mtj@ferrari.UUCP (Michael T. Jones) (02/09/88)
I am in need of a complete set of arbitrary precision rational arithemtic routines in C. By rational, I mean numbers of the form p/q where p and q are arbitrary precision integers. It should be easy to build this as a layer above any arbitrary precision integer package, so that would be fine too. Here is what I have found so far: 1. The libmp (3x) routines from Version 7. These would be fine, but I don't have them. And could not use them if I did (Copyright and Moral restrictions). 2. "VLI-A Very Long Integer Package" These would be fine, but they are in 8080 assembler for BDS C. This is only slightly better than nothing. 3. The source for DC. We're talking ugly code here. VERY ugly. Plus the same restriction situation mentioned above. So please, if you have an arbitrary precision math package, or have any advice on locating same, please respond by email. As I might not survive a trip into the inner workings of DC. ---------------------- Michael T. Jones Tsunami Supercomputing uunet!ferrari!mtj
yuval@taux01.UUCP (Gideon Yuval) (02/09/88)
Sending mail to netlib@anl-mcs.arpa, with a line line inside it " send index from bmp ", will get you back what you need to get Brent's Multiple Precision S/W package E-mailed to you. It's all in Fortran. P.S. A line "send index" will get the top-level index sent back. -- Gideon Yuval, +972-52-522255 (work), -2-690992 (home), yuval@taux01.nsc.com