It seems the l1_* files never get deleted, slowly filling up the disks of hosts until the quota is reached, effectively shutting down work for Einsein@Home after some time. Could be responsible for some hosts dropping out of E@H.
Nope, I didn't notice before. Thanks for pointing out. I wrote a note to the person in charge of the scheduler (which sends out the cleanup requests).
I plan to make this App official in the next few hours, mainly to get a more useful info regarding the client errors. It looks like it will at least make things not worse than the current official App (apart from Intel-SSE2 machines, which will see a small penalty of around 7%, but caugth up by the AMDs).
... I do hope it wasn't you again, Gary :-(. I was lucky to get away alive with it last time...
No it wasn't me this time :). You are obviously getting stuck into some other poor sod :).
These Linux/Windows validation issues just keep going on. Here is one of my AMD/Linux boxes that is suffering with three validation problems in its current list. Two have been finalised and a score of 0.0 awarded and they will probably disappear from the list shortly. A third is still pending but as usual, the "decider" has gone to a windows box so no joy there either.
It turns out there are a number of issues that lead to these cross-platform validation problems, some of which have been addressed recently, some we're still digging for. Solving these problems will probably require both a new validator and a complete set of Apps. I am confident that we will have all these pieces together next week.
.... I am confident that we will have all these pieces together next week.
That's really great news, thank you!! With those issues sorted out soon, and with the prospect of significantly optimised apps to follow, hopefully people will be encouraged to hang on a bit longer or even return if they had already left.
That's really great news, thank you!! With those issues sorted out soon, and with the prospect of significantly optimised apps to follow, hopefully people will be encouraged to hang on a bit longer or even return if they had already left.
I already have some code that should speed up computation significantly, but with the present issues it's simply impossible to validate it.
Would someone post here in this thread suggested procedure for those of us who got on 4.24 during Beta (using app_info.xml) to back out? This would mean we'd get the next official ap change automatically.
Is it as simple as stopping Boincmgr, removing app_info.xml in the Einstein directory, and restarting Boincmgr? (Assuming all results in queue are already tagged as 4.24 because they were downloaded after the previous change?)
Is it as simple as stopping Boincmgr, removing app_info.xml in the Einstein directory, and restarting Boincmgr? (Assuming all results in queue are already tagged as 4.24 because they were downloaded after the previous change?)
My first result with the Windows 4.24 app completed during the night. For a 340-cobblestone workunit, completion time went from 29.86 hours to 22.46 hours. So, a nice little speed-up.
RE: BTW, Bernd, did you
)
Nope, I didn't notice before. Thanks for pointing out. I wrote a note to the person in charge of the scheduler (which sends out the cleanup requests).
BM
BM
I plan to make this App
)
I plan to make this App official in the next few hours, mainly to get a more useful info regarding the client errors. It looks like it will at least make things not worse than the current official App (apart from Intel-SSE2 machines, which will see a small penalty of around 7%, but caugth up by the AMDs).
BM
BM
RE: ... I do hope it wasn't
)
No it wasn't me this time :). You are obviously getting stuck into some other poor sod :).
These Linux/Windows validation issues just keep going on. Here is one of my AMD/Linux boxes that is suffering with three validation problems in its current list. Two have been finalised and a score of 0.0 awarded and they will probably disappear from the list shortly. A third is still pending but as usual, the "decider" has gone to a windows box so no joy there either.
C'est la vie!! :).
Cheers,
Gary.
It turns out there are a
)
It turns out there are a number of issues that lead to these cross-platform validation problems, some of which have been addressed recently, some we're still digging for. Solving these problems will probably require both a new validator and a complete set of Apps. I am confident that we will have all these pieces together next week.
BM
BM
RE: .... I am confident
)
That's really great news, thank you!! With those issues sorted out soon, and with the prospect of significantly optimised apps to follow, hopefully people will be encouraged to hang on a bit longer or even return if they had already left.
Cheers,
Gary.
RE: That's really great
)
I already have some code that should speed up computation significantly, but with the present issues it's simply impossible to validate it.
BM
BM
I understand 4.24 is now the
)
I understand 4.24 is now the official release.
Would someone post here in this thread suggested procedure for those of us who got on 4.24 during Beta (using app_info.xml) to back out? This would mean we'd get the next official ap change automatically.
Is it as simple as stopping Boincmgr, removing app_info.xml in the Einstein directory, and restarting Boincmgr? (Assuming all results in queue are already tagged as 4.24 because they were downloaded after the previous change?)
Thanks
RE: Is it as simple as
)
Yes it is.
BM
BM
The notice on the Home Page
)
The notice on the Home Page regarding 4.24 is dated June 13, 2007. You may want to fix it...
Seti Classic Final Total: 11446 WU.
My first result with the
)
My first result with the Windows 4.24 app completed during the night. For a 340-cobblestone workunit, completion time went from 29.86 hours to 22.46 hours. So, a nice little speed-up.