R311...

[Xyzzy]

Here are some benchmarks and some save files... Hopefully Xilman will drop in here and explain what needs to be done for this number...

(10^311 - 1 ) / 9

Digits   B1     B1 ms
    50 43e6   2319099
    55 11e7   5926153
    60 26e7  14038577
    65 85e7  46068559
    70 29e8 158656304

[xilman]

Originally Posted by Xyzzy

Here are some benchmarks and some save files... Hopefully Xilman will drop in here and explain what needs to be done for this number...

I'm not entirely sure what I'm supposed to say ...

I know that ECM has been run to at least the 50-digit level, because I did 57% of the required work and Bruce Dodson did the other 43%. A relatively small amount of work has been done with B1>43M but I'm not entirely sure how much. My guess is the equivalent of a few hundred curves at the p55 level but I'll see if I can discover more.

Although I haven't calculated the optimum cut-off point, at least in part because no-one really knows how hard it will be by SNFS, most people seem to think that ECM should be run to at least 55 digits and possibly further.

Paul

[akruppa]

From past factoizations I found that usually a reasonable rule of thumb is to test GNFS numbers for factors up to 1/3 the size of the number, and SNFS numbers up to 2/9 of the difficulty. For R311, that would mean ECM up to almost 70 digits, which seems rather a lot and I'm not sure how well this rule of thumb holds up for numbers this large. I'm still pretty sure that ECM well beyond 55 digits is worthwhile, probably even beyond 60.

Alex

[Xyzzy]

9 curves with B1=43e6
32 curves with B1=11e7

[geoff]

Since the new Prime95 can do numbers of this form I tried it out, but even on a P4, stage one with Prime95 24.6 takes about 15% longer than gmp-ecm 5.0.3.

[Mystwalker]

Any progress on this number?

I've just noticed that no 55-digit effort has been reported on the 10- tables so far.

As this number is a possible (or even likely?) candidate for a kilobit SNFS, I think it should reach a proper ECM level within a year or two...

[Xyzzy]

228 curves using B1=11e7 & B2=680270182898...

[fatphil]

From PZ's ECM records page - slotted in at the #2 of all-time:

digits: 64
f: 4344673058714954477761314793437392900672885445361103905548950933
N: 10^311-1
B1: 85e7
sigma: 1917732841
date: 05 Sep 2005
Finder: K. Aoki & T. Shimoyama

Very impressive indeed.

[fatphil]

This forum is broke.

I just got a mail saying:
"""
Jwb52z has just replied to a thread you have subscribed to entitled - R311... -
in the Factoring forum of mersenneforum.org.

This thread is located at:
http://www.mersenneforum.org/showthr...2&goto=newpost

Here is the message that has just been posted:
***************
Am I right in assuming that this P64 means that R311 is finished, meaning no
more work needs to be done now, and I can report it to Kirk Pearson as such with
the P64 being the factor found?
***************

"""

And there's no such post here?!?!?

Anyway - it doesn't matter. Kirk knows.
I know he knows, as I was the one who told him.

[akruppa]

>This forum is broke.

Nope, Jwb52z deleted his post right after posting it - but at that time, the notification email had been sent already. I can still see his deleted posting, but that may be because I have mod powers...

Alex

[Jwb52z]

This is my fault guys, I'm sorry. I was checking on the "older" factoring projects that haven't had much done on them in a while and I forgot that Kirk already had the P64 factor listed.