I've run out of Cuda Wu's at Seti and Seti Beta, so i've got my GTX460 doing BRP4cuda32nv301 Wu's now, another day and i'll be upgrading Boinc and starting the HD7770 too,
I apologize, I didn't mean it to be rude or directed to any specific person... Poor wording. Just wanted to say that fortunately no massive seti-like caches are needed here as far as I've observed.
Also, because this project still relies on DCF (duration correction factor) on local boinc client, it sometimes can overfetch when running CPU+GPU tasks, so small cache is the best.
I allready post about the cache diferences in our private forum, so i belive that situation will not replicate.
Still beliving E@H only needs something like a 0.5 day cache, the WU are allways avaiable for DL, even for the big hungry hosts most of we have.
Four the hosts is easy to handle Smaller caches and faster to see any problem.
Remember years of SETI cruch make us need an trust only in large caches, is dificult to us to acomodate to that new situation.
Don´t know the avaiable band the E@H have, but it must be impressive. I check and the actual DL speed here is >1MBps (the maximum capacity allowed for this host by my gigabyte router configuration, so i´m sure must be a lot more) and i´m in Brazil far from the lab. Just for curiosity anybody knows it´s real capacity?
Don´t know the avaiable band the E@H have, but it must be impressive. I check and the actual DL speed here is >1MBps (the maximum capacity allowed for this host by my gigabyte router configuration, so i´m sure must be a lot more) and i´m in Brazil far from the lab. Just for curiosity anybody knows it´s real capacity?
The BRP search is served from Hannover, Germany, using the internet connection of the local university, which I think is 10GB/s. The technical bottleneck, however, is what the two download servers could ship out - at the time of writing, they both do 20MB/s each, I saw them doing >30MB/s w/o problems. Not yet sure where the limitations are.
FWIW the GW work uses more constant data files, which are served by a handful of download servers hosted by various institutions distributed all over the world (US Wisconsin (UWM), US California (CalTech), GB (Glasgow), Germany (AEI)). Their capacities and connectivity varies, but the one in Hannover serving most of Europe and Asia is rather bored (5MB/s).
RE: Argh,
)
It was an oversight on his part, not intentional.
Newcomer to the Einstein project.
soft spirit wrote:well it
)
Yes indeed--GPU users group has had an impressive onset here.
The image here is a live link to BOINCstats, so should update daily.
I've run out of Cuda Wu's at
)
I've run out of Cuda Wu's at Seti and Seti Beta, so i've got my GTX460 doing BRP4cuda32nv301 Wu's now, another day and i'll be upgrading Boinc and starting the HD7770 too,
Claggy
You are in our team.....
)
You are in our team..... more firepower for our fleet!
Thanks for the warm welcome
)
Thanks for the warm welcome Khangollo.
I apologize, I didn't mean it
)
I apologize, I didn't mean it to be rude or directed to any specific person... Poor wording. Just wanted to say that fortunately no massive seti-like caches are needed here as far as I've observed.
Also, because this project still relies on DCF (duration correction factor) on local boinc client, it sometimes can overfetch when running CPU+GPU tasks, so small cache is the best.
I allready post about the
)
I allready post about the cache diferences in our private forum, so i belive that situation will not replicate.
Still beliving E@H only needs something like a 0.5 day cache, the WU are allways avaiable for DL, even for the big hungry hosts most of we have.
Four the hosts is easy to handle Smaller caches and faster to see any problem.
Remember years of SETI cruch make us need an trust only in large caches, is dificult to us to acomodate to that new situation.
Don´t know the avaiable band the E@H have, but it must be impressive. I check and the actual DL speed here is >1MBps (the maximum capacity allowed for this host by my gigabyte router configuration, so i´m sure must be a lot more) and i´m in Brazil far from the lab. Just for curiosity anybody knows it´s real capacity?
RE: soft spirit wrote:well
)
We allways try to do our best... Thats why i ask about hungry hosts (look the first post on this thread)
Congrats team for the 1.5MM mark. But we could do better!
RE: Don´t know the
)
The BRP search is served from Hannover, Germany, using the internet connection of the local university, which I think is 10GB/s. The technical bottleneck, however, is what the two download servers could ship out - at the time of writing, they both do 20MB/s each, I saw them doing >30MB/s w/o problems. Not yet sure where the limitations are.
FWIW the GW work uses more constant data files, which are served by a handful of download servers hosted by various institutions distributed all over the world (US Wisconsin (UWM), US California (CalTech), GB (Glasgow), Germany (AEI)). Their capacities and connectivity varies, but the one in Hannover serving most of Europe and Asia is rather bored (5MB/s).
BM
BM
Now we do our best, reaches
)
Now we do our best, reaches close to the 2MM mark in one day
http://stats.free-dc.org/stats.php?page=team&proj=eah&team=10217&sort=today#
Congrats team.