New banner ad for the forum...

[quote="ewmayer"]
The present case has r = 24036583 == 7 modulo 8, so admits of a primitive trinomial over GF(2) of form x^r + x^s + 1, 0 < s < r. Actually finding the magic value of s for the middle term will require a lot of compute time - I'll ask Richard B. what his estimate is.[/quote]
Richard sent me the following e-mail about this:

[quote="Richard Brent"][color=green]
"With the algorithm that we (and most other people) use the time per
trinomial of degree n goes up like n^2 so the overall time like n^3.
For degree 6972593 we took an estimated 230,000 Mips-years (about
2.4 years actual time). Scaling by (24036583/6972593)^3
gives 9.4 million Mips-years or about 85 actual years with the same
computing resources.

"However, we have some ideas for speedups (using asymptotically faster GCD
algorithms and sieving further, so there are less "hard" cases to consider).
Preliminary tests on a 3Ghz P4 suggest that a speedup of about 2.9 for
degree 6972593 and speedup of about 4.3 for degree 24036583 is possible.
Thus, the estimate for degree 24036583 is about 2.3 million Mips-years (769
years on a 3Ghz P4). If we had access to a few hundred such machines the
computation would be feasible.

"The expected time to find a new primitive trinomial is about one third
of the time to do an exhaustive search for that degree, since the expected
number is about three. However, for degree 6972593 we only found one!

"At present we are developing the new algorithm and using it to verify results
for smaller Mersenne exponents (currently 756839) before embarking on a very long computation for the new exponent. See the paragraph "Extended format" at [url]http://web.comlab.ox.ac.uk/oucl/work/richard.brent/trinomlg.html[/url] for some results using the new GCD algorithm."[/color][/quote]

I was wondering if I can get a save file to run the last couple thousand interations so I can "feel" like I discovered something. (LOL that sounds so corny) Plus, I want to listen to the software audio :wink:

Here is my save file at 23M...

[url]http://www.mersenneforum.org/pO036583.023[/url]

I'll leave it up there for a week or so but I will have to remove it eventually since it is so big...

The doublecheck run on my system verified the new Prime :banana: .

A side-effect was that the mprime client did report the new prime quite thoroughly, and as the output was piped to a log file, that became very large :shock:.

[CODE]16022092 -rw-rw-r-- 1 Prime95 Prime95 16390592633 May 28 09:55 log.txt [/CODE]

PM

Banner
 

Ahh, being bored last night, I decided to do something with my time... Here's a quick banner :unsure: Comments please! :ermm:

[URL]http://www.alienaa.com/members/46201/banner2.jpg[/URL]

[QUOTE=Mike]Ahh, being bored last night, I decided to do something with my time... Here's a quick banner :unsure: Comments please! :ermm:

[URL]http://www.alienaa.com/members/46201/banner2.jpg[/URL][/QUOTE]

Corperate firewall blocks access to that site. :censored:

[QUOTE=Xyzzy][pre][May 27 15:25] Iteration: 22770000 / 24036583 [94.73%]. Per iteration time: 0.044 sec.
[/pre][/QUOTE]

Is it finished now?

[QUOTE=Uncwilly]Corperate firewall blocks access to that site. :censored:[/QUOTE]

[QUOTE=Uncwilly]Corperate firewall blocks access to that site. :censored:[/QUOTE]

LOL, look at it when you get home... I tried to upload it onto the MBs but it can only take 97.5 kb, or something like that, and it's 125 kb because I jacked the quality up :whistle:

[QUOTE=jinydu]Is it finished now?[/QUOTE]Yes, my run was on Prime Monster's box... Never thought I'd get a 16GB file full of "New Mersenne Prime!" messages...


:rolleyes: