Download and remove!!!

[davidw]

Ok, On my second cow, I downloaded 11 jobs. The weird thing is, the jobs were inserted into the queue but it removed 11 jobs prior to that. For example...

I had jobs 110 to 118 uploaded.
I show jobs 119 running
I have 120 to 136 remaining.
I download 11 jobs - It should read 120 to 147.
Jobs 120 to 130 are removed.

I download 53 more jobs - no problem
So now it shows 119, 131-200

What causes this problem and where did the jobs go to?

[Plum Ugly]

that was the new purge feature.This done because the jobs have already been done by somebody else and turned in.
I have that happen on my slower machines.

[davidw]

I can understand if it is specific jobs and they are randomly removed (such as 122, 131, 135, 156, etc), but this was consecutive. Does this mean that I'm running jobs that need to be deleted and then download a different type of job? They are the 1fgi series.

[Plum Ugly]

It will purge the ones it needs to by itself.
the 1fgi series are faster running jobs.I have 1 or 2 machine running them,I run about 10-15 jobs a day on each of them. on my twin 500 I only run 2 in queue and let it down load automaticly.
The HIV 1A30 are longer running jobs with high hits.On my 2400's they can run 24 hrs and have 250-500 hits. I rarely have anything that takes over 30 hrs per jobs. And on the new cancer jobs I can run 10-15 jobs per day on the faster machines.
On the machines that aren't online I run 10 in queue.Puserver runs on automatic downloads and changes by itself what molies are downloaded.I run Hiv,Cancer,Methodology on it.It can pulldown 20-30 jobs on the faster molies.

[Twinkletoes]

C-C - I wouldn't worry about it. Strange things like that are happening to me frequently, and I can't see any pattern to it.
I've given up trying to work out the logic behind it, and just hope that the software knows what it's doing.

But I do wish one got some kind of warning about it. You see you're sitting there with 10 jobs left, so better download another 10, and when you look later you've still only got a queue of 10, but now they are different jobs ! Very disconcerting...

[davidw]

That's practically what happened.

[PresterJohn]

>Jobs 120 to 130 are removed.

C-C,

as plum mentioned, those jobs were removed because results were already returned. the 1fgi's are very fast running mollies and many many people are running them since it also happens to be a cancer query as well.

this is why i have repeatedly mentioned in previous stats posts that all members are highly encouraged to run a purge (even when new jobs do not need to be downloaded) so as to ensure that we don't do more 'repeated' work than is avoidable.

[davidw]

so should i move on to another type, instead of the 1fgi? if all im doing is redundancy work?

[davidw]

You'd think that FaD would be smart enough to purge the ones that needed to be purged by checking with what has been returned instead of just taking a chunk of consecutive jobs.

[PresterJohn]

Originally posted by Christian-Comp
so should i move on to another type, instead of the 1fgi? if all im doing is redundancy work?

no, you're not doing redundant work. if the purge removed 10 jobs but you are still left w/the 10 new ones you downloaded, then those WU's need to be crunched. ie. they are NOT redundant.

the point i was i trying to make is that if a particular query happens to have very fast running mollies, then it might be advantageous to run purge a little more often (at least every 24hrs) because completed jobs within even a 24 hr period will probably be quite high. those of you who are running short queues probably don't have to do anything if you have it to auto-connect since the purge will be performed automatically when new WU's are downloaded.