Mail Archives: djgpp/1997/02/05/00:38:33
| From: | jasonp AT Glue DOT umd DOT edu (Jason Stratos Papadopoulos) | 
| Newsgroups: | comp.os.msdos.djgpp | 
| Subject: | need help handling enormous size arrays | 
| Date: | 5 Feb 1997 04:09:50 GMT | 
| Organization: | University of Maryland, College Park | 
| Lines: | 18 | 
| Message-ID: | <5d916e$c1e@hecate.umd.edu> | 
| NNTP-Posting-Host: | taylor.isr.umd.edu | 
| To: | djgpp AT delorie DOT com | 
| DJ-Gateway: | from newsgroup comp.os.msdos.djgpp | 
Hello. I'm writing a homebrew arbitrary precision arithmetic package,
and was coding up a multiply that uses fast Fourier transforms. Everything
is nice and fast up to array (of doubles) size 2^14, but from 2^15 on
the computation time hits a brick wall! Before this point (2^10,11,12,etc)
a given power of 2 only takes a bit more than twice as long as the one
before. The program does no memory management, and I need to work with
array sizes perhaps as large as 2^18 or 2^19. 
Am I not doing somethng vital? My hard drive isn't thrashing so it's
not a virtual memory problem; For the record, I have 16 megs of ram
and about 20 or 30 megs of free HD space; I'm using a Win95 DOS box
for dpmi.
Please don't flame me for missing the obvious...I'm an amateur at
high-powered scientific computing.
Thanks in advance,
jasonp
- Raw text -