Just got one of these and I checked all of my computers and this is the only task I have seen so far.
Yup, those are kind of rare beasts :-). As the version number v0.22 indicates, the LAT search is still kind of experimental and not in full production mode.
I just checked by history, I've gotten a total of 23 LAT tasks in the last few days and completed most of them (they appear to be triggering EDF mode on arrival since I have a multi-day cache). THey also appear to be growing in size. The current tasks I have are ~250 credit, but I also have ~125 and ~50 credit tasks.
Unfortunately on my i9xx quadcores the 250 credit WUs take almost 50% longer than similar sized S6GC ones. I think this probably puts them close to what a BRP4 task would take, although I stopped running them since using CPUs to do GPU tasks is wasteful. IIRC the s5/s6 vs brp spread is narrower on AMD boxes; but I don't have any of those currently operational.
I just checked by history, I've gotten a total of 23 LAT tasks in the last few days and completed most of them (they appear to be triggering EDF mode on arrival since I have a multi-day cache). THey also appear to be growing in size. The current tasks I have are ~250 credit, but I also have ~125 and ~50 credit tasks.
Unfortunately on my i9xx quadcores the 250 credit WUs take almost 50% longer than similar sized S6GC ones. I think this probably puts them close to what a BRP4 task would take, although I stopped running them since using CPUs to do GPU tasks is wasteful. IIRC the s5/s6 vs brp spread is narrower on AMD boxes; but I don't have any of those currently operational.
I 've successfully processed a number of FGR WU's, which are now validated and credited. What I can see on all of my machines is, that the gamma ray search needs 25%-30% more CPU seconds to finish, as the GW6 search. But the credit is the same ~250 credits. I think this should be adjusted
What I can see on all of my machines is, that the gamma ray search needs 25%-30% more CPU seconds to finish, as the GW6 search. But the credit is the same ~250 credits. I think this should be adjusted
I have up to 70% more time for the same credits, so I have de-selected the gamma ray search. Let´s face it: I'm only in it for the credits. :)
What I can see on all of my machines is, that the gamma ray search needs 25%-30% more CPU seconds to finish, as the GW6 search. But the credit is the same ~250 credits. I think this should be adjusted
I have up to 70% more time for the same credits, so I have de-selected the gamma ray search. Let´s face it: I'm only in it for the credits. :)
/Ib
As Bernd stated already, the credits will be fine tuned once there is more data on the runtime of the units on different hosts.
Think about it, you are missing the chance to be one of teh few discoverers of a gamma ray pulsar (then again, you leave more units to be crunched by me ;-) ).
My first gamma-ray unit ended with a validate error I have described it in the proper thread at Problems and bug reports, My wingmate had an AMD CPU similar to mine and he had no error.He also had a Linux version, like me.
Tullio
I have finished several Gamma-Ray WUs. I don't have any problems - some are still waiting for validation the rest was validated already. The total runtime seems to vary more than for the gravitational wave WUs - somewhere between 9000 and 24000 seconds. However, the overall average seems to be the same so far. The machine is an AMD Phenom IIx4 940 running a 64Bit Linux
Gamma-ray pulsar search #1 v0.22
)
Yup, those are kind of rare beasts :-). As the version number v0.22 indicates, the LAT search is still kind of experimental and not in full production mode.
EDIT: See Bernd's latest news here: http://einsteinathome.org/node/195841&nowrap=true#113384 . So we are very close to a regular WU supply for teh Fermi search.
HBE
I got my first workunit for
)
I got my first workunit for this new application today. Hopefully, this is the start of a new trend! :)
By the way, I noticed that there are two identical Binary Radio Pulsar headings in the applications menu list. What is the difference between the two?
RE: I got my first workunit
)
The menu isn't able to distinguish between BRP4 (new) and BRP3 (old). BRP3 is winding down rapidly, but still has almost 3k work units to complete.
I just checked by history,
)
I just checked by history, I've gotten a total of 23 LAT tasks in the last few days and completed most of them (they appear to be triggering EDF mode on arrival since I have a multi-day cache). THey also appear to be growing in size. The current tasks I have are ~250 credit, but I also have ~125 and ~50 credit tasks.
Unfortunately on my i9xx quadcores the 250 credit WUs take almost 50% longer than similar sized S6GC ones. I think this probably puts them close to what a BRP4 task would take, although I stopped running them since using CPUs to do GPU tasks is wasteful. IIRC the s5/s6 vs brp spread is narrower on AMD boxes; but I don't have any of those currently operational.
Got my first gamma-ray search
)
Got my first gamma-ray search after a string of SETI units. The rest of BOINC seems to be on vacation, at least in my subset.
Tullio
RE: I just checked by
)
I 've successfully processed a number of FGR WU's, which are now validated and credited. What I can see on all of my machines is, that the gamma ray search needs 25%-30% more CPU seconds to finish, as the GW6 search. But the credit is the same ~250 credits. I think this should be adjusted
RE: RE: What I can see
)
RE: RE: RE: What I can
)
My first gamma-ray unit ended
)
My first gamma-ray unit ended with a validate error I have described it in the proper thread at Problems and bug reports, My wingmate had an AMD CPU similar to mine and he had no error.He also had a Linux version, like me.
Tullio
I have finished several
)
I have finished several Gamma-Ray WUs. I don't have any problems - some are still waiting for validation the rest was validated already. The total runtime seems to vary more than for the gravitational wave WUs - somewhere between 9000 and 24000 seconds. However, the overall average seems to be the same so far. The machine is an AMD Phenom IIx4 940 running a 64Bit Linux
mickydl*