Project: 780 and 785 - super low PPD

[morko]

Hello.

I'm folding these two projects on Linux on my core2duo E6300 system. Is it possible that this frame time I'm getting (1h 40 min) is right? It gives me like super low PPD of about 80 PPD? This box can do much more for sure.

What is wrong here? Benchmarking, my box, something else?

Project: 780 (R3, C53, G1)
Project: 785 (R5, C97, G2)

[morko]

This Gromacs 33 core must have some problem with proper detection of SSE and 3DNow instruction set on my system. (Core2Duo E6600) I have to use -forceasm flag. After that the frame time is 100% shorter.
Now I have 50 minutes time for 1 % (before 100 minutes).

Is there anyone else folding this projects?

Executable: ./fah6

Code: Select all

[18:15:08] - Ask before connecting: No
[18:15:08] - User name: morko (Team 10824)
[18:15:08] - User ID: 3FC1C54242D9AA26
[18:15:08] - Machine ID: 1
[18:15:08]
[18:15:08] Loaded queue successfully.
[18:15:08]
[18:15:08] + Processing work unit
[18:15:08] Core required: FahCore_a0.exe
[18:15:08] Core found.
[18:15:08] Working on Unit 06 [September 30 18:15:08]
[18:15:08] + Working ...
[18:15:09]
[18:15:09] *------------------------------*
[18:15:09] Folding@Home Gromacs 3.3 Core
[18:15:09] Version 1.92 (April 17. 2007)
[18:15:09]
[18:15:09] Preparing to commence simulation
[18:15:09] - Ensuring status. Please wait.
[18:15:26] - Looking at optimizations...
[18:15:26] - Working with standard loops on this execution.
[18:15:26] - Previous termination of core was improper.
[18:15:26] - Files status OK
[18:15:26] - Expanded 1171344 -> 6252409 (decompressed 533.7 percent)
[18:15:26]
[18:15:26] Project: 785 (Run 5, Clone 97, Gen 2)
[18:15:26]
[18:15:26] Entering M.D.
[18:15:48] (Starting from checkpoint)
[18:15:48] Protein: Mini chaperonin
[18:15:48] Writing local files
[18:15:48] Completed 46482 out of 500000 steps  (9%)

Executable: ./fah6 -forceasm

Code: Select all

Arguments: -forceasm

Warning:
 By using the -forceasm flag, you are overriding
 safeguards in the program. If you did not intend to
 do this, please restart the program without -forceasm.
 If work units are not completing fully (and particularly
 if your machine is overclocked), then please discontinue
 use of the flag.

[18:16:25] - Ask before connecting: No
[18:16:25] - User name: morko (Team 10824)
[18:16:25] - User ID: 3FC1C54242D9AA26
[18:16:25] - Machine ID: 1
[18:16:25]
[18:16:25] Loaded queue successfully.
[18:16:25]
[18:16:25] + Processing work unit
[18:16:25] Core required: FahCore_a0.exe
[18:16:25] Core found.
[18:16:25] Working on Unit 06 [September 30 18:16:25]
[18:16:25] + Working ...
[18:16:25]
[18:16:25] *------------------------------*
[18:16:25] Folding@Home Gromacs 3.3 Core
[18:16:25] Version 1.92 (April 17. 2007)
[18:16:25]
[18:16:25] Preparing to commence simulation
[18:16:25] - Assembly optimizations manually forced on.
[18:16:25] - Not checking prior termination.
[18:16:26] - Expanded 1171344 -> 6252409 (decompressed 533.7 percent)
[18:16:26]
[18:16:26] Project: 785 (Run 5, Clone 97, Gen 2)
[18:16:26]
[18:16:26] Assembly optimizations on if available.
[18:16:26] Entering M.D.
[18:16:47] (Starting from checkpoint)
[18:16:47] Protein: Mini chaperonin
[18:16:47] Writing local files
[18:16:47] Completed 46482 out of 500000 steps  (9%)
[18:16:47] Extra 3DNow boost OK.
[18:16:47] Extra SSE boost OK.

[toTOW]

How do you quit your client ?

These message indicates that you're not doing it correctly :

[18:15:09] - Ensuring status. Please wait.
[18:15:26] - Looking at optimizations...
[18:15:26] - Working with standard loops on this execution.
[18:15:26] - Previous termination of core was improper.

[morko]

This time I quit with ctrl-C from console, because I want to show the difference in log with or without using -forceasm.

What is the proper way of quiting, anyway?

[toTOW]

CTRL-C is the best one indeed. By the way, -forceasm is the best way to be sure that your client will always use optimizations (if your machine is stable, you can use it safely).