Meantime: 5 fewer mults and 2 fewer adds. Not much, but ...
Tried to convert FPU-Code to SSE2, gives 'about' the same value, but would never ever validate. Seems to be slower too :-(
My codes also have this part in an optimised FPU code, but I didn't get any improvement when I tried to combine it with the SSE part. But the SSE2 code was a bit faster in S37a, if you remember for that. The other codes use the same optimised routine for this.
As I said 12 days ago in the S41.xx Observation Thread : "No one here ever gonna forget Akosf I guess."
It's nice to see this development. Good for you Akos.
My English is to bad otherwise I would try to make some funny comment over a transfer to Seti but I'm afraid to be mistaken on that.:-)
Akos, congratulations for being hired by Einstein team. I am happy that you have accepted this invitation; it is very positive sign for all of us.
And I generally highly appreciate this progress, it has definitely increased value of this project in my eyes (which already had had been very high).
Not only exciting project from the scientific point of view, but it's also managed by great staff now also hiring very skillful optimizators! A perfect combination.
It has been nearly a year since Akos was last heard from on these message boards. Before that he did some really good work for the project. Does anyone know what became of him?
Verloren ist nur, wer sich selbst aufgibt. - Hans-Ulrich Rudel
RE: Meantime: 5 fewer mults
)
My codes also have this part in an optimised FPU code, but I didn't get any improvement when I tried to combine it with the SSE part. But the SSE2 code was a bit faster in S37a, if you remember for that. The other codes use the same optimised routine for this.
As I said 12 days ago in the
)
As I said 12 days ago in the S41.xx Observation Thread : "No one here ever gonna forget Akosf I guess."
It's nice to see this development. Good for you Akos.
My English is to bad otherwise I would try to make some funny comment over a transfer to Seti but I'm afraid to be mistaken on that.:-)
RE: My English is to bad
)
May I suggest Rosetta instead? ;-)? The science is more important there.
Greetings groundhog
Akos, congratulations for
)
Akos, congratulations for being hired by Einstein team. I am happy that you have accepted this invitation; it is very positive sign for all of us.
And I generally highly appreciate this progress, it has definitely increased value of this project in my eyes (which already had had been very high).
Not only exciting project from the scientific point of view, but it's also managed by great staff now also hiring very skillful optimizators! A perfect combination.
Thanks to all, and Akos - good luck! :-)
It has been nearly a year
)
It has been nearly a year since Akos was last heard from on these message boards. Before that he did some really good work for the project. Does anyone know what became of him?
Verloren ist nur, wer sich selbst aufgibt. - Hans-Ulrich Rudel
His last contact with boinc
)
His last contact with boinc was this year in the middle of april...