FahMon (multi-platform app to monitor various F@h clients)

[uncle_fungus]

It is a bug, and here's the technical explanation as to why.

The "frame" count on psummary is better described as an Intermediate Data Point (IDP) count.
The researchers running these simulations don't just want to know what the state of the WU was at the beginning and the end of the simulation, they also want to know what happens in-between (there was a special case a while back where a particular Gro33 WU had an IDP of zero), and the number of "in-between" states is the IDP.
For example, in the results file for a WU with an IDP of 50 there are 50 individual snapshots of the system state at different times during the simulation.

Now, the confusion arises because the IDP count has no real relationship with how many "frame" markers there are in a given FAHlog.txt. It has just been coincidence that most of the time they match up (this is also an example of the unfortunate use of the word frame to mean many different things).
Simply put you cannot work out how many markers there will be in a logfile by looking at the IDP count because different cores have different logging logic.

To correct this, FahMon 2.3.2 distrusts the psummary frame count and calculates how many markers there are going to be in a log file as it goes. When it can't make a determination it falls back to the psummary value (or another sensible default) but marks the PPD with an asterisk to indicate that it (and the ETA) may not be accurate.

[uncle_fungus]

FahMon 2.3.2beta is ready for testing.

This an unsupported release that internally identifies itself as 2.3.1 so the update checker will still alert you when 2.3.2 is released.

See here: http://fahmon.net/news/archives/2008/01 ... _51_17.txt

[francois]

Thanks!
The ticket 76 seems to have been resolved too, at least here at work.
FahMon doesn't crash anymore when the wallpaper is changed and icon background transparency is set!

Will try tonight at home and report later.

Best regards,
François

[uncle_fungus]

Thanks!
The ticket 76 seems to have been resolved too, at least here at work.
FahMon doesn't crash anymore when the wallpaper is changed and icon background transparency is set!

Will try tonight at home and report later.

Best regards,
François

Ah! I'm not entirely sure how I've fixed it, but if it works...

[ICE_9]

The windows zip says fahmon.exe is password protected. Do I need to request a password?

[uncle_fungus]

The windows zip says fahmon.exe is password protected. Do I need to request a password?

No. What are you using to open the zip file as none of the files are password protected?

[ICE_9]

Windows zip. no external program. just the extract files wizard.

[ICE_9]

nm. I figured it out. I had it starting as a service in the background and it was still on.

[uncle_fungus]

nm. I figured it out. I had it starting as a service in the background and it was still on.

Starting as a service? Does that work?

[bruce]

Is it really a bug or is it just a problem with that the WU actually has 100 frames & not really 50.

It depends on the definition of "frames" Apparently these WUs have 50 frames but 100 progress messages. Externally, all that can be seen are the progress messages and FahMon assumes that there is one message per frame.

(The GUI client uses a third definition of "frames" but let's not go there.)

[TheWolf]

Yea I have always called them frames as this is what I have always seen others say that the percentage steps are called.
I know the GUI client is much different with its progress of steps. Good to see someone is looking into the problem. Thanks

I'll give the Fahmon beta a try, thanks again. Keep up the good work guys.

[Hyperlife]

I just tried running 2.3.2b on Windows after a clean install in its own folder, and when one particular client is added, it consistently crashes on me. All other clients can be added with no problems.

The problem client is running p4117 (R3 C5 G3) right now -- it's the only client of mine that's running a 4117 WU. The messages.log doesn't show anything beyond a "Reloading" message for the problem client.

Anything else I can do to isolate this problem?

[TheWolf]

Looks like the 50 frame bug is fixed in 2.3.2b beta. Everything looks to be showing correct even once it passes the 50 percent point.
Times looks to be correct & PPD is about half what it was.

I just copied my config folder into the new 2.3.2b folder, so I didn't have to resetup all my clients/networks rigs.
I did do the check for updates from the psummary page & everything looks good so far. I'll have to keep a look out on the CPU usage.
I only monitor 9 rigs of 11 with in the program, so I'm not sure if this will be enough to show the CPU usage problem?

Thanks..

[toTOW]

FahMon 2.3.2beta is ready for testing.

This an unsupported release that internally identifies itself as 2.3.1 so the update checker will still alert you when 2.3.2 is released.

See here: http://fahmon.net/news/archives/2008/01 ... _51_17.txt

It solved the 50 frames WU issue

[uncle_fungus]

I just tried running 2.3.2b on Windows after a clean install in its own folder, and when one particular client is added, it consistently crashes on me. All other clients can be added with no problems.

The problem client is running p4117 (R3 C5 G3) right now -- it's the only client of mine that's running a 4117 WU. The messages.log doesn't show anything beyond a "Reloading" message for the problem client.

Anything else I can do to isolate this problem?

OK yes I see the problem, I've just confirmed it here. It seems that there's still an issue I haven't fixed that means missing project data crashes FahMon.
Edit: I've just fixed the problem so it'll work in the final release (and any subsequent betas if there are any).

I'll have to keep a look out on the CPU usage.
I only monitor 9 rigs of 11 with in the program, so I'm not sure if this will be enough to show the CPU usage problem?

I'm not sure but I'd appreciate it if you could keep an eye on it anyway (having the web apps enabled does increase the CPU usage per update though).

Looks like the 50 frame bug is fixed in 2.3.2b beta. Everything looks to be showing correct even once it passes the 50 percent point.
Times looks to be correct & PPD is about half what it was.

It solved the 50 frames WU issue

That's good to hear