NumberFields@home

[odicin]

http://numberfields.asu.edu/NumberFields/

Team

Wiki

NumberFields@home is a research project that uses Internet-connected computers to do research in number theory. You can participate by downloading and running a free program on your computer.

NumberFields@home searches for fields with special properties. The primary application of this research is in the realm of algebraic number theory. Number theorists can mine the data for interesting patterns to help them formulate conjectures about number fields. Ultimately, this research will lead to a deeper understanding of the profound properties of numbers, the basic building blocks of all mathematics.

NumberFields@home is based at the school of mathematics at Arizona State University. The final results of this project will be complete tables of number fields. The results are given in table form or as a searchable database.

[odicin]

Leider hier das gleiche Problem wie bei Correlizer. Der Fortschritt der WU wird im BOINCManager nicht angezeigt. Das ist schon ärgerlich, wenn man mal eine längere WU mit 12h erwischt. BoincTasks hat das Problem nicht und zeigt den Progress an.

Die Credits sind auch noch etwas mau, eine 12h WU brachte nur 160 Credits.

Gruß Odi

[odicin]

Technically speaking, the project could go on forever, as our searches depend on field degree and ramified primes. Right now we are concentrating on degree 10 fields. After that, we could move on to higher degrees. We also sort fields by ramified primes. So for each set of primes, the search will be finite, but there are an unlimited combination of primes.

To give you an idea of the time required for a specific case, part of my dissertation found all degree 10 fields ramified over the set {2,3}, and this would have taken about 5 years on a single cpu (it took 60 days using boinc, but the total run time summed over all hosts was 4.7 years).

Hier noch ein paar Infos zum Projekt vom Admin.

Gruß Odi

[DoctorNow]

24 Mar 2018, 6:21:03 UTC New application coming soon
In the coming days we will be introducing a new number field application. This one will target degree 7 fields (called septics). The current best bound is 5E6, and this app will increase that to 200E6. This means it will find all degree 7 fields whose discriminant is bounded by 200E6.

Here is some additional information:
1. This search is expected to run for only 5 to 6 months (but could take longer).
2. It will run side-by-side with the current decic app.
3. It will only support the 64bit Windows and 64 bit Linux platforms.

Eine neue App ist in Vorbereitung für eine neue Suche, die ca. 5-6 Monate (oder auch länger) dauern soll. Neben der herkömmlichen App wird sie nur für Windows und Linux in 64-Bit verfügbar sein.
Erste Test-WUs gibt es bereits.

[odicin]

Seit November letzten Jahres gibt es beim Projekt auch eine App für Nvidia und für AMD-Karten (Beta). Ist mir bislang noch gar nicht aufgefallen ;)

Gruß Odi

[odicin]

My old thesis advisor who has been backing this project retired in May. I don't actually work for the university, so this means the project has lost is primary benefactor.

The good news is that the university has allowed the project to continue. The bad news is that they will not be supporting any more hardware upgrades. This means when the current hardware fails the project will come to an end.

I'm hoping the current hardware can last several more years. One of the most critical parts is the hard drive, and we just got a new one about 1.5 years ago. Given that the previous hard drive lasted over 4 years, this is a positive.

This news has played into my reasoning on the order of the searches. Given that sf3 is practically guaranteed to be complete, I though it was prudent to use our finite time to first concentrate on sf7. The short term goal is to get sf7 to the same state that sf3 is currently in, which should take us into early next year. At that point, we can either spend the time to finish off sf3 and sf7, or continue on with sf6. Doing sf6 will allow us to find more fields in the finite time we have left, but finishing off sf3 and sf7 will allow us to say with 100% certainty that they are indeed complete. I am leaning towards finishing off sf3 and sf7, but that may change as time goes on.

Neuigkeiten von Numberfields, sobald die Hardware den Geist aufgibt, wird das Projekt an dieser Stelle nicht mehr weiter geführt. Daher werden auch die Suchsettings angepasst, um kurzfristig (>1 Jahr) bestimmte Suchfelder fertig zu bekommen.

Gruß Odi

[odicin]

Es scheint so das die Suche dann ggf. bei Gerasim weiter geht. Zu mindestens ist die Get Decic Fields App für Win, Linux, NVidia und AMD dort aufgetaucht und es gibt Arbeit dafür.

Gruß Odi

[odicin]

2023 Year End Summary
[..]
Going forward we will continue chipping away at subfield 6. Meanwhile, the last remnants of subfield 7 have been farmed out to Gerasim.

Wegen der bekannten Hardwareproblematik (siehe Post vom 11.08.22) werden die Reste von subfield 7 bei Gerasim bearbeitet, nur das subfield 6 wird bei Numberfields zu Ende gebracht. Schon die Reste Subset 3 wurden von Gerasim gemacht, damit sich das Projekt auf ein subset auf dem Server konzentrieren kann. subset 3 ist im Jahr 2023 nun komplett fertig gestellt worden.