3 ms·
OK, after some effort I got the first of the MPQS programs to compile. Unfortunately it fails with a runtime error so I can't evaluate it. The program itself r
by wbhart 15y ago
OK, after some effort I got the first of the MPQS programs to compile. Unfortunately it fails with a runtime error so I can't evaluate it.
The program itself requires two files to exist before it can operate. Then it asks for the input number in base 10000 as a set of digits separated by spaces or returns. This is pretty bizarre by today's standards.
It does succeed in factoring very tiny numbers, but it seems to have bugs which prevent it from working for large numbers.
The double large prime variant of the MPQS seems to allow numbers up to a pretty small bound. This strongly indicates that it is not competitive.
I'll see if I can get anywhere with the double large prime variant version (edit: I tried the second program and it fails with the same runtime error), but given the extremely poor state of the code (no comments, full of gotos, no indication of structure, no documentation, no test suite, buggy, very poor interface, etc.) I would say this code is not going to be particularly interesting to modern day researchers.
At the same time, it is a tremendously amazing accomplishment that one man essentially toiled away in secret and created a massive number theory library like this! When I google his name I find nothing. Was he associated with an institution? Did he publish any papers? I know he didn't use the web, so that explains why he doesn't have a web presence, but it is odd not to find him referred to anywhere. Or am I just looking in the wrong place?
- daoudc 15y agoHe didn't publish any papers that I'm aware of, and he wasn't associated with an institution. I'm hoping to write a short biography at some time.
- achille 15y agoLooking through the code I see a lot of references to a group of japanese researchers. See for example this file which references "A.KUDO & Y.KIDA" and is complete with japanese comments. https://github.com/daoudclarke/wartonlegacy/blob/master/src/ubasic/REALQF4.UB https://github.com/daoudclarke/wartonlegacy/blob/master/src/... Looks like K.Aoki (Kazumaro Aoki) A.KUDO & Y.KIDA used to post to academic factorization related mailing lists see: http://www.loria.fr/~zimmerma/records/6353 http://www.loria.fr/~zimmerma/records/6353
- wbhart 15y agoKida wrote UBASIC I believe.
- daoudc 15y agoThanks for investigating this. I'd be grateful if you could make the changes you made to get it to compile available.
- wbhart 15y agoThe mpqs code actually compiles unmodified (but with many warnings) if you use f95. When the code is run it complains of a missing file. One can create an empty file of this name to keep it happy. In fact it fills the file with data. Then it complains of a second missing file, however the same trick does not work here. It complains of missing records indicating that it expects this file to contain data. I've no idea what data it should contain or in what format.
- daoudc 15y agoI have the files but I haven't uploaded them as they are very big. I imagined one of the other programs would create the files, but I've yet to discover it. I believe these files are precomputations that are meant to speed up the algorithm.
- wbhart 15y agoPerhaps they are tables of precomputed inverses of primes to speed up the trial division. There might be up to 100000 entries in the file if this is the case. I can't think what else would be in the file. I would have guessed the file was generated by the program and would contain relations from a sieving run for the last large factorisation your uncle did. But I also don't see how these files get generated by the code.