mersenneforum.org  

Go Back   mersenneforum.org > Other Stuff > Archived Projects > 3*2^n-1 Search

 
 
Thread Tools
Old 2004-09-16, 19:36   #1
segmtfault
 
segmtfault's Avatar
 
Jun 2004
California

277 Posts
Unhappy Bug in LLR P4

I read this alarming thread today about a bug in LLR P4 which seems to miss primes for k's up to 22 bits :
http://mersenneforum.org/showthread.php?p=40739

As you can see, Jean Penné fixed the bug and will add an option to quickly find the numbers giving wrong results.

Since I am using LLR P4 1 for this project as probably most of the other people involved in the 321 search, I think it is important that we double check all our results since LLR P4 release (according to Jean, the double check should be a LOT faster since the wrong numbers should be detected during the init phase).

I can dedicate a P4 machine to the double check as soon as the new version is released, but I would need the list of all the numbers we tested with LLR P4.
Paul, do you think you could build such a list ?
(The easiest way would be to concatenate all workunits reserved since the release of LLR P4)
segmtfault is offline  
Old 2004-09-16, 21:31   #2
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

2·32·13·17 Posts
Default

Okay, you'll have to test from 1140000 to 1310000. I will send you a list as soon as Jean releases the bug-finding LLR. Hopefully it will not take you very long to check this out

Jean is doing a splendid job and I can hardly wait for the new LLR.
paulunderwood is offline  
Old 2004-09-17, 14:14   #3
scottsaxman
 
scottsaxman's Avatar
 
Aug 2004
way out west

2×13 Posts
Default

So, business as usual then? Or do we need to back off LLRing until the new program comes out?
scottsaxman is offline  
Old 2004-09-17, 16:05   #4
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

2·32·13·17 Posts
Default

Jean said in one message "So, I strongly suggest to all users to avoid using this program for the time beeing, the risk of missing prime numbers is too big."

The problem is that the older LLR is not available publicly.

I am continuing with my ranges until they finish and then run some double checks on LLRP4 work with an old unofficial LLR I have.

If you like you can suspend testing 321 numbers until the new LLRP4 comes out or you can continue and maybe find a primes like the last two which were found with the buggy version.
paulunderwood is offline  
Old 2004-09-17, 18:59   #5
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

2×32×13×17 Posts
Default

In fact what I will do is some double checking with the current blindingly fast albeit buggy LLRP4 on some of the earliest 321 files and see if I can find any mismatches. If anyone else wants a part of this action then email me at paulunderwood@mindless.com

Last fiddled with by paulunderwood on 2004-09-17 at 19:09
paulunderwood is offline  
Old 2004-09-17, 23:24   #6
TTn
 

55568 Posts
Default old

I have all versions of LLR if you would like to use an older one(12/20/02).
others
http://mersenneforum.org/showpost.ph...6&postcount=38

Once LLRp4 is working correctly, it seems to continue to work correctly during a logon(windows). So if you are still going to use the new version , at least check it's accuracy, with some existing known problemed examples first.

Last fiddled with by TTn on 2004-09-17 at 23:26
 
Old 2004-09-18, 17:18   #7
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

2·32·13·17 Posts
Default

Good idea! I will double check Markus Frind's troubled file against the current buggy LLRP4. Then I will test some of the earliest 321 files -- it should rip through those small "n" at 5 times as fast as the original LLR Hopefully we will not have to wait much longer for the new LLR to be verified and then released by Jean
paulunderwood is offline  
Old 2004-09-18, 17:49   #8
Citrix
 
Citrix's Avatar
 
Jun 2003

3×232 Posts
Default

Paul, there was some bugs in LLR2, before Jean fixed it. Is there a way to check these numbers LLRed using LLR2 real fast for bugs. Just to clearify this was the start and stop bug.

Thanks,
Citrix
Citrix is offline  
Old 2004-09-18, 18:47   #9
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

397810 Posts
Default

I get really confused as to which version is which. How many are there exactly and which ones have known bugs? Will the new one have bugs? More importantly will it find our next prime?

Last fiddled with by paulunderwood on 2004-09-18 at 18:49
paulunderwood is offline  
Old 2004-09-18, 19:21   #10
Citrix
 
Citrix's Avatar
 
Jun 2003

3×232 Posts
Default

I am asking about the version that came out earlier in the year( in april I guess).

there was a bug initally but was fixed in a few months down the line.

Citrix
Citrix is offline  
Old 2004-09-18, 22:03   #11
TTn
 

33×113 Posts
Default

Quote:
I get really confused as to which version is which. How many are there exactly and which ones have known bugs? Will the new one have bugs? More importantly will it find our next prime?
Official releases:(to the best of my knowledge)

12/20/2002
Had problems with small k and large k.(George thought it may be combo prob.)

7/19/2003
Doesn't work for small n < 300. Used by prof Caldwell to verify k*2^n-1 primes reported to Top-5000.

4/11/2004
The first stable IBDWT version. One false residue found when running on Pentium-4. No problems found on Athlon. Inefficient on both platforms when initial fft length guesses are wrong.

8/1/2004
LLR2

8/14/2004
LLRP42

8/22/2004
LLRP41

Last fiddled with by TTn on 2004-09-18 at 22:08
 
 

Thread Tools


All times are UTC. The time now is 07:53.


Thu Jan 20 07:53:06 UTC 2022 up 181 days, 2:22, 0 users, load averages: 1.96, 1.56, 1.34

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