much points - short pref. time

[JeansOn]

Hallo,
I've got a large WU Project: 10722 (Run 0, Clone 4731, Gen 0) with 2244 points.

one percent in 1,5 hours - it is too hard for me, because it is too much to reach the preferred deadline.
My cpu-cores are working @2,93GHz (It's a two year old dualcore-pentium)

I think, I read somewhere the final deadline is for the points only and results are not taken by the project.
When I send back my results in the preferred deadline the project takes the results.
I support the pande-project up to 6 hours a day, on weekend often 12 hours, it is not a problem. But only when I'm at home.
I want to support the project, I'm too old to add some points. That would not be enough.

150 hours / 30 days (pref. time) - that is too short to reach the preferred time calculating the PC-up-time.

Now my problem:
I deleted the WU (I gave it to my son. It works well, there.)
Then I deleted queue.dat, unitinfo.txt and, at last the whole work dir.

Well, I started the client again, and I'm astonished: The same Project was downloaded again.
What can I do? I made an other client.cfg "bigpackets=small" but the next download-result was the same again.
How can I get an other WU?

I do not think, my machine is the oldest/slowest one. What does the others do?

EDIT:
I add some log-code, if helpful.
At my son's Quad-core he can get 1% in 1:20 h.

Code: Select all

[16:19:13] + Attempting to send results [June 19 16:19:13 UTC]
[16:19:13] - Reading file work/wuresults_03.dat from core
[16:19:13]   (Read 1897741 bytes from disk)
[16:19:13] Connecting to http://171.67.108.42:8080/
[16:19:35] Posted data.
[16:19:35] Initial: 0000; - Uploaded at ~84 kB/s
[16:19:35] - Averaged speed for that direction ~76 kB/s
[16:19:35] + Results successfully sent
[16:19:35] Thank you for your contribution to Folding@Home.
[16:19:35] + Number of Units Completed: 223

[16:19:39] Trying to send all finished work units
[16:19:39] + No unsent completed units remaining.
[16:19:39] - Preparing to get new work unit...
[16:19:39] + Attempting to get work packet
[16:19:39] - Will indicate memory of 6142 MB
[16:19:39] - Detect CPU. Vendor: GenuineIntel, Family: 6, Model: 7, Stepping: 6
[16:19:39] - Connecting to assignment server
[16:19:39] Connecting to http://assign.stanford.edu:8080/
[16:19:40] Posted data.
[16:19:40] Initial: 43AB; - Successful: assigned to (171.67.108.42).
[16:19:40] + News From Folding@Home: Welcome to Folding@Home
[16:19:40] Loaded queue successfully.
[16:19:40] Connecting to http://171.67.108.42:8080/
[16:19:42] Posted data.
[16:19:42] Initial: 0000; - Receiving payload (expected size: 228361)
[16:19:42] Conversation time very short, giving reduced weight in bandwidth avg
[16:19:42] - Downloaded at ~446 kB/s
[16:19:42] - Averaged speed for that direction ~827 kB/s
[16:19:42] + Received work.
[16:19:42] Trying to send all finished work units
[16:19:42] + No unsent completed units remaining.
[16:19:42] + Closed connections
[16:19:42]
[16:19:42] + Processing work unit
[16:19:42] Core required: FahCore_a4.exe
[16:19:42] Core found.
[16:19:42] Working on queue slot 04 [June 19 16:19:42 UTC]
[16:19:42] + Working ...
[16:19:42] - Calling '.\FahCore_a4.exe -dir work/ -suffix 04 -checkpoint 3 -verbose -lifeline 3032 -version 623'

[16:19:43]
[16:19:43] *------------------------------*
[16:19:43] Folding@Home Gromacs GB Core
[16:19:43] Version 2.27 (Dec. 15, 2010)
[16:19:43]
[16:19:43] Preparing to commence simulation
[16:19:43] - Looking at optimizations...
[16:19:43] - Created dyn
[16:19:43] - Files status OK
[16:19:43] - Expanded 227849 -> 354128 (decompressed 155.4 percent)
[16:19:43] Called DecompressByteArray: compressed_data_size=227849 data_size=354128, decompressed_data_size=354128 diff=0
[16:19:43] - Digital signature verified
[16:19:43]
[16:19:43] Project: 10722 (Run 0, Clone 4731, Gen 0)
[16:19:43]
[16:19:43] Assembly optimizations on if available.
[16:19:43] Entering M.D.
[16:19:49] Mapping NT from 1 to 1
[16:19:49] Completed 0 out of 7000000 steps  (0%)
[17:49:15] Completed 70000 out of 7000000 steps  (1%)
[19:21:18] Completed 140000 out of 7000000 steps  (2%)
[19:52:47] ***** Got a SIGTERM signal (2)
[19:52:47] Killing all core threads

Folding@Home Client Shutdown.


--- Opening Log file [June 19 20:00:44 UTC]


# Windows CPU Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.23

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: D:\Programme\Folding@Home
Executable: D:\Programme\Folding@Home\Folding@home-Win32-x86.exe
Arguments: -local -verbosity 9

[20:00:44] - Ask before connecting: No
[20:00:44] - User name: JeansOn (Team 70335)
[20:00:44] - User ID: 2280A5F039EFA816
[20:00:44] - Machine ID: 1
[20:00:44]
[20:00:44] Work directory not found. Creating...
[20:00:44] Could not open work queue, generating new queue...
[20:00:44] - Preparing to get new work unit...
[20:00:44] + Attempting to get work packet
[20:00:44] - Will indicate memory of 6142 MB
[20:00:44] - Detect CPU. Vendor: GenuineIntel, Family: 6, Model: 7, Stepping: 6
[20:00:44] - Connecting to assignment server
[20:00:44] - Autosending finished units... [June 19 20:00:44 UTC]
[20:00:44] Trying to send all finished work units
[20:00:44] Connecting to http://assign.stanford.edu:8080/
[20:00:44] + No unsent completed units remaining.
[20:00:44] - Autosend completed
[20:00:45] Posted data.
[20:00:45] Initial: 43AB; - Successful: assigned to (171.67.108.42).
[20:00:45] + News From Folding@Home: Welcome to Folding@Home
[20:00:45] Loaded queue successfully.
[20:00:45] Connecting to http://171.67.108.42:8080/
[20:00:47] Posted data.
[20:00:47] Initial: 0000; - Receiving payload (expected size: 228361)
[20:00:47] Conversation time very short, giving reduced weight in bandwidth avg
[20:00:47] - Downloaded at ~446 kB/s
[20:00:47] - Averaged speed for that direction ~446 kB/s
[20:00:47] + Received work.
[20:00:47] + Closed connections
[20:00:47]
[20:00:47] + Processing work unit
[20:00:47] Core required: FahCore_a4.exe
[20:00:47] Core found.
[20:00:47] Working on queue slot 01 [June 19 20:00:47 UTC]
[20:00:47] + Working ...
[20:00:47] - Calling '.\FahCore_a4.exe -dir work/ -suffix 01 -checkpoint 3 -verbose -lifeline 3292 -version 623'

[20:00:47]
[20:00:47] *------------------------------*
[20:00:47] Folding@Home Gromacs GB Core
[20:00:47] Version 2.27 (Dec. 15, 2010)
[20:00:47]
[20:00:47] Preparing to commence simulation
[20:00:47] - Looking at optimizations...
[20:00:47] - Created dyn
[20:00:47] - Files status OK
[20:00:48] - Expanded 227849 -> 354128 (decompressed 155.4 percent)
[20:00:48] Called DecompressByteArray: compressed_data_size=227849 data_size=354128, decompressed_data_size=354128 diff=0
[20:00:48] - Digital signature verified
[20:00:48]
[20:00:48] Project: 10722 (Run 0, Clone 4731, Gen 0)
[20:00:48]
[20:00:48] Assembly optimizations on if available.
[20:00:48] Entering M.D.
[20:00:54] Mapping NT from 1 to 1
[20:00:54] Completed 0 out of 7000000 steps  (0%)
[20:01:20] ***** Got a SIGTERM signal (2)
[20:01:20] Killing all core threads

Folding@Home Client Shutdown.


--- Opening Log file [June 19 20:03:49 UTC]


# Windows CPU Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.23

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: D:\Programme\Folding@Home
Executable: D:\Programme\Folding@Home\Folding@home-Win32-x86.exe
Arguments: -local -verbosity 9

[20:03:49] - Ask before connecting: No
[20:03:49] - User name: JeansOn (Team 70335)
[20:03:49] - User ID: 2280A5F039EFA816
[20:03:49] - Machine ID: 1
[20:03:49]
[20:03:49] Work directory not found. Creating...
[20:03:49] Loaded queue successfully.
[20:03:49]
[20:03:49] + Processing work unit
[20:03:49] Core required: FahCore_a4.exe
[20:03:49] - Autosending finished units... [June 19 20:03:49 UTC]
[20:03:49] Trying to send all finished work units
[20:03:49] + No unsent completed units remaining.
[20:03:49] - Autosend completed
[20:03:49] Core found.
[20:03:49] Working on queue slot 01 [June 19 20:03:49 UTC]
[20:03:49] + Working ...
[20:03:49] - Calling '.\FahCore_a4.exe -dir work/ -suffix 01 -checkpoint 3 -verbose -lifeline 2236 -version 623'

[20:03:49]
[20:03:49] *------------------------------*
[20:03:49] Folding@Home Gromacs GB Core
[20:03:49] Version 2.27 (Dec. 15, 2010)
[20:03:49]
[20:03:49] Preparing to commence simulation
[20:03:49] - Looking at optimizations...
[20:03:49] - Created dyn
[20:03:49] - Files status OK
[20:03:49] Error: Missing work file=<>
[20:03:49]
[20:03:49] Folding@home Core Shutdown: MISSING_WORK_FILES
[20:03:54] CoreStatus = 74 (116)
[20:03:54] The core could not find the work files specified. Removing from queue
[20:03:54] Deleting current work unit & continuing...
[20:03:58] Trying to send all finished work units
[20:03:58] + No unsent completed units remaining.
[20:03:58] - Preparing to get new work unit...
[20:03:58] + Attempting to get work packet
[20:03:58] - Will indicate memory of 6142 MB
[20:03:58] - Detect CPU. Vendor: GenuineIntel, Family: 6, Model: 7, Stepping: 6
[20:03:58] - Connecting to assignment server
[20:03:58] Connecting to http://assign.stanford.edu:8080/
[20:03:58] Posted data.
[20:03:58] Initial: 43AB; - Successful: assigned to (171.67.108.42).
[20:03:58] + News From Folding@Home: Welcome to Folding@Home
[20:03:58] Loaded queue successfully.
[20:03:58] Connecting to http://171.67.108.42:8080/
[20:04:01] Posted data.
[20:04:01] Initial: 0000; - Receiving payload (expected size: 228361)
[20:04:01] Conversation time very short, giving reduced weight in bandwidth avg
[20:04:01] - Downloaded at ~446 kB/s
[20:04:01] - Averaged speed for that direction ~446 kB/s
[20:04:01] + Received work.
[20:04:01] + Closed connections
[20:04:06]
[20:04:06] + Processing work unit
[20:04:06] Core required: FahCore_a4.exe
[20:04:06] Core found.
[20:04:06] Working on queue slot 02 [June 19 20:04:06 UTC]
[20:04:06] + Working ...
[20:04:06] - Calling '.\FahCore_a4.exe -dir work/ -suffix 02 -checkpoint 3 -verbose -lifeline 2236 -version 623'

[20:04:06]
[20:04:06] *------------------------------*
[20:04:06] Folding@Home Gromacs GB Core
[20:04:06] Version 2.27 (Dec. 15, 2010)
[20:04:06]
[20:04:06] Preparing to commence simulation
[20:04:06] - Looking at optimizations...
[20:04:06] - Created dyn
[20:04:06] - Files status OK
[20:04:06] - Expanded 227849 -> 354128 (decompressed 155.4 percent)
[20:04:06] Called DecompressByteArray: compressed_data_size=227849 data_size=354128, decompressed_data_size=354128 diff=0
[20:04:06] - Digital signature verified
[20:04:06]
[20:04:06] Project: 10722 (Run 0, Clone 4731, Gen 0)
[20:04:06]
[20:04:06] Assembly optimizations on if available.
[20:04:06] Entering M.D.
[20:04:12] Mapping NT from 1 to 1
[20:04:12] Completed 0 out of 7000000 steps  (0%)
[20:05:45] ***** Got a SIGTERM signal (2)
[20:05:45] Killing all core threads

Folding@Home Client Shutdown.


--- Opening Log file [June 20 17:07:33 UTC]


# Windows CPU Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.23

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: D:\Programme\Folding@Home
Executable: D:\Programme\Folding@Home\Folding@home-Win32-x86.exe
Arguments: -local -verbosity 9

[17:07:33] - Ask before connecting: No
[17:07:33] - User name: JeansOn (Team 70335)
[17:07:33] - User ID: 2280A5F039EFA816
[17:07:33] - Machine ID: 1
[17:07:33]
[17:07:33] Work directory not found. Creating...
[17:07:34] Could not open work queue, generating new queue...
[17:07:34] - Preparing to get new work unit...
[17:07:34] + Attempting to get work packet
[17:07:34] - Will indicate memory of 6142 MB
[17:07:34] - Detect CPU.[17:07:34] - Autosending finished units... [June 20 17:07:34 UTC]
[17:07:34] Trying to send all finished work units
[17:07:34] + No unsent completed units remaining.
[17:07:34] - Autosend completed
 Vendor: GenuineIntel, Family: 6, Model: 7, Stepping: 6
[17:07:34] - Connecting to assignment server
[17:07:34] Connecting to http://assign.stanford.edu:8080/
[17:07:35] Posted data.
[17:07:35] Initial: 43AB; - Successful: assigned to (171.67.108.42).
[17:07:35] + News From Folding@Home: Welcome to Folding@Home
[17:07:35] Loaded queue successfully.
[17:07:35] Connecting to http://171.67.108.42:8080/
[17:07:37] Posted data.
[17:07:37] Initial: 0000; - Receiving payload (expected size: 228361)
[17:07:37] Conversation time very short, giving reduced weight in bandwidth avg
[17:07:37] - Downloaded at ~446 kB/s
[17:07:37] - Averaged speed for that direction ~446 kB/s
[17:07:37] + Received work.
[17:07:37] + Closed connections
[17:07:37]
[17:07:37] + Processing work unit
[17:07:37] Core required: FahCore_a4.exe
[17:07:37] Core found.
[17:07:37] Working on queue slot 01 [June 20 17:07:37 UTC]
[17:07:37] + Working ...
[17:07:37] - Calling '.\FahCore_a4.exe -dir work/ -suffix 01 -checkpoint 3 -verbose -lifeline 1060 -version 623'

[17:07:37]
[17:07:37] *------------------------------*
[17:07:37] Folding@Home Gromacs GB Core
[17:07:37] Version 2.27 (Dec. 15, 2010)
[17:07:37]
[17:07:37] Preparing to commence simulation
[17:07:37] - Looking at optimizations...
[17:07:37] - Created dyn
[17:07:37] - Files status OK
[17:07:38] - Expanded 227849 -> 354128 (decompressed 155.4 percent)
[17:07:38] Called DecompressByteArray: compressed_data_size=227849 data_size=354128, decompressed_data_size=354128 diff=0
[17:07:38] - Digital signature verified
[17:07:38]
[17:07:38] Project: 10722 (Run 0, Clone 4731, Gen 0)
[17:07:38]
[17:07:38] Assembly optimizations on if available.
[17:07:38] Entering M.D.
[17:07:43] Mapping NT from 1 to 1
[17:07:44] Completed 0 out of 7000000 steps  (0%)
[17:08:26] ***** Got a SIGTERM signal (2)
[17:08:26] Killing all core threads

Folding@Home Client Shutdown.



[P5-133XL]

Since it is a dual core CPU, are you running two uniprocessor clients? You might find that adding the -smp additional parameter to one of the clients and then shutting down the other uniprocessor client may help in decreasing the frame times (1% being done). The -smp parameter allows both cores to work on the same WU cutting the time in half.

Unfortunately, you are going to continue getting the specific WU. Once a WU is assigned to a specific client, the servers are setup to keep reassigning the same WU till it is returned. The goal was to keep people from cherry picking only the good projects by deleting WU's. Giving the project to someone else is not going to work because the servers will also check to see if the WU was assigned to your son's machine when it is returned and he won't get credit for it. Sorry!

Stanford has recently been giving these FAHCore_A4 projects to the uniprocessor clients. It can be a problem because they have relatively short deadlines (7 days) compared to the traditional uniprocessor project (30+ days).

Selecting the packet size to different values as well as choosing -advmethods or not may give you future immunity to these A4 projects for those settings select from a different set of projects. I don't know which setting works and which one doesn't for A4's so some experimentation is in order. Further the project groups for each of those settings change over time so what works now may not work in the future. That being said, those settings only work when getting a new WU and till the current WU is returned or expires you won't get a new WU.

[bruce]

I think, I read somewhere the final deadline is for the points only and results are not taken by the project.
When I send back my results in the preferred deadline the project takes the results.
I support the pande-project up to 6 hours a day, on weekend often 12 hours, it is not a problem. But only when I'm at home.
I want to support the project, I'm too old to add some points. That would not be enough.

After the preferred deadline, the results MAY not be taken by the project, but there's a good chance of that. For example:
Assume that Pref Deadline = 20.00; Final Dealine = 30.00; You start a WU and will complete it in 24 days.
WU is reissued to someone else after 20 days.

There are 3 possibilities.
A) The person receiving the reassignment completes it in 3 days and the project proceeds. Your result arrives on the 4th day and is an unnecessary duplicate.
B) The person receiving the reassignment completes it in 8 days. Your result arrives on the 4th day and the project proceeds. His unnecessary duplicate arrives on the 8th day.
C) The other person fails to complete the WU so your result arrives on the 4th day and the project proceeds.

The real cost to the project is that somebody was wasting their time and could have been doing something else in both cases A and B.

150 hours / 30 days (pref. time) - that is too short to reach the preferred time calculating the PC-up-time.
. . .
I do not think, my machine is the oldest/slowest one. What does the others do?

Understood. There are some internal discussions going on at the Pande Group about how to deal with this issue. You're not the first to report it.

Now my problem:
I deleted the WU (I gave it to my son. It works well, there.)
Then I deleted queue.dat, unitinfo.txt and, at last the whole work dir.

Well, I started the client again, and I'm astonished: The same Project was downloaded again.
What can I do? I made an other client.cfg "bigpackets=small" but the next download-result was the same again.
How can I get an other WU?

The server keeps track of every machine (client) and every WU that has been assigned to it. "Giving" the WU to your son is probably the wrong thing to do and it won't be resolved until the WU is actually returned. The server still knows that the WU is checked out to your machine and expects it to be returned. As far as the server can tell, it doesn't know your son has it; it just knows that you are supposed to have the WU and it has been dumped so it reassigns it.

I don't know what will happen to the uniprocessor deadlines or when. These new projects are a radical change for people like you who fold part-time or on very, very slow machines.

[Amaruk]

Project 10722, GRO-A4, 32 day preferred/56 day final deadline, 2244.00 points, k=0.75

Code: Select all

[16:19:49] Completed 0 out of 7000000 steps  (0%)
[17:49:15] Completed 70000 out of 7000000 steps  (1%)
[19:21:18] Completed 140000 out of 7000000 steps  (2%)

TPF 01:30:45 = total runtime of 151h, 15m. (6 days, 7 h, 15m)

Preferred deadline is 32 days. (768 hours) Average of 5.5 hours/day would meet preferred deadline.

Final deadline is 56 days. (1,344 hours.) Average of 2.75 hours/day would meet final deadline.


If it is possible to average 6 hours per day, you will easily make preferred deadline.

If this is not possible, the only way to get a different WU assignment is to change MID.
However, please note that there is nothing to prevent you from getting another 10722 if you do this.

[JeansOn]

Thank you, for your answers. They are helpful.

I found it out, why I got always the same WU after deleting it:
I changed the machine-ID in client.cfg and then I got an other WU.

I consider, I' ll take back the WU 99% finished at my son's pc and, after re-changing the machine-ID I'll finish the work with -oneunit Option.
You can be sure, it will work, because I shared a WU via usb-stick with my pc@work. ... until I thought the possibility, my employer doesn't want, I do that. ...

But in future I have to change machine-IDs?
I do not want to support the whole project with tricks. And I want to be sure that my results are taken without duplicates (by me or others).
That's, why I decrease my RAM in client.cfg from 6GB to 1GB. (I upgrated RAM three weeks ago and I never had such problems with large WUs the two years I 'm continously folding.)

[MtM]

Look out, people been fired for running Folding@home on their work computers without permission, the faq states you need it for good reasons.

Decreasing ram might help with assignments if there are wu's which fit the profile you're requesting but no one can guarantee there are at any given time. The sneakernet article on the folding wiki should cover all aspects on how to proceed if you want to sneakernet. If you want a definitive answer as to how you would ideally configure your system regarding deadlines and points is hard to say at this point, PG needs to decide what route to take with the clients and points at this time.

[Jan van de Velde]

• I started folding 5½ years ago
• I like the idea and support it wholeheartedly.
• I co-direct a (top-400) team (Wetenschapsforum) of 20-40 active Folding-enthousiasts.
• I have a reasonably up-to-date, standard machine.
• That machine has an up-time of 4-8 hours a day, most days.
• I do not wish / am not capable to start "messing" with software (and messed up software is wat I usually end up with as soon as I try to change something in the innards of my machine )

And now, for the first time in 5½ years I am missing preferred deadlines, e.g. on 7028 and 7600.

I hope that this pressure on "simple" Folders like myself can be relieved. Or, if this pressure is necessary, that Stanford simply but clearly states that they can no longer do something useful with machines like mine. No hard feelings, but it would save me the trouble to try and keep up, and would reduce my PC-power costs.

[VijayPande]

We had a meeting today to work on this issue. Our short term plan is to extend the deadlines for the A4 projects. We'll make an announcement once we settle on details.

[Jan van de Velde]

We had a meeting today to work on this issue. Our short term plan is to extend the deadlines for the A4 projects. We'll make an announcement once we settle on details.

Thank you. I hope that means there is no need in the foreseeable future for this standard owner's standard machine to start feeling useless.

[VijayPande]

Yes, we're trying to find ways to be as inclusive as possible, but still keep our scientific mission moving forward.

[Jan van de Velde]

..//.. , but still keep our scientific mission moving forward.

That's a point that I fully appreciate. On the other hand, if that means that unicore clients are getting relatively useless towards that goal, (perhaps an overstatement, but that is what I read between the lines) why not simply say so?
No hard or bitter feelings on my part. But I'd rather quit gracefully than struggle on uselessly for some time until I finally give up anyway.

[7im]

They simply didn't say so because that is NOT the case. Unicore clients are still very numerous and very productive.

I personally run 2 of those clients on my home computer that doesn't fold 24/7, and I was one of the first to ring the bell when the shorter deadlines were preventing formerly productive computers from meeting the new shorter deadlines.

And now they are working to correct the problem. The emphasis is not of the part of their statement you quoted, but on this part...

Our short term plan is to extend the deadlines for the A4 projects.

Which means they will also take additional steps to correct the problem long term.

[P5-133XL]

We had a meeting today to work on this issue. Our short term plan is to extend the deadlines for the A4 projects. We'll make an announcement once we settle on details.

When the DL's are extended, are the old A4 WU's also going to be re-benchmarked? The A4's have a QRB so extending the DL will significantly affect the bonus points if not re-benchmarked.

[VijayPande]

We will change the k-factor to adjust for the change in deadline time.

[JeansOn]

Coming home from a journey, I was happy to read above posts.
VijayPande wrote, that there will be an announcement.
I think, I found it: http://foldingforum.org/viewtopic.php?f=24&t=19059

But I have doubt about it. It is not my goal, being realized there.
The time-pressure goes on and so I'll work and upload the running WUs next days with -oneunit Option.
This decision is not my last, but I cannot work@PC, fearing to lose my in background running WUs because the deadlines are too small.

I always liked f@h and a great thx to the admis and other dedicated users in this forum. They are very ready to help.