Hi,
in order to perform FFT fast gsl creates some auxiliary datastructures
which depend n the size of the array. What I need to do is cache these
datastructures for later use in case they are needed to save time in
generation. One way to make this transparent is to save them in hash
table hashed by number of array elements (for the aray FFT has already
been aplied to) and then at each FFT call either fetch qux structure
from cache or generate a new one once and then save in cache. Do I have
to write hash DTS myself or it is already there in bigforth. There is
hash.fs but it seems to be not a general purpose hash DTS (bad
factoring?) but rather some tool for vocabularies (documented?). May be
the whole idea with hash is not the most efficient and you can recommend
a faster one in bigforth? I can think of a hidden vocabulary with names
for FFT DTS in it, if lookup is done by hash and not traversing the
compplete list - that is exactly what I need, but is it?
Thanks lot!
--
Sergey
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
