They haven’t added any new FGRPB1G work. The total has been steady at 142,830,632 for many months now.
The days remaining estimate is derived from the 5-day average WU completed per day metric. That 5-day average has been dropping since it looks like the “Atlas Condor Jobs” super computer has tapered off or shut off (Their average has been dropping hard and they were previously doing about 150M/day). The rate in which the average is dropping is in line with keeping the estimated days remaining pretty steady. it will level off and the days remaining will start dropping with real time again provided that there isn’t another large shift in computational power for the project.
Don’t rely on the project estimated flops either, since it’s calculated in an incorrect way, based on credit reward instead of reported or calculated flops.
The Atlas at AEI/MPG was briefly the fastest thing on the planet. E@H work units are used to 'burn-in' new nodes. I walked around inside it on an open day last decade. It is literally like the inside of an air conditioning installation that also happens to compute. I described my observations in Hannover Diary, but sadly the photo links there were mangled when we went to Drupal administration of these boards. From memory 'Condor' referred to GPU arrays, an emerging new joy back then.
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
The photos won't be lost, just which photo refers to what comment. I'm not sure I can edit that far back to fix any links, even if I could re-thread the thread. I may look into that if I have the time.
Yeah, E@H is not Atlas' primary purpose. It gets used by the scientific consortium that paid for it and they generously use E@H work to test their new hardware. So many of the temporary hosts in our statistics are/were their compute nodes. E@H is the DC project based on SETI started by Bruce Allen who was director of AEI at one point.
I fondly remember the delightful welcome I got from Bruce who then introduced me to Carsten Aulbert & Henning Ferhmann ( the Atlas sysadmins ) and a host of others there. I hope to go back some day to visit ..... wars, politics and plagues permitting. ;-)
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
Still no sign of BRP on AMD
)
Still no sign of BRP on AMD GPU. Is it coming?
And why isn't the 8.3 days of Gamma remaining going down? Are you feeding in new work?
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
The days remaining estimate
)
They haven’t added any new FGRPB1G work. The total has been steady at 142,830,632 for many months now.
The days remaining estimate is derived from the 5-day average WU completed per day metric. That 5-day average has been dropping since it looks like the “Atlas Condor Jobs” super computer has tapered off or shut off (Their average has been dropping hard and they were previously doing about 150M/day). The rate in which the average is dropping is in line with keeping the estimated days remaining pretty steady. it will level off and the days remaining will start dropping with real time again provided that there isn’t another large shift in computational power for the project.
Don’t rely on the project estimated flops either, since it’s calculated in an incorrect way, based on credit reward instead of reported or calculated flops.
_________________________________________________________________________
Oh well, I'll keep on gamma
)
Oh well, I'll keep on gamma crunching and get it finished off. I would like to see the BRP on my AMDs though.
What is this Atlas Condor Jobs Supercomputer? The only Atlases I know of are a supercomputer in the late 70s and the LHC project.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
Einstein@home has had at
)
Einstein@home has had at their disposal the use of the Atlas cluster supercomputer for years.
https://www.aei.mpg.de/189177/the-launch-of-atlas
Has it been taken away to do
)
Has it been taken away to do something else?
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
The Atlas at AEI/MPG was
)
The Atlas at AEI/MPG was briefly the fastest thing on the planet. E@H work units are used to 'burn-in' new nodes. I walked around inside it on an open day last decade. It is literally like the inside of an air conditioning installation that also happens to compute. I described my observations in Hannover Diary, but sadly the photo links there were mangled when we went to Drupal administration of these boards. From memory 'Condor' referred to GPU arrays, an emerging new joy back then.
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
Interesting read, pity I
)
Interesting read, pity I can't see the photos, are they elsewhere?
So Einstein only gets use of it when they upgrade it and test new bits?
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
The photos won't be lost,
)
The photos won't be lost, just which photo refers to what comment. I'm not sure I can edit that far back to fix any links, even if I could re-thread the thread. I may look into that if I have the time.
Yeah, E@H is not Atlas' primary purpose. It gets used by the scientific consortium that paid for it and they generously use E@H work to test their new hardware. So many of the temporary hosts in our statistics are/were their compute nodes. E@H is the DC project based on SETI started by Bruce Allen who was director of AEI at one point.
I fondly remember the delightful welcome I got from Bruce who then introduced me to Carsten Aulbert & Henning Ferhmann ( the Atlas sysadmins ) and a host of others there. I hope to go back some day to visit ..... wars, politics and plagues permitting. ;-)
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
If I can see the photos in
)
If I can see the photos in the wrong order that would still be cool.
I guess we need to crank up our own GPUs for now then.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
There's more info
)
There's more info here, it has a Megawatt UPS backup !
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