Hi,

[email protected] wrote:
>Not bad for early results.

see http://www.fftw.org/cell/ps3/dyn94171207.austin.ibm.com.1d.drxx.p2.png
unfotunaly fftw3-r2r is too slow.

complex fftw version.
http://netnews.gotdns.org/WallStreet/6351/gfn/lucas.ps3/MacLucasFFTW.ps3.1.patch

ps3$ time ./MacLucasFFTW 216091
M( 216091 )P, n = 12288, MacLucasFFTW v8.1  Ballester

real    5m19.948s
user    5m16.405s
sys     0m3.412s

ps3$ time ./MacLucasFFTW 1257787
M( 1257787 )P, n = 73728, MacLucasFFTW v8.1  Ballester

real    226m15.787s
user    223m18.505s
sys     2m54.711s

Real Discrete Fourier Transform version.(oour rfdt->complex fftw)
http://netnews.gotdns.org/WallStreet/6351/gfn/lucas.ps3/MacLucasFFTW.ps3.2.patch
ps3$ time ./MacLucasFFTW 216091
M( 216091 )P, n = 16384, MacLucasFFTW v8.1  Ballester

real    10m6.359s
user    9m57.052s
sys     0m9.179s

 .. .  .   .    .     .       .         .          .           .
 Shoichiro Yamada [email protected]
_______________________________________________
Prime mailing list
[email protected]
http://hogranch.com/mailman/listinfo/prime

Reply via email to