16]
[02:09:16] Project: 3852 (Run 4115, Clone 0, Gen 4)
[02:09:16]
[02:09:16] Assembly optimizations on if available.
[02:09:16] Entering M.D.
[02:09:22] Working on p3850_fkbprelative_ligand
[02:09:22] Completed 0 out of 1250000 steps (0)
[02:09:22] Extra SSE2 boost OK
[02:20:42] Completed 12500 out of 1250000 steps (1)
[02:32:03] Completed 25000 out of 1250000 steps (2)
[02:43:22] Completed 37500 out of 1250000 steps (3)
[02:54:42] Completed 50000 out of 1250000 steps (4)

So it's only a matter of time before the X2 gets one of those WU that puts it back at 20+ min/frame.

But still doesn't explain why the HT p4 is going so frickin slow.
 
Just finished installing Xubuntu on my 4800....it's going on 2605(run 8, clone 124, gen 21) on the smp core, so I'll let you know my times when they start coming in to compare against yours in that general WU

Edit: ~23 minutes per frame through 7 percent....looks like i didn't get one of the incredibly fast runs like the one you got, but I'll see once it gets to another WU
 
I believe the very fast one might be either buggy or is using some erroneous parameters. Just be happy to crunch this and I'm sure Stanford will notice the discrepancy when they will analyze the results of this WU.

The other averages is more in line with those processors.
 
schenksmill said:
Just finished installing Xubuntu on my 4800....it's going on 2605(run 8, clone 124, gen 21) on the smp core, so I'll let you know my times when they start coming in to compare against yours in that general WU

Edit: ~23 minutes per frame through 7 percent....looks like i didn't get one of the incredibly fast runs like the one you got, but I'll see once it gets to another WU
Click to expand...

When you started the client, did you launch it with -smp on the end? I made that mistake before and was running WU's with only 50% of my cpu.
 
schenksmill said:
Just finished installing Xubuntu on my 4800....it's going on 2605(run 8, clone 124, gen 21) on the smp core, so I'll let you know my times when they start coming in to compare against yours in that general WU

Edit: ~23 minutes per frame through 7 percent....looks like i didn't get one of the incredibly fast runs like the one you got, but I'll see once it gets to another WU
Click to expand...

23 per frame is dead on for a 4800+ on a P2605 under linux.
X2 3800+ @ 2.0 runs one frame in 27:04 (936 PPD)
Opty170 @ 2.85 runs one frame in 18:48 (1348 PPD)



 
st4rk said:
When you started the client, did you launch it with -smp on the end? I made that mistake before and was running WU's with only 50% of my cpu.
Click to expand...

yea it's definitely running right...the run you nailed is probably just that much faster...if/when i get one in the same run i can update my #s
 
You must log in or register to reply here.
Back
Top