Folding Forum

Couple of days ago I've been assigned a 4437 (82-1-3) unit.
I was taken aback a little when I could not find it in any list, but I surmised it's of the same family as 4438 etc., for 225 points. I was mystified even more when it I couldn't find its server (171.67.108.20) within official WU's list with not a single Wu (4438 family is handed out by 171.67.108.13), but what could I do

Anyway, twenty minutes ago the day came and I hoped to upload it and thus know some more, but alas... server rebuffed me with a message I didn't ever see so far, namely

[02:46:03] + Attempting to send results
[02:46:03] - Presenting message box asking to network.
[02:47:01] - Server reports problem with unit.

And that's it - no further trying, no points, no nothing - it just went on to download yet another WU.
As I could not find such message in wiki or other posts on this forum, I had to ask what this is.
Could anyone enlighten me a bit, please?

Psummary is NOT the official list of work units. It is the list of work units that the servers are currently sending out. The list is updated dynamically. If the servers run out of P4437, it is no longer listed.

http://fah-web.stanford.edu/cgi-bin/fahproject?p=4437

The "[02:46:03] - Presenting message box asking to network." message is logged when you have the client set to prompt for an internect connection. It logs the fact that it did ask if you wanted to connect.

Not really sure what happened after that, as you didn't post any more of the log where it went on to do other things. It also helps if you post the first part of the log showing the client startup so we can see which client you are using, the version, what switches, etc.

7im wrote:Psummary is NOT the official list of work units. It is the list of work
units that the servers are currently sending out.

That's what I meant anyway - since I was looking for details of this WU (points and servers), this is my usual reference.
The only list of past WU's that I know then is the 'full' list given while uploading results at fahinfo.org, and it's funny too - we have a full range of Amber_Natives there, listed from 4433 to 4463, yet with notable exception of 4436 and and 4437.
Truly a mysterious one

7im wrote:Not really sure what happened after that, as you didn't post any more of
the log where it went on to do other things.

I'm afraid that's all I have under current verbosity, there is really nothing relevant there:

[02:46:00] - Ask before connecting: Yes
[02:46:00] - User name: endrik (Team 276)
[02:46:00] - User ID: 23078AC450A98CB8
[02:46:00] - Machine ID: 2
[02:46:00]
[02:46:00] Loaded queue successfully.
[02:46:00] + Benchmarking ...
[02:46:02] - Preparing to get new work unit...
[02:46:02] - Presenting message box asking to network.

[02:46:03] + Attempting to send results
[02:46:03] - Presenting message box asking to network.
[02:47:01] - Server reports problem with unit.

Folding@Home Client Shutdown.

I have everything set on manual, since I run just CPU, there are no industrial quantities and I like to keep control of things. So after that last message I just switched it off without downloading another one (so that not another week will be wasted) and went on searching what exactly does such message mean (still to no avail ...

[02:47:01] - Server reports problem with unit.

There is no clear definition of that error. There can be many problems, and all are reported in that same unhelpful manner.

4436 and 4437 have yet to make it onto any of the Psummary lists. I've been getting these for maybe a month (have 3 running right now) and they are always shown as project 0 with 0 points. This has been mentioned several times in other posts, but either there's some sort of problem or Pande Group have forgotten about it.


Pete

Pette Broad wrote:either there's some sort of problem or Pande Group have forgotten about it.

That roughly amounts to the same effect
As another mod toTOW wrote there is a 4436 unit that was assigned and completed on 18 different computers - now that's a waste, my one week looks bleak in comparison.
Maybe Pande group is neglecting CPU users in favour of GPUs - I wouldn't be surprised given the difference in computing power. Perhaps nothing intentional, it's just that they have priority in solving problems (as it was reported and acknowledged a week ago).

Just to have clear conscience I finished the WU again from a backup file and sent it with high verbosity, to expected result:

[00:51:46] + Attempting to send results
[00:51:46] - Reading file work/wuresults_05.dat from core
[00:51:46] (Read 2352535 bytes from disk)
[00:51:46] - Presenting message box asking to network.
[00:51:59] Connecting to http://171.67.108.20:8080/
[00:52:54] Posted data.
[00:52:54] Initial: 0000; - Uploaded at ~41 kB/s
[00:52:54] - Averaged speed for that direction ~79 kB/s
[00:52:54] - Server reports problem with unit.
[00:52:54] + Sent 0 of 1 completed units to the server
[00:52:54] - Autosend completed

Case is closed for me, I'll change client config somewhat to avoid this series.

Yeah, these are very slow units indeed, mine on reasonable machines are only around the 35-40% mark after 24 plus hours. I'll be switching off my remaining CPU only folders as they finish up their units and just continue with GPU only, just not worth the effort and cost for such a feeble return. Won't make that much difference as I'm only getting 10% of my points from CPU's anyway.

Pete

I noticed I've gotten a few of these new 44xx WUs recently on my lone CPU client. I also have noticed that FAHMON 2.3.2 reports WU credit as unknown for several if not all of these... So I thought I'd search the forum to see if someone had determined what credit ws granted when the WU was returned. Still not sure... Currently working on p4337 (R5,C3,G6). Are p4336/p4437 225 point WUs s/a p4438 ??

No big deal to me, just curious. I'll continue to let it crunch away regardless of how feeble the ppd. This one is on an approx 2.3 day pace. Running on 2nd core of a E2160 @default speed. (The 1st core feeding a GPU). With the system GPU folding, it might burn 180W. With system GPU folding and CPU folding it might then burn 185W. I think I can live with the extra power usage...

FAHMON Benchmark Data:
Project : 4436 (Unknown)
-- E2160 - CPU Console --
Min. Time / Frame : 32mn 56s
Avg. Time / Frame : 32mn 56s
No Cur. Time / Frame
No R3F. Time / Frame
No Eff. Time / Frame

Project : 4437 (Unknown)
-- E2160 - CPU Console --
Min. Time / Frame : 32mn 56s
Avg. Time / Frame : 32mn 56s
Cur. Time / Frame : 33mn 08s
R3F. Time / Frame : 33mn 00s
Eff. Time / Frame : 6h 00mn 08s

Some quick maths tells me these units are taking twice as long as other the 44xx units that I'm running right now. Yes, I'm still actually running 2 of the 3 units from 12th January that's how slow and undervalued they are. Perhaps the person in charge of these units should have a look at the benchmark figures again. I'm presuming here that they are worth 225 points....no way of knowing for sure.

Pete

I recently finished 4436 and in fact it's worth 225 points. I'm also folding the 44xx units for the last time on one of my two CPU-Clients. 4436 took about 24 minutes per frame, the projects >4437 about 13 minutes - for the same credit... And the projects don't appear in the PSummary!

I also am working on WU 4437 (R3,C3,G7), at this time I have completed 37%.
I can not find any info on this WU,
as (FahMon, FahLog, and http://fah-web.stanford.edu/psummary.html) have not been updated to include this WU.

Can someone find an Answer?
Should we just delete the WU and start a new (if it not being accepted back by the servers) why bother folding it, if it is not going to be used?

I'd let them finish then change the settings so you don't get them again until Pande Group sort it out. Unfortunately, I guess nobody's bothered, if it was GPU2 it'd be sorted right away.

Pete

I just want to know, if everything is correct with the projects 4436 and 4437! Why don't they appear? Are they supposed to be that slow (two times slower than the other 44xx!)? A single statement like "Everything's OK" or "Under observation" would be enough for me...

If everything's fine, than of course I would fold them - science matters!

It gets worse.....Nobody from the Pande Group appears to "own" the server these units are coming from at least that's the way it looks to me from looking at the server status. Also, the server status is showing 10,000 units available all the time so something is not right. It would seem that all Windows CPU's are now being sent to this server unless you specify -advmethods and Big WU's....not good. I'm now deleting any 4436/4437's that my machines picks up, I'm not folding for half points on units that nobody seems bothered about fixing.


Pete

Pete, It appears the server stats counts do not get updated for this server. Its been stuck at 10000/ 10000 since Jan-11.

Got a p4437 myself and here are a few details from qd:

Index 3: folding now 225.00 pts (3.350 pt/hr) 22.5 X min speed; 4% complet
server: 171.67.108.20:8080; project: 4437, "p4437"
Folding: run 92, clone 1, generation 8; benchmark 0; misc: 500, 0
issue: Wed Dec 31 18:00:00 1969; begin: Mon Jan 19 09:08:08 2009
expect: Thu Jan 22 04:18:08 2009; due: Mon Mar 23 10:08:08 2009 (63 days)
preferred: Tue Mar 03 09:08:08 2009 (43 days)
core URL: http://www.stanford.edu/~pande/Win32/x86/Core_78.fah (V1.90)
CPU: 1,687 Pentium II/III; OS: 1,8 WinXP
flops: 1064887178 (1064.887178 megaflops)
memory: 900 MB
assignment info (be): Mon Jan 19 09:09:21 2009; BA5D7FCA
P limit: 10484736
user: AnandBhat; team: 41494; ID: xxx; mach ID: xxx
work/wudata_03.dat file size: 238595; WU type: Folding@Home