0
I Use This!
High Activity

Commits : Listings

Analyzed 2 days ago. based on code collected 2 days ago.
Dec 09, 2018 — Dec 09, 2019
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Intermediate step in cleaning procedure Anon32 More... almost 12 years ago
Use rwmax to distinguish between runs Anon32 More... almost 12 years ago
Miscellaneous updates; files sparse.{c,h} appear, to prepare for more code sharing + cleaning w.r.t. sparse matrices. Anon32 More... almost 12 years ago
stupid bug. Anon32 More... almost 12 years ago
more timings. New strategy: small merge levels asaa. Anon32 More... almost 12 years ago
a little more verbose. Added prp tests on factors. Anon32 More... almost 12 years ago
More w(M) use and now n*w(M) is printed, here and there. It has to decrease, I guess. Anon32 More... almost 12 years ago
bw should now compile without a patched gmp Anon32 More... almost 12 years ago
just a better print. Anon32 More... almost 12 years ago
mat->weight is back. Ready for the next step. Anon32 More... almost 12 years ago
Stupid bug corrected. Anon32 More... almost 12 years ago
Use -std=c99 instead of #define __USE_ISOC99 __USE_ISOC99 is an internal glibc name, not meant to be exposed. _ISOC99_SOURCE is more recommended, although still glibc specific. So -std=c99 is preferred. Anon32 More... almost 12 years ago
polyselect.c: removed warnings Makefile: added options -W -Wall Anon32 More... almost 12 years ago
long_poly.c: fixed bug in long_poly_sqr when input is zero utils.h: added prototype for nbits() Makefile: added options -W -Wall Anon32 More... almost 12 years ago
added missing header Anon32 More... almost 12 years ago
added missing include fixed fprint_relation() [apparently not used currently] Anon32 More... almost 12 years ago
less pignoufing is required. More to come. Anon32 More... almost 12 years ago
The square bug patched in new function find_prime_factor. Hope this will not cause problem later on in factoring, that admitting p1 = p2 in the answer. Anon32 More... almost 12 years ago
added some auto-generated files, they will disappear soon anyway. Anon32 More... almost 12 years ago
A lot of cleaning done including simplification of types, variables, etc. Plus the use of rwmax to prevent some dense rows to be used. Clever strategies must be tried now. Anon32 More... almost 12 years ago
using rwmax better. Anon32 More... almost 12 years ago
Take paramater order change in purge.c into account + prepare for the use of rwmax. Anon32 More... almost 12 years ago
Now, light columns can enter the square dance if needed. More to come. Anon32 More... almost 12 years ago
A little bit of cleaning and adapting to multifile input, easier to play with. Anon32 More... almost 12 years ago
The even smarter version. Once I get bigger relation files, I will look at the next smarter one... Anon32 More... about 12 years ago
Cleaned version, expecting the next one to be even smarter. Anon32 More... about 12 years ago
Prepare for multiple jobs on the same data and/or machine. Anon32 More... about 12 years ago
Two major novelties: 1. a faster merge, using intricate data structures reminiscent of that used in graph theory. More in tex one of these days. I hope this will resist gigabytes of data. For the time being, this is ok for c80, where the running time is spent elsewhere. The "slow" merge is still around. Probably, things have to be cleaned and simplified a bit and for huge data, some choices will have to be made. 2. After all, minimal spanning trees can be used. They are more tricky to use correctly that simply pivoting. The current implementation might not be optimal, but seems to work. Prim's algo is easier for our purpose I think. As a result, the coding in name.merge.his evolved, as well as replay.c for obvious compatibility reasons. Anon32 More... about 12 years ago
ndepmin and ndepmax are in the same boat Anon32 More... about 12 years ago
typos + new param in allsqrt.c Anon32 More... about 12 years ago