In order to cover the widest possible range of hardware and software, we need your help. If you go to the web site and find that your machine or compiler are not represented, please consider submitting benchmark results from your computer to the archive.
Submitting results to the archive is very easy. Just run the
benchmark program with the "-submit
" flag or choose the
analogous option in interactive mode. On a Unix system, we recommend
using make submit
instead. The program will then prompt
you for a variety of information concerning you and your computer, run
a canonical set of benchmarks, and output everything to a plain ASCII
file called "submit.txt
". When the benchmark is done,
simply email this file to benchfft@theory.lcs.mit.edu
with "SUBMIT" in the email subject line. make submit
emails the results to us automatically. (Don't worry if your email
program adds stuff to the beginning or the end of the
email...extraneous text of this sort will be ignored.)
The complete benchmark will probably take several hours to run, so be patient. You may want to let it run overnight.
Note that when you run the program with the "-submit
",
most other options are ignored. The exception is the
"-maxmem
" flag, which allows you to set the maximum
amount of RAM that the program will use. (See the "Usage" section for more information.) You
should use this option to ensure that the program knows how large a
transform it can perform. (In interactive mode, you will
automatically be prompted for this information.)