mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > Riesel Prime Search

Reply
 
Thread Tools
Old 2010-05-02, 23:37   #1
Oddball
 
Oddball's Avatar
 
May 2010

1111100112 Posts
Default What's up with LLR?

When I was testing the number 30387*2^500000-1, this message comes up:

30387*2^500000-1, iteration : 41000/500000 [8.19%] Time per...

41000/500000 should be exactly 8.20000%, not 8.19%. Is there any explanation for that?
Oddball is offline   Reply With Quote
Old 2010-05-03, 02:20   #2
Kosmaj
 
Kosmaj's Avatar
 
Nov 2003

2×1,811 Posts
Default

Because log2(30387) is about 15 and log2 of your number is 500015.
Kosmaj is offline   Reply With Quote
Old 2010-05-03, 11:57   #3
Mini-Geek
Account Deleted
 
Mini-Geek's Avatar
 
"Tim Sorbera"
Aug 2006
San Antonio, TX USA

427910 Posts
Default

Quote:
Originally Posted by Kosmaj View Post
Because log2(30387) is about 15 and log2 of your number is 500015.
So does LLR really do ~500015 iterations in this scenario? I know that for a +1 Proth test LLR would show (and, presumably, run) ~500015 iterations. Odd that it might be inconsistent like that, if that's the case...

Last fiddled with by Mini-Geek on 2010-05-03 at 11:59
Mini-Geek is offline   Reply With Quote
Old 2010-05-03, 12:19   #4
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

2·3·7·71 Posts
Default

Here're the differences on that test for LLR V3.71c and V3.8.
Attached Thumbnails
Click image for larger version

Name:	LLR_37-38.jpg
Views:	175
Size:	40.6 KB
ID:	5116  
kar_bon is offline   Reply With Quote
Old 2010-05-03, 12:51   #5
Mini-Geek
Account Deleted
 
Mini-Geek's Avatar
 
"Tim Sorbera"
Aug 2006
San Antonio, TX USA

11·389 Posts
Default

I'm seeing inconsistent behavior.
At least in LLR 3.8.0, the Windows GUI version, running 30387*2^500000-1, iteration 54000 says "54000 / 500000 [10.80%]" and iteration 63000 says "63000 / 500000 [12.59%]".

Karsten, I think that what you're seeing is because 3.8 is using a PRP test, (there's an option related to that in 3.8, do you have it set so it will run PRP tests even when it can run LLR tests?) which displays the bits, and 3.7.1c is using the LLR test, which just displays the n (whether it really runs n iterations or bit_length_of_number iteration)

Last fiddled with by Mini-Geek on 2010-05-03 at 12:59
Mini-Geek is offline   Reply With Quote
Old 2010-05-03, 13:39   #6
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

298210 Posts
Default

Quote:
Originally Posted by Mini-Geek View Post
Karsten, I think that what you're seeing is because 3.8 is using a PRP test, (there's an option related to that in 3.8, do you have it set so it will run PRP tests even when it can run LLR tests?) which displays the bits, and 3.7.1c is using the LLR test, which just displays the n (whether it really runs n iterations or bit_length_of_number iteration)
Run LLRV3.8 (without any ini-file) and an inputfile with such testpair. You got the "Strong Fermat PRP test" with 500015 bits to test without any settings!

PS: I've tested the version from 2010-02-08 and this is unsing the 'normal' LLR-test with 500000 iterations to test! Seems Jean changed something in the versions from Jan to Feb!

Last fiddled with by kar_bon on 2010-05-03 at 13:44
kar_bon is offline   Reply With Quote
Old 2010-05-03, 17:52   #7
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

3·2,083 Posts
Default

Quote:
Originally Posted by kar_bon View Post
Run LLRV3.8 (without any ini-file) and an inputfile with such testpair. You got the "Strong Fermat PRP test" with 500015 bits to test without any settings!

PS: I've tested the version from 2010-02-08 and this is unsing the 'normal' LLR-test with 500000 iterations to test! Seems Jean changed something in the versions from Jan to Feb!
Are you sure you don't have ForcePRP=1 set in llr.ini? That would make it do the Fermat PRP test instead of the usual LLR.
mdettweiler is offline   Reply With Quote
Old 2010-05-03, 18:20   #8
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

1011101001102 Posts
Default

Quote:
Originally Posted by mdettweiler View Post
Are you sure you don't have ForcePRP=1 set in llr.ini? That would make it do the Fermat PRP test instead of the usual LLR.
I've tested it with all three settings (one by the other) and with none: The Jan-version tests PRP for standard, the Feb-version tests LLR (with that example given).

Last fiddled with by kar_bon on 2010-05-03 at 18:20
kar_bon is offline   Reply With Quote
Old 2010-05-03, 21:27   #9
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

186916 Posts
Default

Quote:
Originally Posted by kar_bon View Post
I've tested it with all three settings (one by the other) and with none: The Jan-version tests PRP for standard, the Feb-version tests LLR (with that example given).
Ah, I see. Perhaps it was a bug in the pre-release version that was since fixed, then? (I didn't upgrade from 3.7.1c until 3.8 was fully released, i.e. the Feb version.)

Last fiddled with by mdettweiler on 2010-05-03 at 21:27
mdettweiler is offline   Reply With Quote
Reply

Thread Tools


All times are UTC. The time now is 08:54.


Sun Dec 4 08:54:30 UTC 2022 up 108 days, 6:23, 0 users, load averages: 1.07, 0.87, 0.84

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 𝛴𝜎𝜍 𝛵𝜏 𝛶𝜐 𝛷𝜙𝜑 𝛸𝜒 𝛹𝜓 𝛺𝜔