New Bonus Amber Core WU's [-advmethods option]

Moderators: Site Moderators, FAHC Science Team

vvoelz
Pande Group Member
Posts: 552
Joined: Sun Dec 02, 2007 8:07 pm
Location: Temple University, Philadelphia PA

Re: New Bonus Amber Core WU's [-advmethods option]

Post by vvoelz »

4411 is not one of my "bonus" AMBER projects, but I'll forward this link onto the person in charge of this project... Hopefully they will have some insight --Vince
xhuang
Pande Group Member
Posts: 81
Joined: Fri Nov 30, 2007 9:38 pm
Location: Hong Kong

Re: New Bonus Amber Core WU's [-advmethods option]

Post by xhuang »

Leoslocks, the results for WU you have for p4411 has exceeded the 5MB limit, so that it does not upload. p4411 has been running for a while, and we did not see this issue before. This might be something related to this particular WU, and we are looking into it now.
toTOW
Site Moderator
Posts: 6349
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France
Contact:

Re: Length of work exceeds packet limit

Post by toTOW »

Leoslocks wrote:Is this a Core issue or a network issue? (or another issue)

[12:01:13] - Error: Length of work/wuresults_01.dat (44968604) exceeds packet limit set (26213376)

Code: Select all

[09:40:29] Project: 4411 (Run 26, Clone 14, Gen 80)
[09:40:29] 
[09:40:29] Assembly optimizations on if available.
[09:40:29] Entering M.D.
[09:40:35] Protein: p4411_Seq_46_unf_AMBER
[09:40:35] 
[09:40:35] Writing local files
[09:42:36] Extra SSE boost OK.
[09:42:36] Writing local files
[09:42:36] Completed 0 out of 150000 steps  (0%)
[09:43:58] Writing local files
[09:43:58] Completed 1500 out of 150000 steps  (1%)
[09:45:19] Writing local files
[09:45:19] Completed 3000 out of 150000 steps  (2%)
[09:46:41] Writing local files
[09:46:41] Completed 4500 out of 150000 steps  (3%)
[09:48:03] Writing local files
[09:48:03] Completed 6000 out of 150000 steps  (4%)
[09:49:26] Writing local files
[09:49:26] Completed 7500 out of 150000 steps  (5%)
[09:50:48] Writing local files
[09:50:48] Completed 9000 out of 150000 steps  (6%)
[09:52:10] Writing local files
[09:52:10] Completed 10500 out of 150000 steps  (7%)
[09:53:32] Writing local files
[09:53:32] Completed 12000 out of 150000 steps  (8%)
[09:54:55] Writing local files
[09:54:55] Completed 13500 out of 150000 steps  (9%)
[09:56:16] Writing local files
[09:56:16] Completed 15000 out of 150000 steps  (10%)
[09:57:40] Writing local files
[09:57:40] Completed 16500 out of 150000 steps  (11%)
[09:59:02] Writing local files
[09:59:02] Completed 18000 out of 150000 steps  (12%)
[10:00:24] Writing local files
[10:00:24] Completed 19500 out of 150000 steps  (13%)
[10:01:46] Writing local files
[10:01:46] Completed 21000 out of 150000 steps  (14%)
[10:03:09] Writing local files
[10:03:09] Completed 22500 out of 150000 steps  (15%)
[10:04:31] Writing local files
[10:04:31] Completed 24000 out of 150000 steps  (16%)
[10:05:53] Writing local files
[10:05:53] Completed 25500 out of 150000 steps  (17%)
[10:07:15] Writing local files
[10:07:15] Completed 27000 out of 150000 steps  (18%)
[10:08:38] Writing local files
[10:08:38] Completed 28500 out of 150000 steps  (19%)
[10:10:01] Writing local files
[10:10:01] Completed 30000 out of 150000 steps  (20%)
[10:11:23] Writing local files
[10:11:23] Completed 31500 out of 150000 steps  (21%)
[10:12:45] Writing local files
[10:12:45] Completed 33000 out of 150000 steps  (22%)
[10:14:08] Writing local files
[10:14:08] Completed 34500 out of 150000 steps  (23%)
[10:15:31] Writing local files
[10:15:31] Completed 36000 out of 150000 steps  (24%)
[10:16:53] Writing local files
[10:16:53] Completed 37500 out of 150000 steps  (25%)
[10:18:15] Writing local files
[10:18:15] Completed 39000 out of 150000 steps  (26%)
[10:19:36] Writing local files
[10:19:36] Completed 40500 out of 150000 steps  (27%)
[10:20:58] Writing local files
[10:20:58] Completed 42000 out of 150000 steps  (28%)
[10:22:19] Writing local files
[10:22:19] Completed 43500 out of 150000 steps  (29%)
[10:23:42] Writing local files
[10:23:42] Completed 45000 out of 150000 steps  (30%)
[10:25:06] Writing local files
[10:25:06] Completed 46500 out of 150000 steps  (31%)
[10:26:28] Writing local files
[10:26:28] Completed 48000 out of 150000 steps  (32%)
[10:27:49] Writing local files
[10:27:49] Completed 49500 out of 150000 steps  (33%)
[10:29:11] Writing local files
[10:29:11] Completed 51000 out of 150000 steps  (34%)
[10:30:33] Writing local files
[10:30:33] Completed 52500 out of 150000 steps  (35%)
[10:31:55] Writing local files
[10:31:55] Completed 54000 out of 150000 steps  (36%)
[10:33:17] Writing local files
[10:33:17] Completed 55500 out of 150000 steps  (37%)
[10:34:41] Writing local files
[10:34:41] Completed 57000 out of 150000 steps  (38%)
[10:36:03] Writing local files
[10:36:03] Completed 58500 out of 150000 steps  (39%)
[10:37:26] Writing local files
[10:37:26] Completed 60000 out of 150000 steps  (40%)
[10:38:47] Writing local files
[10:38:47] Completed 61500 out of 150000 steps  (41%)
[10:40:10] Writing local files
[10:40:10] Completed 63000 out of 150000 steps  (42%)
[10:41:32] Writing local files
[10:41:32] Completed 64500 out of 150000 steps  (43%)
[10:42:55] Writing local files
[10:42:55] Completed 66000 out of 150000 steps  (44%)
[10:44:17] Writing local files
[10:44:17] Completed 67500 out of 150000 steps  (45%)
[10:45:36] Writing local files
[10:45:36] Completed 69000 out of 150000 steps  (46%)
[10:46:52] Writing local files
[10:46:52] Completed 70500 out of 150000 steps  (47%)
[10:48:16] Writing local files
[10:48:16] Completed 72000 out of 150000 steps  (48%)
[10:49:38] Writing local files
[10:49:38] Completed 73500 out of 150000 steps  (49%)
[10:51:01] Writing local files
[10:51:01] Completed 75000 out of 150000 steps  (50%)
[10:52:24] Writing local files
[10:52:24] Completed 76500 out of 150000 steps  (51%)
[10:53:47] Writing local files
[10:53:47] Completed 78000 out of 150000 steps  (52%)
[10:55:10] Writing local files
[10:55:10] Completed 79500 out of 150000 steps  (53%)
[10:56:33] Writing local files
[10:56:33] Completed 81000 out of 150000 steps  (54%)
[10:57:57] Writing local files
[10:57:57] Completed 82500 out of 150000 steps  (55%)
[10:59:21] Writing local files
[10:59:21] Completed 84000 out of 150000 steps  (56%)
[11:00:45] Writing local files
[11:00:45] Completed 85500 out of 150000 steps  (57%)
[11:02:07] Writing local files
[11:02:07] Completed 87000 out of 150000 steps  (58%)
[11:03:30] Writing local files
[11:03:30] Completed 88500 out of 150000 steps  (59%)
[11:04:52] Writing local files
[11:04:52] Completed 90000 out of 150000 steps  (60%)
[11:06:15] Writing local files
[11:06:15] Completed 91500 out of 150000 steps  (61%)
[11:07:37] Writing local files
[11:07:37] Completed 93000 out of 150000 steps  (62%)
[11:09:01] Writing local files
[11:09:01] Completed 94500 out of 150000 steps  (63%)
[11:10:24] Writing local files
[11:10:24] Completed 96000 out of 150000 steps  (64%)
[11:11:47] Writing local files
[11:11:47] Completed 97500 out of 150000 steps  (65%)
[11:13:11] Writing local files
[11:13:11] Completed 99000 out of 150000 steps  (66%)
[11:14:33] Writing local files
[11:14:33] Completed 100500 out of 150000 steps  (67%)
[11:15:57] Writing local files
[11:15:57] Completed 102000 out of 150000 steps  (68%)
[11:17:19] Writing local files
[11:17:19] Completed 103500 out of 150000 steps  (69%)
[11:18:42] Writing local files
[11:18:42] Completed 105000 out of 150000 steps  (70%)
[11:20:05] Writing local files
[11:20:05] Completed 106500 out of 150000 steps  (71%)
[11:21:29] Writing local files
[11:21:29] Completed 108000 out of 150000 steps  (72%)
[11:22:52] Writing local files
[11:22:52] Completed 109500 out of 150000 steps  (73%)
[11:24:15] Writing local files
[11:24:15] Completed 111000 out of 150000 steps  (74%)
[11:25:37] Writing local files
[11:25:37] Completed 112500 out of 150000 steps  (75%)
[11:27:00] Writing local files
[11:27:00] Completed 114000 out of 150000 steps  (76%)
[11:28:23] Writing local files
[11:28:23] Completed 115500 out of 150000 steps  (77%)
[11:29:45] Writing local files
[11:29:45] Completed 117000 out of 150000 steps  (78%)
[11:31:07] Writing local files
[11:31:07] Completed 118500 out of 150000 steps  (79%)
[11:32:30] Writing local files
[11:32:30] Completed 120000 out of 150000 steps  (80%)
[11:33:52] Writing local files
[11:33:52] Completed 121500 out of 150000 steps  (81%)
[11:35:15] Writing local files
[11:35:15] Completed 123000 out of 150000 steps  (82%)
[11:36:37] Writing local files
[11:36:37] Completed 124500 out of 150000 steps  (83%)
[11:37:59] Writing local files
[11:37:59] Completed 126000 out of 150000 steps  (84%)
[11:39:22] Writing local files
[11:39:22] Completed 127500 out of 150000 steps  (85%)
[11:40:45] Writing local files
[11:40:45] Completed 129000 out of 150000 steps  (86%)
[11:42:07] Writing local files
[11:42:07] Completed 130500 out of 150000 steps  (87%)
[11:43:30] Writing local files
[11:43:30] Completed 132000 out of 150000 steps  (88%)
[11:44:52] Writing local files
[11:44:52] Completed 133500 out of 150000 steps  (89%)
[11:46:16] Writing local files
[11:46:16] Completed 135000 out of 150000 steps  (90%)
[11:47:38] Writing local files
[11:47:38] Completed 136500 out of 150000 steps  (91%)
[11:49:01] Writing local files
[11:49:01] Completed 138000 out of 150000 steps  (92%)
[11:50:25] Writing local files
[11:50:25] Completed 139500 out of 150000 steps  (93%)
[11:51:47] Writing local files
[11:51:47] Completed 141000 out of 150000 steps  (94%)
[11:53:09] Writing local files
[11:53:09] Completed 142500 out of 150000 steps  (95%)
[11:54:31] Writing local files
[11:54:32] Completed 144000 out of 150000 steps  (96%)
[11:55:54] Writing local files
[11:55:54] Completed 145500 out of 150000 steps  (97%)
[11:57:17] Writing local files
[11:57:17] Completed 147000 out of 150000 steps  (98%)
[11:58:39] Writing local files
[11:58:39] Completed 148500 out of 150000 steps  (99%)
[12:00:01] Writing local files
[12:00:01] Completed 150000 out of 150000 steps  (100%)
[12:00:01] Writing final coordinates.
[12:00:01] Past main M.D. loop
[12:01:01] 
[12:01:01] Finished Work Unit:
[12:01:01] - Reading up to 377616 from "work/wudata_09.arc": Read 377616
[12:01:01] - Reading up to 169592 from "work/wudata_09.xtc": Read 169592
[12:01:01] goefile size: 0
[12:01:01] logfile size: 19222
[12:01:01] Leaving Run
[12:01:02] - Writing 1191293 bytes of core data to disk...
[12:01:02] Done: 1190781 -> 592033 (compressed to 49.7 percent)
[12:01:02]   ... Done.
[12:01:02] - Shutting down core
[12:01:02] 
[12:01:02] Folding@home Core Shutdown: FINISHED_UNIT
[12:01:04] CoreStatus = 64 (100)
[12:01:04] Sending work to server


[12:01:04] + Attempting to send results
[12:01:09] + Results successfully sent
[12:01:09] Thank you for your contribution to Folding@Home.
[12:01:09] + Number of Units Completed: 186

[12:01:13] - Error: Length of work/wuresults_01.dat (44968604) exceeds packet limit set (26213376)
[12:01:13] - Error: Could not transmit unit 01 (completed May 9) to work server.
[12:01:13] - Error: Length of work/wuresults_01.dat (44968604) exceeds packet limit set (26213376)
[12:01:13]   Could not transmit unit 01 to Collection server; keeping in queue.
[12:01:13] - Preparing to get new work unit...
[12:01:13] + Attempting to get work packet
[12:01:13] - Connecting to assignment server
[12:01:13] - Successful: assigned to (171.64.65.65).
[12:01:13] + News From Folding@Home: Welcome to Folding@Home
[12:01:13] Loaded queue successfully.
[12:01:25] - Error: Length of work/wuresults_01.dat (44968604) exceeds packet limit set (26213376)
[12:01:25] - Error: Could not transmit unit 01 (completed May 9) to work server.
[12:01:25] - Error: Length of work/wuresults_01.dat (44968604) exceeds packet limit set (26213376)
[12:01:25]   Could not transmit unit 01 to Collection server; keeping in queue.
[12:01:25] + Closed connections
[12:01:25] 
[12:01:25] + Processing work unit
[12:01:25] Core required: FahCore_78.exe
[12:01:25] Core found.
[12:01:25] Working on Unit 00 [May 13 12:01:25]
[12:01:25] + Working ...
[12:01:25] 
[12:01:25] *------------------------------*
qfix will fix your queue and you'll be able to send the WU ;)
Image

Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.
vvoelz
Pande Group Member
Posts: 552
Joined: Sun Dec 02, 2007 8:07 pm
Location: Temple University, Philadelphia PA

Re: New Bonus Amber Core WU's [-advmethods option]

Post by vvoelz »

The -advmethods "bonused" AMBER Core WUs just got more bonus.

Starting with projects 4534-6 (to come online shortly), the bonus points will be x2.0. We hope that this will attract even more -adv users, and commensurately reward our current -adv users for this high-priority project.

Let me remind everyone that these "bonused" projects are for a limited time (until ~August). As before, point info at: http://fah-web.stanford.edu/cgi-bin/fahproject?p=4534

Vince
sortofageek
Site Admin
Posts: 3110
Joined: Fri Nov 30, 2007 8:06 pm
Location: Team Helix
Contact:

Re: New Bonus Amber Core WU's [-advmethods option]

Post by sortofageek »

Nice. :)
sneakers55
Posts: 94
Joined: Sun Dec 02, 2007 10:41 pm
Location: Texas, USA

Re: New Bonus Amber Core WU's [-advmethods option]

Post by sneakers55 »

vvoelz wrote:The -advmethods "bonused" AMBER Core WUs just got more bonus.

Starting with projects 4534-6 (to come online shortly), the bonus points will be x2.0. We hope that this will attract even more -adv users, and commensurately reward our current -adv users for this high-priority project.

Let me remind everyone that these "bonused" projects are for a limited time (until ~August).
I'm resisting the SMP siren song (although I'm thinking about replacing the slowest machine with a C2Q as my biggest expense has to be electricity). Recently had a small ($50K) windfall. Bank most of it but I could spring for more folding hardware!

My folding window garden has six cores all doing AMBER.
AMD Athlon X2 Dual Core 4200+ (2.2 GHz)
Intel C2D 6400 (2.13 GHz)
Intel C2D T7800 (2.6 GHz)
PS3
Leoslocks
Posts: 120
Joined: Fri Jan 25, 2008 3:20 am
Hardware configuration: Q6600 | P35-DQ6 | Crucial 2 x 1 GB ram | VisionTek 3870
GPU2 Version 6.20| CPU three 6.20 Clients

Re: New Bonus Amber Core WU's [-advmethods option]

Post by Leoslocks »

Got one today. Hope to see all my cores on them.

Code: Select all

[18:00:26] Project: 4535 (Run 18, Clone 15, Gen 0)
[18:00:26] 
[18:00:26] Assembly optimizations on if available.
[18:00:26] Entering M.D.
[18:00:32] Protein: p4535_T0396-16-5ns_minout
vvoelz
Pande Group Member
Posts: 552
Joined: Sun Dec 02, 2007 8:07 pm
Location: Temple University, Philadelphia PA

Re: New Bonus Amber Core WU's [-advmethods option]

Post by vvoelz »

the animal wrote:Are the Amber WU also benchmarked for 110ppd ( x bonus)?

I have a P4 1.7 that is getting just over 70ppd on 4533 (should be about 100 by my calculations) and 100 on 4534 (should be 118), and a PD 3.14 getting 125 on 4533 (should be 185).

Those numbers seem quite low for bonus WU to me.
Hmmm -- I'll do some double-checking on this and get back to you.

In the meantime it might be worth checking which projects have the old bonus (x1.5) and which projects have the new bonus (x2.0). Proejcts 4531, 4532, and 4533 have x1.5 bonus, while projects 4534 onward have the x2.0 bonus.

Again, see for example (scroll to bottom):
http://fah-web.stanford.edu/cgi-bin/fahproject?p=4533
and
http://fah-web.stanford.edu/cgi-bin/fahproject?p=4534

Further adding to general confusion may be the fact that 4534-4536 are half the length (5ns) of 4531-4533 (10 ns).

Vince
toTOW
Site Moderator
Posts: 6349
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France
Contact:

Re: New Bonus Amber Core WU's [-advmethods option]

Post by toTOW »

I read somewhere on the forums that these projects are subject to variation in folding times within a single project ... I don't know if remember well, but it might be the case here ;)
Image

Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.
the animal
Posts: 41
Joined: Mon Feb 25, 2008 5:04 pm
Location: TeAm Anandtech

Re: New Bonus Amber Core WU's [-advmethods option]

Post by the animal »

Yes, I read about the variation, but it seems pretty extreme.

Take the ppd of my PD for example. If the average is 185ppd, and I'm getting 125, then presumably there would also be a WU that would average 245 to balance it out. That would make almost a 100% difference, which I wouldn't really call "variation".
the animal
Posts: 41
Joined: Mon Feb 25, 2008 5:04 pm
Location: TeAm Anandtech

Re: New Bonus Amber Core WU's [-advmethods option]

Post by the animal »

I made a mistake in the post quoted above. 4534 on my P4 should average 133 not 118 (I had calculated at 1.5 not 2).
toTOW
Site Moderator
Posts: 6349
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France
Contact:

Re: New Bonus Amber Core WU's [-advmethods option]

Post by toTOW »

My P4-m 1.4 GHz is at 77 PPD on a 4533 ... it was at 83 PPD on the previous WU (p4532).
Image

Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.
vvoelz
Pande Group Member
Posts: 552
Joined: Sun Dec 02, 2007 8:07 pm
Location: Temple University, Philadelphia PA

Re: New Bonus Amber Core WU's [-advmethods option]

Post by vvoelz »

the animal wrote:Yes, I read about the variation, but it seems pretty extreme.

Take the ppd of my PD for example. If the average is 185ppd, and I'm getting 125, then presumably there would also be a WU that would average 245 to balance it out. That would make almost a 100% difference, which I wouldn't really call "variation".
There are two things here that I want to make clear:

1) There is *lots* of variation (if you'll allow me to call it that). I have confirmed that this is the culprit for multiple WUs now that have been suspected of under-performing with respect to the benchmark. (Strangely, I have received *no* reports about WUs suspected of over-performing. ;))

This variation is unfortunate, but there really is no way around it. (Like I have said before, we would run out of project numbers). If you run enough of these, your PPD will average out, but there are lots of fluctuations along the way.

Here's the cumulative distribution of the number of atoms for project p4534:
http://www.stanford.edu/~vvoelz/p4534_natoms.tiff
You can see there is quite a spread. For projects on the larger end of the spectrum, things do tend to slow down, as the non-bonded interactions scale as O(N^2).

2) The PPD will vary as the project progresses, more so than usual work units. We think this is due in part to more non-bonded interactions as time progress, exacerbated by our (very non-dense) initial starting states. You may see projects for which the bonus seems to dwindle. This is normal. Please keep in mind that there is a x2.0 bonus is in part designed to compensate for these idiosyncrasies.

Thanks again for donating your time and resources to helping us fold these targets, and for your patience with us trying some new things on FAH.

Vince
Pette Broad
Posts: 128
Joined: Mon Dec 03, 2007 9:38 pm
Hardware configuration: CPU folding on only one machine a laptop

GPU Hardware..
3 x 460
1 X 260
4 X 250

+ 1 X 9800GT (3 days a week)
Location: Chester U.K

Re: New Bonus Amber Core WU's [-advmethods option]

Post by Pette Broad »

Strangely, I have received *no* reports about WUs suspected of over-performing.
i've seen close to 400ppd on an AMD 6400+ running Windows but only the once. In General I'm seeing 200-280ppd on my 6000/6400+ machines and around 150-200 on a couple of 4600's. My Intel Q6600 is getting less than that, around 130-160 (per core) and my New Laptop a T7250 gives me 120-140. My older laptop a T5600, 65-75. Very stable units, done over 100 without any sign of an EUE.

Pete
Image
the animal
Posts: 41
Joined: Mon Feb 25, 2008 5:04 pm
Location: TeAm Anandtech

Re: New Bonus Amber Core WU's [-advmethods option]

Post by the animal »

vvoelz wrote: There are two things here that I want to make clear:

1) There is *lots* of variation (if you'll allow me to call it that). I have confirmed that this is the culprit for multiple WUs now that have been suspected of under-performing with respect to the benchmark. (Strangely, I have received *no* reports about WUs suspected of over-performing. ;))

This variation is unfortunate, but there really is no way around it. (Like I have said before, we would run out of project numbers). If you run enough of these, your PPD will average out, but there are lots of fluctuations along the way.

Here's the cumulative distribution of the number of atoms for project p4534:
http://www.stanford.edu/~vvoelz/p4534_natoms.tiff
You can see there is quite a spread. For projects on the larger end of the spectrum, things do tend to slow down, as the non-bonded interactions scale as O(N^2).

2) The PPD will vary as the project progresses, more so than usual work units. We think this is due in part to more non-bonded interactions as time progress, exacerbated by our (very non-dense) initial starting states. You may see projects for which the bonus seems to dwindle. This is normal. Please keep in mind that there is a x2.0 bonus is in part designed to compensate for these idiosyncrasies.

Thanks again for donating your time and resources to helping us fold these targets, and for your patience with us trying some new things on FAH.

Vince
Thanks for your response.

<Seinfeld>
I told you not to sell.
You did not tell me not to sell.
I said the market fluctuates. Remember?
Look, Vanessa, of course the market fluctuates. Everybody knows that. I just got fluctuated out of four thousand dollars!
</Seinfeld>
Post Reply