I bet your computer sees nothing whatsoever off the credits claimed.
You were correct. The third computer's results came in and my results were rejected. That was 17.5 hours down the drain. So, if there is a conflict between Windows and Mac OS X, the OS on the third computer determines which result is accepted, just like for Linux and Unix vs. Windows. And given that Macs are a much smaller proportion of the overall pool of crunchers, it's likely that my results will be rejected in almost every case. What a system. I can only hope that the E@H team will correct this bug.
This should be interesting to those engaged in the current bug discussion. I've just gotten credit for a 421.25 cobblestone WU (WU ID 33659845). My wingman did not reply in time and the WU was sent out to a third computer. Now, my wingman has replied AND been given credit for the WU. So, where does that leave the third cruncher? I'd assume they are out of luck.
This should be interesting to those engaged in the current bug discussion. I've just gotten credit for a 421.25 cobblestone WU (WU ID 33659845). My wingman did not reply in time and the WU was sent out to a third computer. Now, my wingman has replied AND been given credit for the WU. So, where does that leave the third cruncher? I'd assume they are out of luck.
as long as the 3rd cruncher returns in time and validates he'll get credit because he didn't do anything wrong.
in which both me and my wingman have a pentium with XP and 4.17 client version.
A third was sent to a pentium/W2K, who did not reply. Now a fourth was sent to another pentium/XP.
Can it be a precision problem, as a 32bit computer computes differently from a 64bit computer?
in which both me and my wingman have a pentium with XP and 4.17 client version.
A third was sent to a pentium/W2K, who did not reply. Now a fourth was sent to another pentium/XP.
Can it be a precision problem, as a 32bit computer computes differently from a 64bit computer?
Always wondered if it`s true...
No, the "32bit" and "64bit" refer to different instruction sets that are supported, the second features wider registers and the ability to address a larger address space. E@H is currently only available as a 32 bit application and if run on a 64bit capable CPU will only use the "32bit" instruction set.
Even a native 64bit E@H app would not make a difference.
The "32bit" CPUs are capable of using floating point numbers that are 64 bit wide and use up to 80 bit wide floating points internally.
Still though, it is rather odd that two very similar hosts failed to validate.
Lets see what happens when the third (fourth) host returns its result.
There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman
It looks like there are two WUs were started (26 and 27 May)
and then a third was started 13 later (10 Jun).
Now, the first two have, apparently validated OK.
The third will take a another day or so to complete...
It looked unusual that the 3rd was started one day before the completion - or
deadline(?) - of the second WU...
It looks like there are two WUs were started (26 and 27 May)
and then a third was started 13 later (10 Jun).
Now, the first two have, apparently validated OK.
The third will take a another day or so to complete...
It looked unusual that the 3rd was started one day before the completion - or
deadline(?) - of the second WU...
Any ideas??
Thanks,
Jay E.
That's exactly it, the 2nd host missed the deadline of the 10th and a third WU was sent out. The 2nd host returned its result a day late but was still granted credit (a nice feature of the validater). When the 3rd host returns its result it too will be granted credit.
The quorum is still 2...
There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman
That's exactly it, the 2nd host missed the deadline of the 10th and a third WU was sent out. The 2nd host returned its result a day late but was still granted credit (a nice feature of the validater). When the 3rd host returns its result it too will be granted credit.
The quorum is still 2...
Ah-HAh.....
Thanks for the info,
Jay
(Orlando, Florida)
RE: I bet your computer
)
Most likely. I think I've lost no more than maybe 10 % of claimned credit to this issue. YMMV.
This is by no means an Einstein@Home specific thing, e.g. over at SETI@HOME zthere's a dedicated thread for validation problems: http://setiathome.berkeley.edu/forum_thread.php?id=39588.
CU
BRM
RE: I bet your computer
)
You were correct. The third computer's results came in and my results were rejected. That was 17.5 hours down the drain. So, if there is a conflict between Windows and Mac OS X, the OS on the third computer determines which result is accepted, just like for Linux and Unix vs. Windows. And given that Macs are a much smaller proportion of the overall pool of crunchers, it's likely that my results will be rejected in almost every case. What a system. I can only hope that the E@H team will correct this bug.
This should be interesting to
)
This should be interesting to those engaged in the current bug discussion. I've just gotten credit for a 421.25 cobblestone WU (WU ID 33659845). My wingman did not reply in time and the WU was sent out to a third computer. Now, my wingman has replied AND been given credit for the WU. So, where does that leave the third cruncher? I'd assume they are out of luck.
RE: This should be
)
as long as the 3rd cruncher returns in time and validates he'll get credit because he didn't do anything wrong.
This is NOT a problem of
)
This is NOT a problem of AMD/LINUX PENTIUM/XP MAC/OsX disagreement.
I have a workunit waiting to get validated,
http://einsteinathome.org/workunit/33732455
in which both me and my wingman have a pentium with XP and 4.17 client version.
A third was sent to a pentium/W2K, who did not reply. Now a fourth was sent to another pentium/XP.
Can it be a precision problem, as a 32bit computer computes differently from a 64bit computer?
Always wondered if it`s true...
RE: This is NOT a problem
)
No, the "32bit" and "64bit" refer to different instruction sets that are supported, the second features wider registers and the ability to address a larger address space. E@H is currently only available as a 32 bit application and if run on a 64bit capable CPU will only use the "32bit" instruction set.
Even a native 64bit E@H app would not make a difference.
The "32bit" CPUs are capable of using floating point numbers that are 64 bit wide and use up to 80 bit wide floating points internally.
CU
BRM
Still though, it is rather
)
Still though, it is rather odd that two very similar hosts failed to validate.
Lets see what happens when the third (fourth) host returns its result.
There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman
Greetings, Has the quorum
)
Greetings,
Has the quorum changed from two to three??
(This looked similar to your original post on the subject.)
On 10 June, I picked up a Work Unit that now has an unusual status.
http://einsteinathome.org/workunit/33822683
It looks like there are two WUs were started (26 and 27 May)
and then a third was started 13 later (10 Jun).
Now, the first two have, apparently validated OK.
The third will take a another day or so to complete...
It looked unusual that the 3rd was started one day before the completion - or
deadline(?) - of the second WU...
Any ideas??
Thanks,
Jay E.
RE: Greetings, Has the
)
That's exactly it, the 2nd host missed the deadline of the 10th and a third WU was sent out. The 2nd host returned its result a day late but was still granted credit (a nice feature of the validater). When the 3rd host returns its result it too will be granted credit.
The quorum is still 2...
There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman
RE: That's exactly it, the
)
Ah-HAh.....
Thanks for the info,
Jay
(Orlando, Florida)