• 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.

QMD Core, is this normal?

L

Leadman584

Limp Gawd
Joined
Jun 10, 2004
Messages
448
Got this strange looking logfile segment.

[04:15:06] Entering QMD...
[04:15:28] System: p1902_32_water_molecules
[04:15:28]
[04:15:28] Performing initial WF calculations
[04:15:54] Completed 0 out of 2500 steps (0)
[04:22:14] Completed 25 out of 2500 steps (1)
[04:23:23] Writing checkpoint files
[04:23:52] WF converged, jumping to MD
[04:23:52] Verifying checksum
[04:24:13] Completed 500 out of 2500 steps (20)
[04:32:22] Completed 525 out of 2500 steps (21)

WTF!

 
I believe this is the new core 96 that went out this week. According to the Stanford site, those are water molocules for that core, before the real heavy hitters come out. It looks like it might be a different method for the folding.

 
p[H]ant0m said:
I believe this is the new core 96 that went out this week. According to the Stanford site, those are water molocules for that core, before the real heavy hitters come out. It looks like it might be a different method for the folding.

Click to expand...

Its a double precision Gromac.
Stock up ram boys the heavy hitters are comming out! :eek:

 
My question really is. Why the heck it skipped 19 steps, and kept right on going, like nothing happened?

 
Leadman584 said:
My question really is. Why the heck it skipped 19 steps, and kept right on going, like nothing happened?

Click to expand...
Whoa... I just noticed that... maybe its just a new core glitch?

 
Leadman584 said:
My question really is. Why the heck it skipped 19 steps, and kept right on going, like nothing happened?

Click to expand...

Was this right after it downloaded it and started...that is strange!
I noticed that but thought it had run a bit and was picking back up after a restart or something!

[H] Folding
 
Found a GB of PC3200 RAM, for $100 here:

http://shop3.outpost.com/product/4366625

Don't even wanna imagine the SPD on it. 5,9,9,15 would be my best guess. Decent RAM usually costs $300+ per GB. If someone here goes for this deal, I'd be curious to hear the actual SPD. These newest cores will require us to have at least a GB of RAM. Silly thing is using 194 MB of RAM, right now.

 
The odd behavior you see with the QMD core is normal.
The way it works is a lot different...
That's one of the reasons EM finds it difficult to track.

I'm working with the Stanford team to try and make sure tracking of some kind is available for this work unit type.

Keep on folding!
 
LPerry said:
The odd behavior you see with the QMD core is normal.
The way it works is a lot different...
That's one of the reasons EM finds it difficult to track.

I'm working with the Stanford team to try and make sure tracking of some kind is available for this work unit type.

Keep on folding!
Click to expand...

Cool thanks for the info!

 
Leadman584 said:
Found a GB of PC3200 RAM, for $100 here:

http://shop3.outpost.com/product/4366625

Don't even wanna imagine the SPD on it. 5,9,9,15 would be my best guess. Decent RAM usually costs $300+ per GB. If someone here goes for this deal, I'd be curious to hear the actual SPD. These newest cores will require us to have at least a GB of RAM. Silly thing is using 194 MB of RAM, right now.

Click to expand...

Kingston valueram can be had for 110 for 1gb and it runs 3-3-3 which isn't to shabby...not great but for 110.....you know.

Or this corsair for 117 @cas2.5:
http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=80098-21&ps=ho1
 
Whoa, just picked up the 450 point QMD core, project 1910. FAHcore 96 using 330MB of RAM. I'd hate to see one of these puppies hit a system with 512MB or less RAM.

 
Leadman584 said:
Whoa, just picked up the 450 point QMD core, project 1910. FAHcore 96 using 330MB of RAM. I'd hate to see one of these puppies hit a system with 512MB or less RAM.

Click to expand...

Do you have any sort of information on time per frame, or estimated time to completion? I haven't gotten any yet, but I am really curious as to how long these monsters are going to take.


 
I'll have to check the RAM usage when I get home tonight, but when I left for work, my P4 had 2 of these 450 point monsters! good thing I have a GB of RAM in there!!

I'll have to post the log file when I get home... I looked this morning and it had some weird things going on... different from the one I posted here.... interesting to say the least...


Keep on Folding!!

 
This was on a 2.4 Celeron 1.5 gigs ram in single channel (don't ask....)

Check my math, I think this is 240 PPD off a celeron!!!!

327 meg mem used


[13:30:22] Core required: FahCore_96.exe
[13:30:23] Folding@Home QMD Core
[13:30:23] Version 1.02 (Mar 7, 2005)
[13:30:23] Project: 1910 (Run 5, Clone 128, Gen 0)
[13:30:23] Writing local files
[13:30:23] Entering QMD...
[13:31:13] System: p1910_ALA-depeptide
[13:31:13] Performing initial WF calculations
[13:33:03] Completed 0 out of 2500 steps (0)
[13:45:48] Writing local files (checkpoint)
[13:59:56] Completed 25 out of 2500 steps (1)
[14:00:59] Writing local files (checkpoint)

edit: cut some junk out
 
It skipped 17 frames and is proceding normally. 12:20 per step on a 3.2 P4. Off to work.

 
More weirdness

[17:48:25] Performing initial WF calculations
[17:49:44] Completed 0 out of 2500 steps (0)
[18:02:36] Writing local files (checkpoint)
[18:09:58] Completed 25 out of 2500 steps (1)
[18:17:58] Writing local files (checkpoint)
[18:30:00] Completed 50 out of 2500 steps (2)
[18:33:10] Writing local files (checkpoint)
[18:48:31] Writing local files (checkpoint)
[18:50:13] Completed 75 out of 2500 steps (3)
[19:04:03] Writing local files (checkpoint)
[19:05:45] Writing local files (checkpoint)
[19:05:50] WF converged, jumping to MD
[19:05:50] Verifying checksum
[19:07:01] Completed 500 out of 2500 steps (20)
[19:21:31] Writing local files (checkpoint)
 
Don't watch this QMD pot too closely...

All this stuff is "Normal" for QMD. Weird, but normal. :cool:
 
I'm guessing you have to tell fah you want these massive WU in the config?? I only have about 4 of the 15+ machines that have 512mb ram or better...:eek:
 
Yup... during the config, you have to say "yes" to the big packets...
once you have that set, you can get the big 600 pointers... the 450 pointers,etc... F@H won't give your PC something it can't handle (memory-wise)... it's part of the benchmark


Keep on Folding!!

 
These process pretty quickly on my Xeon. I only have two of the four F@H processes set up to recieve these, since I only have 1 gig of mem. I believe that you need big packets and advmethods enable to get these QMD wus. Food on!

MN Scout
 
yup.. I forgot about the -advmethods flag...


Keep on Folding!!

 
Leadman584 said:
Found a GB of PC3200 RAM, for $100 here:

http://shop3.outpost.com/product/4366625

Don't even wanna imagine the SPD on it. 5,9,9,15 would be my best guess. Decent RAM usually costs $300+ per GB. If someone here goes for this deal, I'd be curious to hear the actual SPD. These newest cores will require us to have at least a GB of RAM. Silly thing is using 194 MB of RAM, right now.

Click to expand...

As much as Stanford makes a lot of silly assumptions about what we fold on or with even they can’t be so off the wall as to assume most people will actually have a gig-o-ram in their machines.

The whole idea is predicated on still being able to use your machine in real life as well as fold. Those of us with dedicated folding machines are a bonus for them.

I find it strange that Stanford keeps pushing for work units to be returned ASAP but at the same time they keep pushing the envelope on bigger and slower projects.

 
BillR said:
As much as Stanford makes a lot of silly assumptions about what we fold on or with even they can’t be so off the wall as to assume most people will actually have a gig-o-ram in their machines.

The whole idea is predicated on still being able to use your machine in real life as well as fold. Those of us with dedicated folding machines are a bonus for them.

I find it strange that Stanford keeps pushing for work units to be returned ASAP but at the same time they keep pushing the envelope on bigger and slower projects.

Click to expand...


I think they are pushing becasue they know they have people like us who have dedicated folding boxes so they keep putting those out for the addicted ones and the rest of the wu's are there for the casual groups.
 
You must log in or register to reply here.
Back
Top