mersenneforum.org  

Go Back   mersenneforum.org > Search Forums

Showing results 1 to 25 of 25
Search took 0.01 seconds.
Search: Posts Made By: danaj
Forum: Computer Science & Computational Number Theory 2018-02-06, 05:48
Replies: 59
Views: 22,174
Posted By danaj
Supplanting Primo for practical use is really...

Supplanting Primo for practical use is really unlikely. It's nice to have a decent open source version though, and now we have another one with Pari/GP. I admit I rather hoped for more immediate...
Forum: Computer Science & Computational Number Theory 2018-02-02, 19:51
Replies: 59
Views: 22,174
Posted By danaj
Wow, that's great! The standalone ecpp is...

Wow, that's great!

The standalone ecpp is generated from https://github.com/danaj/Math-Prime-Util-GMP by running xt/create-standalone.sh.

I recently added the ability to use Ramanujan class...
Forum: Computer Science & Computational Number Theory 2015-03-13, 18:21
Replies: 59
Views: 22,174
Posted By danaj
No. Typically I use mpz_tdiv_r because it gives...

No. Typically I use mpz_tdiv_r because it gives better performance.


That's a nice idea -- no I haven't. A simple google search shows this one (http://www.christelbach.com/ECCalculator.aspx)...
Forum: Computer Science & Computational Number Theory 2014-10-22, 14:33
Replies: 59
Views: 22,174
Posted By danaj
Thank you very much! That's immensely helpful,...

Thank you very much! That's immensely helpful, and should let me efficiently make this type of certificates. I'm doing the standard way you mention at the beginning. It works well, but of course...
Forum: Computer Science & Computational Number Theory 2014-10-21, 08:22
Replies: 59
Views: 22,174
Posted By danaj
Good point. While François Morain (author of...

Good point. While François Morain (author of numerous ECPP papers) would have lots of good ideas, for this purpose Marcel Martin (author of Primo) would be the person to talk to.
Forum: Computer Science & Computational Number Theory 2014-10-21, 07:00
Replies: 59
Views: 22,174
Posted By danaj
I'm sure there is a chance, but I'm not the...

I'm sure there is a chance, but I'm not the person to ask :). The other option is one I have control over (or anyone else since it is open source), which is adjusting the code to create Primo...
Forum: Computer Science & Computational Number Theory 2014-10-21, 06:55
Replies: 59
Views: 22,174
Posted By danaj
I don't have any special insight into Primo's...

I don't have any special insight into Primo's internals. I believe it is better algorithms for the first step. Certainly Primo has been worked on for much longer. Small implementation...
Forum: Computer Science & Computational Number Theory 2014-10-20, 08:30
Replies: 59
Views: 22,174
Posted By danaj
Apologies, but no, it cannot. factordb is...

Apologies, but no, it cannot. factordb is running this verifier, but has a pre-step that ensures the uploaded certificate is in the Primo format. This is arguably a good thing in that we don't want...
Forum: Computer Science & Computational Number Theory 2014-08-21, 05:08
Replies: 59
Views: 22,174
Posted By danaj
I find it very useful for my own work, but...

I find it very useful for my own work, but mpz_aprcl (http://sourceforge.net/projects/mpzaprcl/), Pari (http://pari.math.u-bordeaux.fr/), and Primo (http://www.ellipsa.eu/public/primo/primo.html) are...
Forum: Computer Science & Computational Number Theory 2014-08-21, 00:53
Replies: 59
Views: 22,174
Posted By danaj
Link to latest version: ecpp-dj.tar.gz...

Link to latest version: ecpp-dj.tar.gz (http://sti15.com/nt/ecpp-dj.tar.gz)

Permission to edit the first post would be nice to allow it to show the latest information.

Version 1.04...
Forum: Computer Science & Computational Number Theory 2013-12-24, 18:11
Replies: 59
Views: 22,174
Posted By danaj
I forwarded it to you. S=1 is allowed if all...

I forwarded it to you. S=1 is allowed if all other conditions are met. His email has more details.
Forum: Computer Science & Computational Number Theory 2013-12-23, 23:01
Replies: 59
Views: 22,174
Posted By danaj
It does, but I'll first say that I discussed the...

It does, but I'll first say that I discussed the Primo proof for that number with Marcel Martin and got a reply. He is going to update his verifier, and I'll update mine and send to Syd so we can...
Forum: Computer Science & Computational Number Theory 2013-08-10, 22:29
Replies: 59
Views: 22,174
Posted By danaj
I know my native code is much faster on x86_64...

I know my native code is much faster on x86_64 and Power7 64-bit machines, but that's due to the mulmod code that isn't used by GMP. Almost every machine I have access to now is 64-bit, but I'll dig...
Forum: Computer Science & Computational Number Theory 2013-08-06, 17:48
Replies: 59
Views: 22,174
Posted By danaj
This is more a math-of-ECPP question, but to the...

This is more a math-of-ECPP question, but to the best of my knowledge, no. We can combine the proofs, but only really in the sense that for any proof that reads "N is prime if Q [and Q2 and Q3...]...
Forum: Computer Science & Computational Number Theory 2013-07-29, 07:21
Replies: 59
Views: 22,174
Posted By danaj
It shouldn't actually require affine coordinates....

It shouldn't actually require affine coordinates. I've thought about changing my code to do it in projective, which should be much faster (at certainly is for ECM). This won't change the proof...
Forum: Computer Science & Computational Number Theory 2013-07-21, 10:29
Replies: 59
Views: 22,174
Posted By danaj
Performance graph: [link to image...

Performance graph: [link to image (http://sti15.com/nt/ecpp-perf1.png)]

Because of the scale (from under 0.0001 seconds to over 40,000) I've used a log-log graph. This does require some extra...
Forum: Computer Science & Computational Number Theory 2013-07-21, 07:33
Replies: 59
Views: 22,174
Posted By danaj
Version 1.01 is available: ecpp-dj-1.01.tar.gz...

Version 1.01 is available: ecpp-dj-1.01.tar.gz (http://sti15.com/nt/ecpp-dj-1.01.tar.gz)

Some of the changes:


A much better defined text certificate format (described in the...
Forum: Computer Science & Computational Number Theory 2013-06-15, 09:47
Replies: 59
Views: 22,174
Posted By danaj
I'm not following what you're asking. With the...

I'm not following what you're asking. With the FPS version of my code run with verbose 2, if it can't find any easy factors, it will show the progress of each candidate, e.g. "q no luck (820...
Forum: Computer Science & Computational Number Theory 2013-06-14, 17:59
Replies: 59
Views: 22,174
Posted By danaj
This page...

This page (http://primes.utm.edu/prove/merged.html) on the "prime pages" website is pretty useful. Here follows a rambling text with much handwaving, ovesimplification, and not reviewed for...
Forum: Computer Science & Computational Number Theory 2013-06-14, 16:48
Replies: 59
Views: 22,174
Posted By danaj
This is interesting in that the structure would...

This is interesting in that the structure would be substantially different than a program meant for proving many "small" numbers.

Some thoughts off the top of my head. Using ECPP and going purely...
Forum: Computer Science & Computational Number Theory 2013-06-14, 16:04
Replies: 59
Views: 22,174
Posted By danaj
Updated table, including Primo (one core) and...

Updated table, including Primo (one core) and reruns with GMP-ECPP after playing more with parameters. Times are in seconds.


Primo v 4.0.3 LX64, 22 bits (4.0.1 with 26 bits is a little faster...
Forum: Computer Science & Computational Number Theory 2013-06-13, 18:41
Replies: 59
Views: 22,174
Posted By danaj
With one core or 12? I can make arguments both...

With one core or 12? I can make arguments both ways in terms of fairness. If Boinc is the goal, then it's easy enough to load up the machine with N numbers all individually running in serial. But...
Forum: Computer Science & Computational Number Theory 2013-06-13, 18:39
Replies: 59
Views: 22,174
Posted By danaj
I tried running that in early May and I didn't...

I tried running that in early May and I didn't get it to run (it segfaults after entering the number). With some help from strace I've got it running now (v6.4.5a). I wrote in the comments at the...
Forum: Computer Science & Computational Number Theory 2013-06-13, 08:22
Replies: 59
Views: 22,174
Posted By danaj
I'm not involved in factordb so maybe I'm not...

I'm not involved in factordb so maybe I'm not following this. You could stop partway, but why not just do the full proof and enter every q value? That way factordb gets the whole chain entered, but...
Forum: Computer Science & Computational Number Theory 2013-06-12, 17:27
Replies: 59
Views: 22,174
Posted By danaj
ECPP-DJ

As mentioned on the Primo (http://mersenneforum.org/showthread.php?t=16643) thread, I wrote an ECPP implementation for my Math::Prime::Util (https://metacpan.org/release/Math-Prime-Util) Perl module,...
Showing results 1 to 25 of 25

 
All times are UTC. The time now is 10:41.


Sun Jan 16 10:41:49 UTC 2022 up 177 days, 5:10, 0 users, load averages: 0.90, 0.86, 0.90

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.

≠ ± ∓ ÷ × · − √ ‰ ⊗ ⊕ ⊖ ⊘ ⊙ ≤ ≥ ≦ ≧ ≨ ≩ ≺ ≻ ≼ ≽ ⊏ ⊐ ⊑ ⊒ ² ³ °
∠ ∟ ° ≅ ~ ‖ ⟂ ⫛
≡ ≜ ≈ ∝ ∞ ≪ ≫ ⌊⌋ ⌈⌉ ∘ ∏ ∐ ∑ ∧ ∨ ∩ ∪ ⨀ ⊕ ⊗ 𝖕 𝖖 𝖗 ⊲ ⊳
∅ ∖ ∁ ↦ ↣ ∩ ∪ ⊆ ⊂ ⊄ ⊊ ⊇ ⊃ ⊅ ⊋ ⊖ ∈ ∉ ∋ ∌ ℕ ℤ ℚ ℝ ℂ ℵ ℶ ℷ ℸ 𝓟
¬ ∨ ∧ ⊕ → ← ⇒ ⇐ ⇔ ∀ ∃ ∄ ∴ ∵ ⊤ ⊥ ⊢ ⊨ ⫤ ⊣ … ⋯ ⋮ ⋰ ⋱
∫ ∬ ∭ ∮ ∯ ∰ ∇ ∆ δ ∂ ℱ ℒ ℓ
𝛢𝛼 𝛣𝛽 𝛤𝛾 𝛥𝛿 𝛦𝜀𝜖 𝛧𝜁 𝛨𝜂 𝛩𝜃𝜗 𝛪𝜄 𝛫𝜅 𝛬𝜆 𝛭𝜇 𝛮𝜈 𝛯𝜉 𝛰𝜊 𝛱𝜋 𝛲𝜌 𝛴𝜎 𝛵𝜏 𝛶𝜐 𝛷𝜙𝜑 𝛸𝜒 𝛹𝜓 𝛺𝜔