Why do you think that?? They will be being sent out again and if it's a problem with the validator rather than the result, it will most likely get 'fixed'. The last result in Holmis' list has already been 'fixed' so it does look promising.
On the GPUs ("O2MDF") we have another chunk of the "V2" workunits. Based on previous experience these should run about twice as long as expected (e.g. like the "G2" ones). I doubled the credit and flops estimation to make up for that, hope that this helps.
You have the dreaded CL_MEM_OBJECT_ALLOCATION_FAILURE on some of those Vela Junior tasks for your computer with a 2GB Nvidia card. See this thread for more information.
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
https://einsteinathome.org/workunit/447314688 I have 22 in a row of these. Thos was running S@H just fine and runs pulsars fine also. This card is a GTX10603GB. Methinks it is bad data not the host. Most fail in ~ 1min.
I rebooted the offending box and have now successfully completed 7 in a row so the problem seems to have been the host. The are in a pending status so time will tell.
https://einsteinathome.org/workunit/447314688 I have 22 in a row of these. Thos was running S@H just fine and runs pulsars fine also. This card is a GTX10603GB. Methinks it is bad data not the host. Most fail in ~ 1min.
Betreger wrote:
I rebooted the offending box and have now successfully completed 7 in a row so the problem seems to have been the host. The are in a pending status so time will tell.
Unless the problem is the immediate consequence of the release of a new or modified app that has just been announced here, Technical News is NOT the best forum to report new problems in longer running searches. There is a Problems forum specifically for that purpose. There is also no use in reporting a problem in multiple places. You just create more work for the already overworked Devs in trying to keep up with all reports that are coming in. You just encourage the 'me too' and the 'maybe it could be me too (but it's actually different)' reports to be in different places as well.
At the start of every day, I check the problems forum first and try to deal with any overnight problem reports, if I can. When I checked your report, it must have been just before you rebooted because there were only failed tasks, and none in progress at the instant I looked. It's always a good idea to try a reboot before declaring that a problem exists.
It is still quite possible that there really could be memory allocation issues with these higher frequency tasks so if you see further examples, please report it in the Problems forum.
Betreger wrote:It seems odd
)
Why do you think that?? They will be being sent out again and if it's a problem with the validator rather than the result, it will most likely get 'fixed'. The last result in Holmis' list has already been 'fixed' so it does look promising.
Cheers,
Gary.
Seems all of them have been
)
Seems all of them have been fixed now. Thanks for taking care of it!
On the GPUs ("O2MDF") we have
)
On the GPUs ("O2MDF") we have another chunk of the "V2" workunits. Based on previous experience these should run about twice as long as expected (e.g. like the "G2" ones). I doubled the credit and flops estimation to make up for that, hope that this helps.
BM
Validate errors have
)
Validate errors have returned.
https://einsteinathome.org/workunit/438858562 https://einsteinathome.org/workunit/438909767 https://einsteinathome.org/workunit/438611172I'm getting a fair number of
)
I'm getting a fair number of "Error while computing" on both boxes.
https://einsteinathome.org/workunit/441408802Validate errors are
)
Validate errors are back
https://einsteinathome.org/workunit/443513006 https://einsteinathome.org/workunit/442048724 https://einsteinathome.org/workunit/445576021You have the
)
You have the dreaded CL_MEM_OBJECT_ALLOCATION_FAILURE on some of those Vela Junior tasks for your computer with a 2GB Nvidia card. See this thread for more information.
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
https://einsteinathome.org/wo
)
I rebooted the offending box
)
I rebooted the offending box and have now successfully completed 7 in a row so the problem seems to have been the host. The are in a pending status so time will tell.
Betreger
)
Unless the problem is the immediate consequence of the release of a new or modified app that has just been announced here, Technical News is NOT the best forum to report new problems in longer running searches. There is a Problems forum specifically for that purpose. There is also no use in reporting a problem in multiple places. You just create more work for the already overworked Devs in trying to keep up with all reports that are coming in. You just encourage the 'me too' and the 'maybe it could be me too (but it's actually different)' reports to be in different places as well.
At the start of every day, I check the problems forum first and try to deal with any overnight problem reports, if I can. When I checked your report, it must have been just before you rebooted because there were only failed tasks, and none in progress at the instant I looked. It's always a good idea to try a reboot before declaring that a problem exists.
It is still quite possible that there really could be memory allocation issues with these higher frequency tasks so if you see further examples, please report it in the Problems forum.
Cheers,
Gary.