Folding Forum

I am just curious to find out why there have been no A7 core WU's assigned lately...

Paul

Once a new Core is out, like Core_a8, all new work is started in a8. However, any work already in progress is completed in the Core it was started in, so the results are consistent. So for some time, often a couple years, Work Units will be assigned for both Cores.

As the old a7 work draws to a close, they become rarer, until some day all project started before the new Core are complete. Later there maybe dribbles of old WUs as the results are analyzed and some folding paths are discovered that were not fully traversed.

Then an all now Core is released (a9?) and the process repeats. Core_a8 takes advantage of new CPUs with AVX2_256 instructions like FMA3. Since then, many server class Intel CPUs have implemented AVX_512. Will that be part of Core_a9? Maybe.

I figured the A7 jobs are going to the older & slower systems while they are usable.

Paul

psaam0001 wrote:I figured the A7 jobs are going to the older & slower systems while they are usable.

Paul

I don't believe that's the case.

I'll fold on none the less... Let's wrap those A7 projects up, and move forward.

Paul

Looking at the A7 projects listed as active on the Project Summary page, most are on two servers that are currently not actively assigning WUs. They are in
Accept status. These projects may have reached the limits set for runs, clones and gens set by the researcher and their data is being analyzed to see enough has been collected. If more data is needed, then either more runs, clones or gens will be added to the existing project, or a new project on the same protein system would be created to get that data. The new project would still be A7 to simplify data verification.

Joe_H wrote:Looking at the A7 projects listed as active on the Project Summary page, most are on two servers that are currently not actively assigning WUs. They are in Accept status. These projects may have reached the limits set for runs, clones and gens set by the researcher and their data is being analyzed to see enough has been collected. If more data is needed, then either more runs, clones or gens will be added to the existing project, or a new project on the same protein system would be created to get that data. The new project would still be A7 to simplify data verification.

That status from the server stats page is what peaked my curiosity, when I saw nothing being assigned, and 72k units that could be.

Thank you all... Now back to my regular folding in progress.

Paul

psaam0001 wrote:I figured the A7 jobs are going to the older & slower systems while they are usable.

So far as I know, all CPUs newer than the Pentium 4 can do the calculations, mind you the speed may not meet deadlines. F@H needs at least SSE2 math units.