request_reschedule_cpus:
7 May 2005 3:50:30 UTC
Topic 189121
(moderation:
I can't get any work, I was running okay until I loaded 4.35?? on Windows XP.
5/6/2005 11:46:41 PM|Einstein@Home|Sending request to scheduler: http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
5/6/2005 11:46:42 PM|Einstein@Home|Scheduler RPC to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi succeeded
5/6/2005 11:47:01 PM|Einstein@Home|Resetting project
5/6/2005 11:47:01 PM||request_reschedule_cpus: exit_tasks
5/6/2005 11:47:01 PM||request_reschedule_cpus: project op
5/6/2005 11:47:01 PM||schedule_cpus: must schedule
request_reschedule_cpus:
)
4.35 and 4.36 have known scheduling and benchmark issues. I upgraded to 4.37 and this seems to have resolved itself. Give it a try...
4.35 and 4.36 have known
)
4.35 and 4.36 have known scheduling and benchmark issues. I upgraded to 4.37 and this seems to have resolved itself. Give it a try...
Do you have a link to where these "known issues" are documented, thanks. I'm running a number of boxes on 4.36, so far without issue. If there's a widespread problem then I'll jump ship immediately :).
With this bleeding edge stuff you could be jumping out of the frying pan and into the fire anyway. I must admit that the so-called "stable" releases seem to be really no better than the bleeding edge so you may as well be at the edge :).
Edit: Before anyone takes that as a serious recommendation, please realise it's IMHO :).
Cheers,
Gary.
Do you have a link to where
)
Do you have a link to where these "known issues" are documented, thanks. I'm running a number of boxes on 4.36, so far without issue. If there's a widespread problem then I'll jump ship immediately :).
There are numerous posts explaining 'oddities' on the S@H boards.
The benchmark bug was that the part of the code that halts the science application prior to benchmarking was borked, so if the CC (or you) decided to benchmark in the middle of processing, that WU would error out. This is supposed to be fixed in 4.38. Some people also reported benchmarks locking up their systems. Not sure if this is related, but it seems likely.
As for the scheduler, 4.36 had several oddities. It would do things like request additional CPDN WUs when you still had hundreds of hours of processing time left on your current one, as well as requesting additional WUs from the ROC projects, when those WUs run for a year. 4.37 fixed some of the problems, but introduced a few more. ;)
Now lets see what 4.38 is all about. :)