My two Linux hosts have been going through periodic and alternating outages of v2.10 GW GPU tasks. When one host runs dry, I'll change the Project Prefs to allow non-preferred GPU apps; so far only FGRBPG1 and O2AS20-500 v1.09 tasks have been downloaded. The host will run through those for a half day or a day until it begins to pick up v2.10 tasks again; then I switch Prefs back to exclusive v2.10 work and off it goes with plenty of those tasks. Within a day or so the other host will dry up and I'll go through the same procedure with that one. Anyone else with this issue?
Ideas are not fixed, nor should they be; we live in model-dependent reality.
My Einstein host allows non- preferred tasks but gets mostly GW work. So I would say I don't have that problem. Finding a GW would be first prize but a pulsar would not be terrible.
I can't comment on that running dry issue as I currently can't run 24/7. However when running FGRBPG1 and O2MD1 2.01 on one host, the estimated remaining time seems to suffer - a matter of the DCF ?
I can't comment on that running dry issue as I currently can't run 24/7. However when running FGRBPG1 and O2MD1 2.01 on one host, the estimated remaining time seems to suffer - a matter of the DCF ?
Some users have reported that the Gamma Ray GPU task doesn't play well with the older GW tasks. That might be the case with the new tasks, also. It was discussed at length in this thread.
I can't comment on that running dry issue as I currently can't run 24/7. However when running FGRBPG1 and O2MD1 2.01 on one host, the estimated remaining time seems to suffer - a matter of the DCF ?
Some users have reported that the Gamma Ray GPU task doesn't play well with the older GW tasks. That might be the case with the new tasks, also. It was discussed at length in this thread.
Thanks for that hint, but it doesn't fit to my observation in this case. I should have been clearer, since I ran FGRBP1G and O2MD1 on two different GPU in one box. Your hint though inspired me to search in the forum. That search returned a comment by MARKJ in this thread:
200 valid 0 invalid so far,
)
200 valid 0 invalid so far, things are looking good here.
Up tp 250 valids and no
)
Up tp 250 valids and no invalids. It continues to look good.
340 valid 0 invalid with a
)
340 valid 0 invalid with a GTX1060 plus whatever the CPU does. Is it ready for prime time?
My two Linux hosts have been
)
My two Linux hosts have been going through periodic and alternating outages of v2.10 GW GPU tasks. When one host runs dry, I'll change the Project Prefs to allow non-preferred GPU apps; so far only FGRBPG1 and O2AS20-500 v1.09 tasks have been downloaded. The host will run through those for a half day or a day until it begins to pick up v2.10 tasks again; then I switch Prefs back to exclusive v2.10 work and off it goes with plenty of those tasks. Within a day or so the other host will dry up and I'll go through the same procedure with that one. Anyone else with this issue?
Ideas are not fixed, nor should they be; we live in model-dependent reality.
My Einstein host allows non-
)
My Einstein host allows non- preferred tasks but gets mostly GW work. So I would say I don't have that problem. Finding a GW would be first prize but a pulsar would not be terrible.
I can't comment on that
)
I can't comment on that running dry issue as I currently can't run 24/7. However when running FGRBPG1 and O2MD1 2.01 on one host, the estimated remaining time seems to suffer - a matter of the DCF ?
solling2 wrote:I can't
)
Some users have reported that the Gamma Ray GPU task doesn't play well with the older GW tasks. That might be the case with the new tasks, also. It was discussed at length in this thread.
https://einsteinathome.org/content/o2-all-sky-gravitational-wave-search-gpus-discussion-thread
Matt White schrieb:solling2
)
Thanks for that hint, but it doesn't fit to my observation in this case. I should have been clearer, since I ran FGRBP1G and O2MD1 on two different GPU in one box. Your hint though inspired me to search in the forum. That search returned a comment by MARKJ in this thread:
https://einsteinathome.org/de/content/are-there-settings-allow-running-gpu-cpu-wus-wo-intervention#comment-169743
Thus it seems when the DCF plays against you, you got to adjust your downloads manually from time to time. :-)
I am now showing 460 valids
)
I am now showing 460 valids and 0 invalids. I'm liking it.
win10 amd r9 270x O2MD1
)
win10
amd r9 270x
O2MD1 works not
FGRPB1G works