[FaD] 11 TechIMO members in the top 100

[Rovastar]

Good crunching by all.

And the way PresterJohn is crunching atm it will soon be 12.

(He's our fastest cruncher atm)

[Kc2iLq]

Wow! It's really cool to see so many fast crunchers on our team!


-Kc2iLq

[PresterJohn]

i'm starting to see some very strange problems with the 1.21k client. i had 8-9 machines that froze on a WU over the Memorial Day weekend...they just did nothing for 3+ days.

i'm a little upset by that...i don't know why (nor do i like) the fact that FAD forces the client to self-upload new versions. i had no problems with the J version...K is giving me nothing but grief.

if this continues i may have to rethink my level of fad participation...i don't like having my cows doing nothing.

[win2kuser]

i'm a little upset by that...i don't know why (nor do i like) the fact that FAD forces the client to self-upload new versions. i had no problems with the J version...K is giving me nothing but grief.

Correct me if I am wrong, but unselecting BETA testing in the projects tab of the control panel will stop the automatic upgrades.

If it is that bad, you could always download the current 121J version and reinstall that. How you would do that of course I wouldn't have a clue

Nothing worse than having a cruncher sat idle

[Wizzard~Of~Ozz]

Are you sure it froze? I haven't seen think freeze, it does take some time on some molecules (my local machines haven't put anything out for over 24hrs and one is a 2.1Ghz XP. )

If there was an error message, feel free to post it here and I will have it addressed.

One of the nice things about FaD is if there is a problem or a bug it's addressed with good speed.

As for the j version vs K, the only real change was to the different jobs you can select, there were also minor modifications to the way it crunches molecules.

There are only 2 outstanding bugs that are to be fixed.

1: excessive CPU rating for P4level celerons (1400 gets a CPU score of 210 under linux)

2: floating point error while running the benchmark routine on occasion (only happened to me once)

[PresterJohn]

>Correct me if I am wrong, but unselecting BETA testing in the projects tab of the control panel will stop the automatic upgrades.

i don't have BETA selected...i primarily only crunch cancer & respiratory disease.

i still have the J version...the problem is that the FAD client is designed to check its own version # and to upgrade itself when the FAD servers informs the client that a newer version is being pushed out. so running J will mean that it upgrades to K whenever I d/l'ed my next set of WUs.

Wizz, the error that i'm seeing consisted of two popups windows: the first i don't have written down, the second show an error msg (27) Formatted read error file: IOSTAT= 0

basically at that point, FAD needs to be shut down because if i click on continue, FAD will quickly run thru the rest of the queued jobs (spends about 10secs per WU) and show 'finished' without actually working on the WUs.

the 'freezing' isn't really the PC freezing, what happens is the FAD client just seems to reject the current wu (it shows a status code '3 Failed' in red on the queue) but it stays on the corrupted WU instead of moving to a new unit.

[Find-a-Drug]

I suspect this is a file corruption (27) file read error. To affect multiple clients it might be the *.pdb file (if they are all running the same job). Is there any text indicating the name of the file name? Is the problem still happening? What happens if you click More?

One of the requested changes which is being implemented is to terminate THINK if it encounters an error when running Hidden. tray will then restart it and hopefully the problem will not re-occur. This would avoid wasting CPU cycles.

[PresterJohn]

Originally posted by Find-a-Drug
I suspect this is a file corruption (27) file read error. To affect multiple clients it might be the *.pdb file (if they are all running the same job). Is there any text indicating the name of the file name? Is the problem still happening? What happens if you click More?

thanks for asking, Keith.

i don't think i saw a More button (i never clicked on one). the two choices i tried were 'Continue' and 'Exit'. clicking continue just proceeded to corrupt the remaining WU's in the queue. BTW, I am running individual queues on each workstation, ie. no centralized queue.

i haven't checked all the machines today...but the one on my desk which had the error seems OK for now.

i appreciate the personal response...i was pretty bummed out when i typed my initial post (complaining of the problem). i don't like to see wasted cpu cycles on my machine, .esp when i'm going to the extra expense & effort of running the client on many machines.

i don't think the UD people would have given me the same level of response (although i've never had to post a technical problem to them). but thanks again for dropping in here to answer...

cheers.

[PresterJohn]

also...thanks, Rovastar for posting this thread in the FAD forum.

[PresterJohn]

doesn't look good...my desktop machine got that file corruption (27) error. clicking on more shows the following


(27) Formatted read error file: IOSTAT= 0

[FCLOSE](120) Successful close of unallocated FORTRAN i/o unit 14
[GLOGIC] (200) Logical or environment name THINK_NAME: not found
[GLOGIC] (200) Logical or environment name d: not found
[GLOGIC] (200) Logical or environment name d: not found
[TXMOLE] (34) Reference to molecule -1 which does not exist
[SEARCH] (27) Error on return
[COMMDS] (279) File: j29-1e7u-q1.log Line:7
[FSYMBL] (140) Symbol not found ON_ERROR