mersenne.ca

[James Heinrich]

Originally Posted by chalsall

Could you please put this in a public-facing report that "only human" can understand?
With pretty graphs too? On a dashboard?

I have put a very rough public-facing report. With no graphs or controls of any kind. But you can choose between plaintext and really-plaintext.
https://www.mersenne.ca/ecmcurves.php
https://www.mersenne.ca/ecmcurves.php?f=t

If someone actually uses this and wishes it has some more filtering controls, please let me know and I'll see about adding them.

[chalsall]

Originally Posted by James Heinrich

If someone actually uses this and wishes it has some more filtering controls, please let me know and I'll see about adding them.

Thanks, James et al.

Few appreciate the depth of peer review around these here parts... 8^>

[lycorn]

Originally Posted by James Heinrich

I have updated the curve count for 35..65 from your ECM page. I didn't follow from what you posted how to extend my lookup table to 80 digits (both bounds and curve count).

Yes it would be nice to have the table in GIMPS ECM page extended to something like 80 digits, as for several of the lowest exponents it is just showing the curves were "done" up to 65 digits, but several people are still running larger curves that are not recorded (or at least their count not displayed). There is a rumour that significant improvements are coming to the GIMPS ECM software so many of us will certainly be wishing to record fresh curves for the lowest exponents...

[SethTro]

Originally Posted by Prime95

Note there are updated curve counts on the GIMPS ECM page as of last week (yes I should have done this years ago). These come from GMP-ECM -- these guys know more about ECM than I ever will.

Using B1=2400000000, B2=240000000000, polynomial Dickson(12), sigma=0:5846338435282416409
dF=17280, k=72, d=180180, d2=17, i0=13304
Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
22      67      226     845     3407    14988   69914   350860  1845420 1e+07

Using B1=7200000000, B2=720000000000, polynomial Dickson(30), sigma=0:17190718471612911695
dF=17280, k=216, d=180180, d2=17, i0=39944
Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
16      44      135     455     1660    6464    27094   120646  567615  2809445

Using B1=21600000000, B2=2160000000000, polynomial Dickson(30), sigma=0:14360955280911240780
dF=17280, k=647, d=180180, d2=17, i0=119864
Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
12      32      89      272     905     3227    12256   49294   209024  930668

If anyone knows how to get GMP-ECM to show proper curve counts for t25 and t30 please let me know.

https://gitlab.inria.fr/zimmerma/ecm...ecm.c#L853-880

Also if you build from source you can use `timeout 0.1 ecm -v` as the curves are now printed before any work is done.

There's also a rho_curve tool that I'm working on updating but it's largely not needed if you don't mind parsing the ecm -v output.

[SethTro]

Originally Posted by James Heinrich

I have updated the curve count for 35..65 from your ECM page. I didn't follow from what you posted how to extend my lookup table to 80 digits (both bounds and curve count).

I don't know where George pulled the bounds from but they used 2.4e9, 7.2e9, and 21.6e9 and the counts are from the relevant columns: 350860, 567615, 930668.

Thanks for the updated data! I attached the bounds for all 322 unique curves from 2020, 2021, 2022.

[James Heinrich]

New side project if anyone wants to try some YAFU factoring for mersenne.ca (its own internal use):
https://www.mersenneforum.org/showthread.php?t=27919

[kriesel]

James has been notified by PM.
Example: try to retrieve https://www.mersenne.ca/exponent/121883029
Another example: try to update https://www.mersenne.ca/obd.php for % completion on TF.
Another example: reload https://www.mersenne.ca/obd

There might be more.

Those return currently in response, only the error message following, and no expected content.

Parse error:  syntax error, unexpected ')' in /var/www/vhosts/mersenne.ca/httpdocs/factorize.inc.php on line 200

Some pages of mersenne.ca function normally.

[James Heinrich]

Originally Posted by kriesel

syntax error, unexpected ')' in factorize.inc.php on line 200

Fixed.

[kriesel]

Originally Posted by James Heinrich

Fixed.

In ~34 minutes from first PM notification. Swift service, and can't beat the price either! :)

[lycorn]

https://www.mersenne.ca/status/tf/
Stuck at 2022-07-16T23:59:59+00:00.

[James Heinrich]

Hmm, there are (still) 438,566 factor candidates (~7000M range) in the factor queue from end-of-day on 2022-07-17, and the TF status processing script won't run if there's unprocessed factors in the queue (to make the numbers accurate for the day; it wasn't designed for an influx of nearly a million new factors...)

I'll have to see about expediting the processing of the factor queue and getting the status script running again. Unfortunately it means the data for the 17th won't be accurate (just a copy of the 18th) but nothing I can do about that now.