New SMP Core's

[16:23:51] + Downloading new core: FahCore_a5.exe
[16:23:51] Downloading core (/~pande/Linux/AMD64/Core_a5.fah from www.stanford.edu)
[16:23:51] - Error: HTTP GET returned error code 404
[16:23:51] + Error: Could not download core
[16:23:51] + Core download error (#4), waiting before retry..

Bah
 
*sigh* Back to A3...

It is said, “To err is human,”
That quote from alt.times.lore,
Alas, you have made an error,
So I say, “404.”

Double-check your URL,
As we all have heard before.
You ask for an invalid filename,
And I respond, “404.”

Perhaps you made a typo—
Your fingers may be sore—
But until you type it right,
You’ll only get 404.

Maybe you followed a bad link,
Surfing a foreign shore;
You’ll just have to tell that author
About this 404.

I’m just a lowly server
(Who likes to speak in metaphor),
So for a request that I don’t know,
I must return 404.

Be glad I’m not an old mainframe
That might just dump its core,
Because then you’d get a ten-meg file
Instead of this 404.

I really would like to help you,
But I don’t know what you’re looking for,
And since I don’t know what you want,
I give you 404.

Remember Poe, insane with longing
For his tragically lost Lenore.
Instead, you quest for files.
Quoth the Raven, “404!”
 
It downloaded a P6901 and now the core is missing to run it - still 404'ing. Sigh.
 
Is it legal / possible to post the a5 core so those of us who get the 404 can dl it and run the wu's we have received?
 
Nice I love the poem.

I'll be sticking with A3 for the forseeable future.


 
nwrtarget said:
Is it legal / possible to post the a5 core so those of us who get the 404 can dl it and run the wu's we have received?
Click to expand...

I highly doubt it...
but you can dl them yourself lol...
 
The A5 on the beta page didn't want to execute
Is there anything I have to do to get it to execute once I manually download it?
 
6.23 can download a5 but cant run it... do you have the 6.24 client?
 
It is smaller than the A3 core and we know code almost never gets smaller so I would assume it is zipped/packed/compressed.
 
By drop-in binary he means the new version of the client executable file, not the core.
 
In the client.cfg file I see one of the last lines is core so I put the path to the beta folder in there and am giving that a try.
 
nwrtarget said:
Yep that I got and it is working fine.
Click to expand...

sweet.... I kept trying to get to the forum page with a link but it kept locking up my browser...
glad its working... let us know how good :)
 
I digged a bit around and got a nativ a5 core running on linux now.

Not sure if TPD/PPD are correct. Lets see :D

A5Core.png
 
TheMTZ said:
I digged a bit around and got a nativ a5 core running on linux now.

Not sure if TPD/PPD are correct. Lets see :D

A5Core.png
Click to expand...

pretty...falls between my 2.26 quads and 2.8s on tpf
 
TheMTZ how did you get the a5 core up and running? Did it download automatically? Did you download it manually and then do something to it?
 
nwrtarget said:
TheMTZ how did you get the a5 core up and running? Did it download automatically? Did you download it manually and then do something to it?
Click to expand...

Here is what i did to get it running:

- >Download http://www.stanford.edu/~pande/Linux/AMD64/beta/Core_a5.fah
-> Manually converting it to a BZ File by:
-> tail -c +513 Core_a5.fah > temp.exe.bz
-> Unpack the BZ2 archive : bzip2 -d temp.exe.bz
-> rename it to FahCore_a5.exe

and run the client again. Its on 1% now and still having about the same TPF/PPD as in the screenshot earlier.
 
TheMTZ
Nice work. I found someone who did it and posted the end result. Thanks for the instructions for next time!
 
On the DL580 G7 running RHEL 5.6 4 x 8 at 2.0 GHz with HT
First frame 8 minutes


[20:13:19] - Digital signature verified
[20:13:19]
[20:13:19] Project: 6901 (Run 5, Clone 4, Gen 0)
[20:13:19]
[20:13:19] Assembly optimizations on if available.
[20:13:19] Entering M.D.
:) G R O M A C S (-:

Groningen Machine for Chemical Simulation

:) VERSION 4.5.3 (-:

Written by Emile Apol, Rossen Apostolov, Herman J.C. Berendsen,
Aldert van Buuren, Pär Bjelkmar, Rudi van Drunen, Anton Feenstra,
Gerrit Groenhof, Peter Kasson, Per Larsson, Pieter Meulenhoff,
Teemu Murtola, Szilard Pall, Sander Pronk, Roland Schulz,
Michael Shirts, Alfons Sijbers, Peter Tieleman,

Berk Hess, David van der Spoel, and Erik Lindahl.

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2010, The GROMACS development team at
Uppsala University & The Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.


:) Gromacs (-:

Reading file work/wudata_07.tpr, VERSION 4.0.99_development_20090605 (single precision)
[20:13:27] Mapping NT from 64 to 64
Note: tpx file_version 70, software version 73
Starting 64 threads
Making 2D domain decomposition 8 x 8 x 1
starting mdrun 'SINGLE VESICLE in water'
250000 steps, 1000.0 ps.
[20:13:32] Completed 0 out of 250000 steps (0%)

NOTE: Turning on dynamic load balancing

[20:21:48] Completed 2500 out of 250000 steps (1%)
 
Update to the above



[20:13:27] Mapping NT from 64 to 64
Note: tpx file_version 70, software version 73
Starting 64 threads
Making 2D domain decomposition 8 x 8 x 1
starting mdrun 'SINGLE VESICLE in water'
250000 steps, 1000.0 ps.
[20:13:32] Completed 0 out of 250000 steps (0%)

NOTE: Turning on dynamic load balancing

[20:21:48] Completed 2500 out of 250000 steps (1%)
[20:30:04] Completed 5000 out of 250000 steps (2%)
[20:37:59] Completed 7500 out of 250000 steps (3%)
[20:46:07] Completed 10000 out of 250000 steps (4%)
[20:53:51] Completed 12500 out of 250000 steps (5%)
[21:01:51] Completed 15000 out of 250000 steps (6%)
[21:09:41] Completed 17500 out of 250000 steps (7%)
[21:17:40] Completed 20000 out of 250000 steps (8%)
[21:25:30] Completed 22500 out of 250000 steps (9%)
[21:33:35] Completed 25000 out of 250000 steps (10%)
 
A little bit slower here:

[19:09:08] Completed 0 out of 250000 steps (0%)
[19:35:32] Completed 2500 out of 250000 steps (1%)
[20:02:07] Completed 5000 out of 250000 steps (2%)
[20:28:34] Completed 7500 out of 250000 steps (3%)
[20:55:03] Completed 10000 out of 250000 steps (4%)
[21:21:30] Completed 12500 out of 250000 steps (5%)
Click to expand...
 
mplee said:
Using the calculator for a 6900, that equates to somewhere around 250k ppd. That'll be a nice boost while you have the machine.
Click to expand...

That is real nice considering it was cranking sub 100K on SMP.

Maybe I will fire up the DL580G5 again....
 
I will have to swap some units on tuesday... no way I am leaving a new core alone with my rack from now till then...

but a test run between tuesday and thursday...ok...
 
Warning - new core fails with Langouste - none of the 3 units that finished a few hours ago uploaded.... checking...:mad:
 
The DL580 G5 is up and running. This one has 4x4 3.0GHz and 40 gigs of ram.

First frame 19 minutes. It isn't a G7 thats for sure!!!


[22:41:54] - Machine ID: 1
[22:41:54]
[22:41:54] Loaded queue successfully.
[22:41:54]
[22:41:54] + Processing work unit
[22:41:54] Core required: FahCore_a5.exe
[22:41:54] Core found.
[22:41:54] Working on queue slot 02 [February 24 22:41:54 UTC]
[22:41:54] + Working ...
[22:41:54] - Calling './FahCore_a5.exe -dir work/ -nice 19 -suffix 02 -np 16 -checkpoint 15 -forceasm -verbose -lifeline 6426 -version 634'

[22:41:54] - Autosending finished units... [February 24 22:41:54 UTC]
[22:41:54] Trying to send all finished work units
[22:41:54] + No unsent completed units remaining.
[22:41:54] - Autosend completed
[22:41:54]
[22:41:54] *------------------------------*
[22:41:54] Folding@Home Gromacs SMP Core
[22:41:54] Version 2.27 (Thu Feb 10 09:46:40 PST 2011)
[22:41:54]
[22:41:54] Preparing to commence simulation
[22:41:54] - Assembly optimizations manually forced on.
[22:41:54] - Not checking prior termination.
[22:41:56] - Expanded 20064892 -> 30796292 (decompressed 153.4 percent)
[22:41:56] Called DecompressByteArray: compressed_data_size=20064892 data_size=30796292, decompressed_data_size=30796292 diff=0
[22:41:56] - Digital signature verified
[22:41:56]
[22:41:56] Project: 6901 (Run 5, Clone 4, Gen 0)
[22:41:56]
[22:41:56] Assembly optimizations on if available.
[22:41:56] Entering M.D.
:) G R O M A C S (-:

Groningen Machine for Chemical Simulation

:) VERSION 4.5.3 (-:

Written by Emile Apol, Rossen Apostolov, Herman J.C. Berendsen,
Aldert van Buuren, Pär Bjelkmar, Rudi van Drunen, Anton Feenstra,
Gerrit Groenhof, Peter Kasson, Per Larsson, Pieter Meulenhoff,
Teemu Murtola, Szilard Pall, Sander Pronk, Roland Schulz,
Michael Shirts, Alfons Sijbers, Peter Tieleman,

Berk Hess, David van der Spoel, and Erik Lindahl.

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2010, The GROMACS development team at
Uppsala University & The Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.


:) Gromacs (-:

Reading file work/wudata_02.tpr, VERSION 4.0.99_development_20090605 (single precision)
[22:42:03] Mapping NT from 16 to 16
Note: tpx file_version 70, software version 73
Starting 16 threads
Making 2D domain decomposition 4 x 4 x 1
starting mdrun 'SINGLE VESICLE in water'
250000 steps, 1000.0 ps.
[22:42:07] Completed 0 out of 250000 steps (0%)
[23:01:22] Completed 2500 out of 250000 steps (1%)
 
eeek - HELP!

NOTHING has appeared on stats - 3 units lost.:confused:

Hard to tell, lots of problems here. It is Langouste, or the work server, or a local bug in corrupting file data?

Units finished, langouste forks them, and moves on to next. Uploads have errors. I go back and try to redo uploads with the FAH.exe -local -send all, (this works when a unit has failed to send after a -oneunit shutdown) but this fails too.



Code: 
[23:36:12] + Attempting to send results [February 24 23:36:12 UTC]
[23:51:15] - Couldn't send HTTP request to server
[23:51:15] + Could not connect to Work Server (results)
[23:51:15]     (130.237.232.237:8080)
[23:51:15] + Retrying using alternative port
[00:06:16] - Couldn't send HTTP request to server
[00:06:16] + Could not connect to Work Server (results)
[00:06:16]     (FAH.exe -local -send all:80)
[00:06:16] - Error: Could not transmit unit 04 (completed February 24) to work server.
[00:06:16] - Read packet limit of 540015616... Set to 524286976.


[00:06:16] + Attempting to send results [February 25 00:06:16 UTC]
[00:06:17] - Couldn't send HTTP request to server
[00:06:17] + Could not connect to Work Server (results)
[00:06:17]     (130.237.165.141:8080)
[00:06:17] + Retrying using alternative port
[00:06:18] - Couldn't send HTTP request to server
[00:06:18] + Could not connect to Work Server (results)
[00:06:18]     (130.237.165.141:80)
[00:06:18]   Could not transmit unit 04 to Collection server; keeping in queue.
[00:06:18] - Failed to send all units to server

Folding@Home Client Shutdown.

On one machine trying again just results in fast exits of FAH - the log shows that it can't read some of the files.

Code: 
[23:14:34] Loaded queue successfully.
[23:14:34] Attempting to return result(s) to server...
[23:14:34] Project: 6900 (Run 17, Clone 10, Gen 25)
[23:14:34] - Read packet limit of 540015616... Set to 524286976.
[23:14:34] - Error: Could not get length of results file work/wuresults_05.dat
[23:14:34] - Error: Could not read unit 05 file. Removing from queue.
[23:14:34] - Failed to send all units to server

Folding@Home Client Shutdown.
 
I am seeing the same thing - using -send all with langouste hangs for about 15 minutes, then shows a failed to connect message. I am trying to manually send now. I am afraid this has nothing to do with langouste, but rather that the results server can't receive completed 6901 units...

Edit: Manual uploading apears to have worked for me. I'll shut off langouste on my other machines that are running the new client/core and see if there are still issues uploading. I'll have another update in ~8 hours.
 
Last edited:
I have manually sent several times - and it seems to complete - but the bytes sent is too low, and nothing shows up on stats, and the temp folder is still there...

posted to FF.
 
Patriot said:
try without using langouste
Click to expand...

Yeah, we think alike. Have been trying that, with same result.

EDIT: it is a server problem. Does not look like they are ready for prime time. Or maybe they were not expecting new WU back 17 hours after the client was released. ;)

Apollo - why did I not listen to you? :rolleyes:

Oh well, 375,000 points up in smoke, less what peanut credit they give me when the server is back up.
 
Last edited:
You must log in or register to reply here.
Back
Top