View Single Post
Old 2005-09-26, 13:19   #5
Wacky's Avatar
Jun 2003
The Texas Hill Country

108910 Posts

If you consider the contribution of computing resources and permitting one of their employees to work on the project from their office as sponsorship, then Microsoft Research Cambridge "sponsored" the project for a part of the time that Paul was employed there.

However, the same can also be said of a number of other corporations and institutions who allow their idle computers to be used.

No company, other than perhaps those of Don and myself as independent consultants, has contributed equipment or funds expressly to the NFSNET project. In each case, any contributions have been in support of a volunteer who wanted to participate.

In reality, we are presently in the third dynasty of the project. Back in the early '90s, I got interested in factoring and "signed up" to help Dr. Silverman sieve. At that time I had a 486 computer, but soon upgraded to a "state-of-the-art" Pentium.

I soon took over the administrative chore of assigning sieving ranges and collecting results so that Dr. Silverman could spend his time working on the more complex "math" part of the project. Then, to make my time spent more effective, I made some minor changes to the sieving program and wrote some scripts to automate the assignment process. Thus NFSNET was born.

Bob determined the sieving parameters. I made the assignments and collected the results and, through Peter, CWI did the post processing.
We were able to set some new records in factoring because we brought together a diverse array of computing resources. I recall a bank in Brisbane, the "build" computer at a large computer company in California, about 100 workstations at a company in Chicago, and many others. Then, as now, we are somewhat limited in participants because of the resources required for an individual machine to participate efficiently. Those which do not have adequate capability are better suited to adjunct efforts such as ECM which require far less memory.

This went on for a few years until I lost my 24hr Internet connection and had to drop out. Thus ended the first dynasty.

Conrad Curry picked up the effort. I know little about what happened then because I was not available or involved. This was the second dynasty.

In 2002, Paul was approached about trying to revive the then inactive project. Although I don't remember the details, with the encouragement of a number of "the names" in the factoring community, he and I and Don got together and reactivated NFSNET. Paul was working at MSRC and had access to a computing cluster of 16 dual processor machines. This was an invaluable resource for the matrix manipulation that follows the sieving.

Because Paul has left MSRC, we have had to cut back on the difficulty of the numbers that we are attempting since the matrix manipulation is now being done on our personal workstations rather than a high speed cluster funded by some corporation or institution.
Wacky is offline