just a question, Does the ps3 come with an os.
If so which one. Could be a nice addition to the farm.
I do not know if the FPUs in the PS3's CPU are accurate enough to allow Einstein@home to run properly. They are optimized for fast game play and multimedia applications and are not IEEE 754 compliant. The IEEE 754 floating point mathematics standard was designed so that accuracy took precedence over performance whenever possible. This led to slow and complicated FPU designs needed for compliance with the standards. Games and multimedia applications need speed, not accuracy. They also prefer that when an overflow (when the result is too large in magnitude to represent in the chosesn floating point format) or underflow (when the result is so small that the result would round to zero even though the true result is not zero) occur that the closest value is used instead of generating an interrupt to handle it to have the program decide on how to handle the situation.
The reason that the IEEE 754 standard results in slow but accurate FPUs is that the scientific, mathematics, and accounting communities obviously will reject a standard that trades off accuracy for speed. Some scientists need as much accuracy as possible to avoid getting their work rejected or to avoid bad results. Mathematicians do not want to publish papers only to get embarassed when the math is shown to be suspect. Honest accountants need good math (unlike the ones who worked for Enron, WorldCom, or Parmalat) because they do not want their work shot down.
just a question, Does the ps3 come with an os.
If so which one. Could be a nice addition to the farm.
I do not know if the FPUs in the PS3's CPU are accurate enough to allow Einstein@home to run properly. They are optimized for fast game play and multimedia applications and are not IEEE 754 compliant. The IEEE 754 floating point mathematics standard was designed so that accuracy took precedence over performance whenever possible. ...
Thanks for this valuable word of caution.
I guess the real test will be suck it and see. If the results pass validation against windows and other flavours of linus than the box will be accepted for BOINC - this may mean it is OK on some projects but not on others.
I am going to leave it to some of the game players to try it first - those who want the box anyway for its games have nothing to lose by trying it on BOINC.
Myself I'd be gutted to spend all that money and not have a box that was good for crunching... I might actually have to start gaming ;-)
Assuming the OS recognizes it as 7 CPUs BOINC would run 7 workunits at a time if your preferences allowed it.
And quota.
Up to four cpus your quota increases pro rata - above 4 cpus it stays at 4x the standard quota spread over you whole machine. On some projects that would not matter, but here it might (the fastest boxes already run out on the shortest ALberts).
Be another incentive to go crunching multiple projects.
It seems like Linux support won't be a problem for the PS3 and IBM is working on the compilers (they have greater plans with Cell than just gaming). Coding an app for Cell won't be easy, but I'm sure BOINC could just launch 7 clients on a PS3, as long as the 512MB RAM are enough.
So thanks for bringing the fp-precision up as an issue. I didn't think of that before. I don't know how fp is implemented in the Cell and how severe this limitation is, but I guess no one here can judge that now?
While we're at it: I think the fp-precision is a huge problem for GP-GPU as well.
Playstation 3 as crunching machine
)
This Link provide you with the info.
Regards
"The FUTURE is only a PARTICLE away from the PRESENT and the PAST."
RE: just a question, Does
)
I do not know if the FPUs in the PS3's CPU are accurate enough to allow Einstein@home to run properly. They are optimized for fast game play and multimedia applications and are not IEEE 754 compliant. The IEEE 754 floating point mathematics standard was designed so that accuracy took precedence over performance whenever possible. This led to slow and complicated FPU designs needed for compliance with the standards. Games and multimedia applications need speed, not accuracy. They also prefer that when an overflow (when the result is too large in magnitude to represent in the chosesn floating point format) or underflow (when the result is so small that the result would round to zero even though the true result is not zero) occur that the closest value is used instead of generating an interrupt to handle it to have the program decide on how to handle the situation.
The reason that the IEEE 754 standard results in slow but accurate FPUs is that the scientific, mathematics, and accounting communities obviously will reject a standard that trades off accuracy for speed. Some scientists need as much accuracy as possible to avoid getting their work rejected or to avoid bad results. Mathematicians do not want to publish papers only to get embarassed when the math is shown to be suspect. Honest accountants need good math (unlike the ones who worked for Enron, WorldCom, or Parmalat) because they do not want their work shot down.
Aren't those the same Cell
)
Aren't those the same Cell CPUs that are mentioned here (I might be wrong though, haven't read much about ps3) :
http://techon.nikkeibp.co.jp/english/NEWS_EN/20050525/105050/ ?
If they are good for a server, DC applications should be happy with them too.
I think that it is the same
)
I think that it is the same processor.
But WOW, 200GFLOPS at 3 GHZ. Now were playing with power.
Sounds like the ps3 will be a beaut cruncher with a lunix os.
RE: RE: just a question,
)
Thanks for this valuable word of caution.
I guess the real test will be suck it and see. If the results pass validation against windows and other flavours of linus than the box will be accepted for BOINC - this may mean it is OK on some projects but not on others.
I am going to leave it to some of the game players to try it first - those who want the box anyway for its games have nothing to lose by trying it on BOINC.
Myself I'd be gutted to spend all that money and not have a box that was good for crunching... I might actually have to start gaming ;-)
River~~
~~gravywavy
The question is, is BOINC
)
The question is, is BOINC threaded to use 7 CPUs?
Or would you just run 7 WU?
Or does linux even support the cell architecture w/o too much modification?
RE: The question is, is
)
Supposedly with the newest kernel. There are references in the changelog.
One entry points to a linuxoncell project.
Michael
Team Linux Users Everywhere
RE: The question is, is
)
Assuming the OS recognizes it as 7 CPUs BOINC would run 7 workunits at a time if your preferences allowed it.
BOINC WIKI
BOINCing since 2002/12/8
RE: Assuming the OS
)
And quota.
Up to four cpus your quota increases pro rata - above 4 cpus it stays at 4x the standard quota spread over you whole machine. On some projects that would not matter, but here it might (the fastest boxes already run out on the shortest ALberts).
Be another incentive to go crunching multiple projects.
~~gravywavy
It seems like Linux support
)
It seems like Linux support won't be a problem for the PS3 and IBM is working on the compilers (they have greater plans with Cell than just gaming). Coding an app for Cell won't be easy, but I'm sure BOINC could just launch 7 clients on a PS3, as long as the 512MB RAM are enough.
So thanks for bringing the fp-precision up as an issue. I didn't think of that before. I don't know how fp is implemented in the Cell and how severe this limitation is, but I guess no one here can judge that now?
While we're at it: I think the fp-precision is a huge problem for GP-GPU as well.
MrS
Scanning for our furry friends since Jan 2002