gb009761 is correct. We needed to over schedule the work units. Normally we only send 2 results and require a quorum of 2 for Help Conquer Cancer. If we waited for the results to return on their own it would normally take 13-14 days for a batch to complete. Doing it this way, caused the batches to complete in about 8 days.

Server aborts only happen if the work unit has been determined valid and your computer has not started running the work unit. This will help your computer from starting to run a work unit that has already been considered completed. No time is wasted in this case. All that happened was some extra network download.

Please remember that over scheduling is not common. On a normal day there are more than 400,000 work units in progress for Help Conquer Cancer. We only started the over schedule when this number was close to 50,000.

Hope this helps.

Thanks,
-Uplinger

Thanks uplinger and gb009761 for yours explanations.
But in a my case i'am sure ( 99% - we are never sure 100%) my job was canceled after I did it. My computer ( preference parameter: change job after = 6000 minutes) takes jobs by its beginning date and stops those one they older even if they are running. My jobs class was after 10/4/11 when i received this HCC job
In this cas Job 0 and job 1 were sent on 1/01/11, and were sent back 8/4/11.
The two others jobs job 2 and 3 were send on 6/4/11 and 7/4/11 respectively.
So I think the logical process you assume to save wasted time, didn't operate in this case. This the reason why I send this message because i'm quite sure i had completely done my job.
I don't worry about points, I only ask if the system does as well as you tell me.
Thanks for your patience to read an "old french grumpy man" :))) All I want is to be wrong: 1% of chance ?

bernie99, please feel free to send me your messages log to support@worldcommunitygrid.org. Please look in your Boinc Data directory and send both the stdoutdae.txt and stderrdae.txt.

Atlantis555, it should run faster with other WCG projects.

Thanks,
-Uplinger

I myself thought that only hcc was speeded up. Now it turns out that all projects are set in motion. Somehow strange. Let's see what Uplinger will answer.

Atlantis555,

I had thought your question was in regard to running multiple work units from different projects at once and if the speed up would still be noticed.

As the change in compiler may work for other projects it would not work for others. HCC runs C++ code and other projects run Fortran. Some require a specific compiler based on the requirements sent in by the researchers. Also, the speed up due to a compiler change may not work on the science code due to the routines that are called. So there are a few things to evaluate we do not have plans to change the compiler for any projects right now due to time constraints.

Hope this helps.

Thanks,
-Uplinger