The latest (mathematical) factoring technology |
Sieving over sets of "Q-values" (i.e., very large ranges of integers) |
Task specifications are small (the integer bounds of the Q-range) |
Task results are large (collections of "partial relations") |
These results must be accumulated, sorted, and run through a neat graph algorithm before they are fed into the final bitmatrix reduction step (done on Maspar in previous factoring efforts but any reasonable machine will handle) |
Forms an ideal, embarrassingly parallel application for Web-wide coordination: very large, naturally distributed database computation which can be done with PC's up. |
Note we build on and still support factoring by email preferred by users in secure/uncommunicative environments! |