Project: 6013 (Run 0, Clone 160, Gen 98) extremly slow!

[Wrish]

One of my dual-cores had this unfortunate assignment. With an expected completion time of 107 hours versus a 72-hour hard deadline, I had to delete it. What's with the rash of Run-0 6013's that take almost as long as bigadv units?

Code: Select all

[08:33:38] - Connecting to assignment server
[08:33:39] - Successful: assigned to (130.237.232.140).
[08:33:39] + News From Folding@Home: Welcome to Folding@Home
[08:33:39] Loaded queue successfully.
[08:33:49] + Closed connections
[08:33:49] 
[08:33:49] + Processing work unit
[08:33:49] Core required: FahCore_a3.exe
[08:33:49] Core found.
[08:33:49] Working on queue slot 04 [June 17 08:33:49 UTC]
[08:33:49] + Working ...
[08:33:49] 
[08:33:49] *------------------------------*
[08:33:49] Folding@Home Gromacs SMP Core
[08:33:49] Version 2.19 (Mar 12, 2010)
[08:33:49] 
[08:33:49] Preparing to commence simulation
[08:33:49] - Assembly optimizations manually forced on.
[08:33:49] - Not checking prior termination.
[08:33:50] - Expanded 979403 -> 10427873 (decompressed 1064.7 percent)
[08:33:50] Called DecompressByteArray: compressed_data_size=979403 data_size=10427873, decompressed_data_size=10427873 diff=0
[08:33:50] - Digital signature verified
[08:33:50] 
[08:33:50] Project: 6013 (Run 0, Clone 160, Gen 98)
[08:33:50] 
[08:33:50] Assembly optimizations on if available.
[08:33:50] Entering M.D.
[08:34:13] Completed 0 out of 250000 steps  (0%)
[09:38:30] Completed 2500 out of 250000 steps  (1%)
[10:42:23] Completed 5000 out of 250000 steps  (2%)

[Mr.Nosmo]

Normaly I'm NOT complaning or picking WU's that I don't like, but now I'm thinking of doing it!

I have read the posts about 6013, 6040/6041 & 6701 and I still don't undertstand PG!

I'm running on a REAL 8-core Xeon (5430@3100MHz) and normaly I get about 6600 Points (17K+ PPD - was 20+K), but now I get 65 PPD (says FahSpy)!!!

[16:22:08] Project: 6013 (Run 0, Clone 160, Gen 98)
[16:22:08]
[16:22:08] Assembly optimizations on if available.
[16:22:08] Entering M.D.
[16:22:30] Completed 0 out of 250000 steps (0%)
[17:50:48] Completed 2500 out of 250000 steps (1%)
[19:16:59] Completed 5000 out of 250000 steps (2%)

Is this good for science?

I think PG need to re-think the whole thing about points and start to select the systems that works best for that WU!

There are many systems available for tests before a WU is send out, so please use them! If you want to benchmark all your WU's before they are released on my HW, just say so!

I was thinking of being a beta-tester, but I don't have programming skills and some of the other things needed to be invited, so I didn't apply, but if my HW & time can be used to benchmark WU's, so they will be send to the best (efficent) machines for this WU, I'll be happy to help (and I'm sure a lot of others are too).

Yes, I konw there are issues with "out of WU's" for this machine - So issue the WU the second best machine!

Am I completly wrong?

[7im]

There is a section of the forum dedicated for reporting problems with WUs. I don't think this is a problem with the SMP client in particular. I think a Mod can move this for you...

Edit by Mod: Topics merged/moved.

[bruce]

I've reported Project: 6013 (Run 0, Clone 160, Gen 98) as a bad WU.