On Wed, Mar 23, 2011 at 3:27 PM, Charles G Montgomery < c...@physics.utoledo.edu> wrote:
> On Wednesday 23 March 2011 9:52:27 am Carsten Aulbert wrote: > > Hi > > > > we are just discussing FFTW in our institute and we hit this bug > > reported to Ubuntu which also holds true for Debian: > > > > https://bugs.launchpad.net/ubuntu/+source/fftw3/+bug/602586 > > > > Is there a specific reason why SSE is enabled for 32bit and not > > enabled for 64bit as the difference in performance can easily be a > > factor of 2 (on 64bit). > > > > As this is very often used in science, I thought people might know > > why this was chosen. > > > > Cheers > > > > Carsten > > I wonder if this might have anything to do with the bug I just submitted > for fftw3-3 > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=619324 > That seems a 64bit problem not seen for 32bit. > > regards cgm > > I have just tested this with fftw3 built explicitly with sse enabled on 64 bit. It does not fix the issue. backtrace: #0 0x0000000000401903 in dfftw_execute_ (p=0x7fffffffe460) at f77funcs.h:27 #1 0x0000000000401730 in badfftw3 () at test.f90:40 #2 0x00000000004014ba in main () #3 0x00007ffff72f1c4d in __libc_start_main (main=<value optimized out>, argc=<value optimized out>, ubp_av=<value optimized out>, init=<value optimized out>, fini=<value optimized out>, rtld_fini=<value optimized out>, stack_end=0x7fffffffe578) at libc-start.c:226 #4 0x00000000004013c9 in _start ()
