AH! I'm infected!

Roman79

[H]ard|Gawd
Joined
Oct 2, 2000
Messages
1,393
All of but one my CPU clients (6) are now running either P2124 or P2125 (those 5 way melt ones). PPD is in the toilet :mad:

Anyone else getting a large number of these in the last few days?

 
My PPD/GHz for 2124 seem a bit lower than for other proteins (except the 7xx's, which really suck). I'm getting 55.6 ppd/ghz on my Semperon @2200. I've folded one 2125 and it was 54 ppd/ghz. My favorite protein for this boxen is the 2098 where I got 103 ppd/ghz.

 
All but 3 of my machines are running those 5 way melt ones. They take me 78hr/ unit to complete compared to 48 hrs for a 2414. I am running 2/3 the PPD that I was 5 days ago
 
P2124 732
P2125 203

huh, gues I've done one or two of these buggers... :D
 
p2124 : 63 folded
p2154 : 0 folded

total wus folded 6633
The wu that I have folded the most: p1808 : 224 folded.
Description:
These projects concern the folding of the triple helix of collagen. Collagen is a major component of bone, skin, and other organs. It is in fact the most abundant protein in humans, taking a quarter of the protein mass. Collagen has a unique structure consisting of three chains wrapped around each other, hence the name triple helix. Mutations in collagen cause misfolding of the triple helix, leading to diseases such as Osteogenesis imperfecta and Ehlers-Danlos syndrome. We are studying the folding mechanism and how mutations affect it.

 
Interesting, also most I've folded.


P1808 2755
P1809 1807
P2106 1458
P991 1226
 
3 of my boxen are running 2124's right now.
the Sempron 3200 doesn't mind it so much but the AXP 1800+ isn't too happy about it.
The 1.0 P3 thinks it's funny and has decided to take 7-8 days to return this WU to provide comedy relief.......
 
[BRO]Alaskan said:
It would be fine if I got 800points for them.

Heh yeah really!

It kills my PPD on all my newer CPUs (Athlon64's and Core Duos) but my older AXP 2500 is still maintaining a decent PPD.
 
I've got 3 5 way melt's running...not nearly as bad as this Protein 1666: p996_longfkcomp. It's on my 2.53Ghz P4 and will take 497 Hrs. EMIII Lists it at 6.2 PPD. 5 freaking hours per frame :( Should be worth loads more points.
 
Are you guys checking these after they've folded a couple frames and EMIII has taken the correct averages? It's interesting to see how the points are falling, but I'm not really a ppd person. I just let my computers chug along. I really only have access to 2 of my computers, and I never really check them. I once let my main computer(Xeon 2.6ghz) run for 2 months, and I looked at it and it was running without SSE boost :eek: . Not good for ppd - I never checked it after it shutdown and restarted incorrectly.

 
MN Scout said:
Are you guys checking these after they've folded a couple frames and EMIII has taken the correct averages? It's interesting to see how the points are falling, but I'm not really a ppd person. I just let my computers chug along. I really only have access to 2 of my computers, and I never really check them. I once let my main computer(Xeon 2.6ghz) run for 2 months, and I looked at it and it was running without SSE boost :eek: . Not good for ppd - I never checked it after it shutdown and restarted incorrectly.

Do me a favor.... let more of your boxen run that way....
It'll take you longer to pass me up that way. :)
 
I just checked to make sure all of mine were running correctly (and they are). The 5 way Melts are taking between 103hr and 13 min (92ppd) for an XP 2400 laptop and 68hr 55min (138ppd) on my X2 3800 OC to 2.5ghz
 
[BRO]Alaskan said:
It would be fine if I got 800points for them.

Why ????
They are just bog standard small Gromacs and are awarding the right points for the folding times.
So no big bonuses for folding them.
I've dropped from around 3k per day with the loss of all the big Gromacs, down to just over 2k per day, but still not dropped to my minimum speed of around 1.6k per day.

p2124, 80 done.
p2125, 14 done.

My max is p563, 553 folded of them.

Luck ......... :D
 
The top two for me:
P991 - 119
P1152 - 118

P2124 - 33
P2125 - 4

Here's what's running:
Box1.jpg

Box2.jpg

Box3.jpg

Box4.jpg

Box5.jpg

Box6.jpg

Box7.jpg
 
MN Scout said:
Are you guys checking these after they've folded a couple frames and EMIII has taken the correct averages? It's interesting to see how the points are falling, but I'm not really a ppd person. I just let my computers chug along. I really only have access to 2 of my computers, and I never really check them. I once let my main computer(Xeon 2.6ghz) run for 2 months, and I looked at it and it was running without SSE boost :eek: . Not good for ppd - I never checked it after it shutdown and restarted incorrectly.


Yes I've checked PPD at several stages through the WU and it still seems low.

How exactly does one insure SSE boost is enabled?
 
Roman79 said:
How exactly does one insure SSE boost is enabled?
With a big hammer of course.

you just unmount the HSF while it's under load and give the CPU a good smack. This will tell it to use SSE and the unknown.

</me prays that he's stupid enough to do this :)>
 
Roman79 said:
Yes I've checked PPD at several stages through the WU and it still seems low.

How exactly does one insure SSE boost is enabled?

Look at the log file in the Folding directory.

Code:
[02:56:02] *------------------------------*
[02:56:02] Folding@Home Gromacs Core
[02:56:02] Version 1.90 (March 8, 2006)
[02:56:02] 
[02:56:02] Preparing to commence simulation
[02:56:02] - Assembly optimizations manually forced on.
[02:56:02] - Not checking prior termination.
[02:56:02] - Expanded 41122 -> 263285 (decompressed 640.2 percent)
[02:56:02] - Starting from initial work packet
[02:56:02] 
[02:56:02] Project: 2130 (Run 1, Clone 86, Gen 9)
[02:56:02] 
[02:56:02] Assembly optimizations on if available.
[02:56:02] Entering M.D.
[02:56:08] Protein: p2130_ww2130
[02:56:08] 
[02:56:08] Writing local files
[02:56:08] Using table 3
[COLOR=Yellow][02:56:08] Extra SSE boost OK.[/COLOR]
[02:56:08] Writing local files
[02:56:08] Completed 0 out of 50000000 steps  (0)

Fold on!

 
Roman79 said:
How exactly does one insure SSE boost is enabled?

Use the -forceasm switch.
If folding stop due to either Windoze freezing/reboot or a power cut/reboot then when it restarts it runs without the boost unless that flag is set.
To turn the boost back on just stop-start the client by hand.

Luck .............. :D
 
Back
Top