Michael Pippig

Thanks for the hint. I will fix it at the end of this week. -------- Ursprüngliche Nachricht -------- Von: Yu Feng [email protected] Datum: An: mpip/pfft [email protected] Betreff: [pfft] Memory Layout...

PFFT already uses a dense representation of the entire volume. This is what page 46 of the manual states. Maybe, I have to explain it more clearly there. The ghost...

Looks like you forgot to divide in[l] by (n[0]*n[1]*n[2]) before you compare it with data[l]. Then, the test works on my desktop.

Right now, I added two inplace c2c checks to the test suite.

Can it be something going wrong with the bitwise operations (my_pe \* n_pes + pe) & 0xffff in line 73 of transpose-pairwise.c? They are used to generate the message tags....

If you compile FFTW on your own, the easiest way is to remove the other transpose algorithms from the solvtab list in mpi/conf.c and recompile. Otherwise, we can look at...

Did I get it right, that the problems only occur in transpose-pairwise-transposed.c?

You are right. At the moment we try to figure out how good the OpenMP support of FFTW works. We have some problems with the scaling of threaded matrix transposition...

kernel/sertrafo.c is the right place. The local data tranpositions are hidden in the calls of fftw_plan_guru64... Those are happening in plan_remap and plan_trafo. The guru interface takes some input strides...

PFFT uses FFTW for the computation of local FFTs, local memory transpositions (that are necessary to ensure the right memory order for MPI Alltoall) and global (parallel) memory transpositions. The...