How can I...

[fredex]

How can I tell whether I'm being given "big" WUs or not? (Linux ver 6.x client)

tks!

[P5-133XL]

There are several questions you need to ask. The first is if you are using the SMP client (All SMP WU's are "Big"). If you are using the uniprocessor client then go to the configuration and then look to see if you allow Big WU's because you won't get them if you don't explicitly allow them. Third, you can look at the logs specifically for the lines that specify how much data was written to disk: The big ones write a lot more data (20+MB) than the small ones (~1.5MB).

[fredex]

NOT smp client, and yes it's configured for big WUs.

but I don't see anything in any of the log files (that I can find) that mentions the size, nor how many bytes are written to disk either when initially downloading the WU, or uploded when it's completed.

What am I overlooking?

Thanks!

[P5-133XL]

Are you using the -verbosity 9 flag? I suggest adding it.

Then you should get output that looks somewhat like:

[16:59:20] Completed 500000 out of 500000 steps (100%)
[16:59:21] DynamicWrapper: Finished Work Unit: sleep=10000
[16:59:31]
[16:59:31] Finished Work Unit:
[16:59:31] - Reading up to 20317968 from "work/wudata_09.trr": Read 20317968
[16:59:31] trr file hash check passed.
[16:59:31] edr file hash check passed.
[16:59:31] logfile size: 57029
[16:59:31] Leaving Run
[16:59:36] - Writing 20409273 bytes of core data to disk...
[16:59:37] ... Done.
[16:59:47] - Shutting down core
[16:59:47]

Just note that not all cores send the same data to the log so your may look slightly different.

[toTOW]

The "size" of the WU can relate to three things (simultaneously or not) :

- download size
- upload size
- memory usage

You could also post your PRCG numbers : we might know the status of the WU if someone else already folded it ...

[P5-133XL]

The "size" of the WU can relate to three things (simultaneously or not) :

- download size
- upload size
- memory usage

You could also post your PRCG numbers : we might know the status of the WU if someone else already folded it ...

At no point did he say he had a problem or something he needed checked. It was just asked how to distinguish between big and small WU's.

[toTOW]

I know ... and I'm asking for the WU references because I (or someone else) might know whether the project is big or not ...

[7im]

BigWUs also receive the Big WU bonus, so if your PPD is significantly higher on some WUs than others, you know those are the big work units.

Note that using the Big setting does not guarantee you a BigWU every time. It only states a preference to receive those WUs, if they are available to download for your client type, settings, etc.

[fredex]

from the FAHlog.txt for a recent WU:

• [11:58:30] Completed 499300 out of 499375 steps (99%)
  [11:58:36] Saving result file logfile_05.txt
  [11:58:36] Saving result file checkpt
  [11:58:36] Saving result file checkpt.crc
  [11:58:36] Saving result file log.txt
  [11:58:40] Saving result file protomol.conf
  [11:58:40] Saving result file ww.36.pos
  [11:58:40] Saving result file ww.36.vel
  [11:58:40] Saving result file ww.dcd
  [11:58:41] Folding@home Core Shutdown: FINISHED_UNIT
  [11:58:41] CoreStatus = 64 (100)
  [11:58:41] Sending work to server
  [11:58:41] - Read packet limit of 540015616... Set to 524286976.
  
  
  [11:58:41] + Attempting to send results
  [11:59:59] + Results successfully sent
  [11:59:59] Thank you for your contribution to Folding@Home.
  [11:59:59] + Number of Units Completed: 320
  
  [11:59:59] - Preparing to get new work unit...
  [11:59:59] + Attempting to get work packet
  [11:59:59] - Connecting to assignment server
  [12:00:00] - Successful: assigned to (129.74.85.15).
  [12:00:00] + News From Folding@Home: Welcome to Folding@Home
  [12:00:00] Loaded queue successfully.
  [12:00:01] + Closed connections

I don't see anything there about the WU size, so this particular core must not be "divulging" that info.

I'm running two separate instances of FAH on a dual-core Phenom II under 32-bit Linux. I'm also using fahmon, and it shows these things for the two WUs currently being processed:

Code: Select all

PPD                  PRCG
262.32               P10017 (R843, C1, G1)
244.64               P10017 (R2811, C0, G1)

and the "work unit information" pane says "Credit: 126 points" for each WU.

[7im]

Are you using the -verbosity 9 switch as recommended above?

Plus I don't think any of the Protomol WUs are BigWUs. So you're probably right about it not writing that size info to the log.

P.S. I recommend putting a bit more info in your title next time. A lot of people will skip threads with titles like that.

[kzaske]

Given the title of this thread I think I will post my question here: How can I tell when I first signed up with F@H? I know the data has to be hiding somewhere...

[7im]

My exact point about thread titles...

First signed up info? That info is not hiding. It's just not published anywhere. Some of the 3rd party stats sites publish the date of the first record they have regarding your first returned work unit. For example, EOC shows your first WU as being 02.06.07 http://folding.extremeoverclocking.com/ ... =&u=207325 top right corner