C41 also has a special (very fast) rounding method.
I developed it to prepare the code for X41.
Don't hesitate if you have any questions about win-x86 optimised applications.
Thanks for thinking of the old boxes. Many of us still use them,
and it is nice to see that they can still contribute to science.
Is X41 a mixture of SSE3 and 3DNOW?
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
Thanks for thinking of the old boxes. Many of us still use them,
and it is nice to see that they can still contribute to science.
Yes, I know. They are also important members of the EAH community.
I'm sorrow that I cannot help for some other members too. (Linux, Macs, etc.)
Quote:
Is X41 a mixture of SSE3 and 3DNOW?
Probably it won't use SSE3 instructions, because SSE2 + 3DNow! combination seems to be enough.
Much more difficult to combine SSE2 + 3DNow! than FPU + SSE (S38-S41.xx).
That is more then a 10 minute per WU decrease in time.
Thanks for putting in the time to optimize for these older rigs.
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
C41.xx Observation Thread
)
Thanks for thinking of the old boxes. Many of us still use them,
and it is nice to see that they can still contribute to science.
Is X41 a mixture of SSE3 and 3DNOW?
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: Thanks for thinking of
)
Yes, I know. They are also important members of the EAH community.
I'm sorrow that I cannot help for some other members too. (Linux, Macs, etc.)
Probably it won't use SSE3 instructions, because SSE2 + 3DNow! combination seems to be enough.
Much more difficult to combine SSE2 + 3DNow! than FPU + SSE (S38-S41.xx).
PII 447 MHz: Short
)
PII 447 MHz:
Short WUs
C41.01 : 9650s avg.
C41.01/.02 : 9089s
C41.02 : 9000s avg.
That is more then a 10 minute per WU decrease in time.
Thanks for putting in the time to optimize for these older rigs.
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
With C41.02 my Celeron 300 @
)
With C41.02 my Celeron 300 @ 463Mhz does a WU in time comparable to P4 2.6Ghz running stock client. :D
Akos, you do magic!
Many thanks,