mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Data

Reply
 
Thread Tools
Old 2003-10-03, 02:21   #1
dave_0273
 
dave_0273's Avatar
 
Oct 2003
Australia, Brisbane

2×5×47 Posts
Default 60bit, 64 bit - what is the difference??

I am sorry if this is a really dumb question.

I have a pentium 4 2.4Ghz, with a gig of ram.

From what i was reading in the forums i should have my testing above 64 bit. However, when i looked at the number that i was testing M33506581, it was doing it at 60 bit.

What is the difference. How do i change it so that it is working at 64 bit. Is 64 bit the best for my computer or would something else be better. I see that some people use up to 69 bit.

I am sorry for such a dumb question, i applogise, i am new.

Anyway, any help would be appreciated.

Last fiddled with by dave_0273 on 2003-10-03 at 02:22
dave_0273 is offline   Reply With Quote
Old 2003-10-03, 02:58   #2
dsouza123
 
dsouza123's Avatar
 
Sep 2002

2·331 Posts
Default

If you are trial factoring, the mersenne number you get is already partially tested by trial factoring upto a lower bit depth ( the high 50s or to 60 ) because no factor showed up it then is available to be factored to a some predetermined bit depth ( the size of the number determines it).

So when you start, it is at 60 it will keep going until it either finishes to the predetermined bit depth ( 69 is probably correct for a 33 million number ) or it finds a factor.

For P-1 factoring somebody else would have to answer.
dsouza123 is offline   Reply With Quote
Old 2003-10-04, 02:24   #3
cheesehead
 
cheesehead's Avatar
 
"Richard B. Woods"
Aug 2002
Wisconsin USA

170148 Posts
Default Re: 60bit, 64 bit - what is the difference??

Quote:
Originally posted by dave_0273
I am sorry if this is a really dumb question.
No, it's quite reasonable.

The user documentation for GIMPS is about as good as computer programmers and mathematicians who are not also professional technical writers usually write. That might be sufficient if it were gathered all in one place -- but it's not; it's scattered.

It's quite common for newbies, and even not-so-newbies, not to pick up the correct understanding of some things discussed in this forum.

Quote:
I have a pentium 4 2.4Ghz, with a gig of ram.

From what i was reading in the forums i should have my testing above 64 bit.
What that meant was that for the particular task of trial factoring, P4s are most the efficent current models to use above 2^64, and also that P4s are relatively less efficient than other models for trial factoring below 2^64.

But when you're assigned a LL test, trial factoring is only a small preliminary step of your total work, and it's not worth worrying about a P4's relative inefficency for the small part of the trial factoring that involves testing divisors up to 2^64.

Quote:
However, when i looked at the number that i was testing M33506581, it was doing it at 60 bit.
That just means that previous trial factoring had been performed only to 2^60, so your assignment's first step will be to continue trial factoring up to the default limit (2^68) for your number.

In other words, the 60 is just a starting point, not some parameter of how the program operates otherwise.

Quote:
What is the difference. How do i change it so that it is working at 64 bit.
You don't need to change anything!

Quote:
Is 64 bit the best for my computer or would something else be better.
It's not a matter of making a better or worse choice. You don't need to make any such choice.

Quote:
I see that some people use up to 69 bit.
That just means that for their numbers, trial factoring has already proceeded to 2^69 (or, actually, 2^68 -- sometimes the 69 really means 68, but that's a separate long story!). It has nothing to do with any choice they made or any different operation of their software, and you needn't be concerned about it.
cheesehead is offline   Reply With Quote
Old 2003-10-04, 07:18   #4
dave_0273
 
dave_0273's Avatar
 
Oct 2003
Australia, Brisbane

2×5×47 Posts
Default

thanks alot. that has cleared a lot up for me.

i have just one other question (well for the moment, i am sure that in time there will be many many more), i have a dual cpu so i am running 2 copies of prime 95. For both i requested to work on 10,000,000 digit numbers but the programs seem to be working differently.

one's output is...

factoring M33506581 to 2^68 is 61% complete....


the others screen output is....

interation 517000/33376487 [1.55%] Per iteration time.....

why are the two programs doing different things. i wanted to set them up the same but have i actually set them up differently???

thankyou to anyone who can help....
dave_0273 is offline   Reply With Quote
Old 2003-10-04, 13:33   #5
smh
 
smh's Avatar
 
"Sander"
Oct 2002
52.345322,5.52471

29×41 Posts
Default

The first one is doing trial factoring. It's currently at 61% of factoring up to 2^68.

This is the prelimanary step that cheesehead was talking about.

When reaches 100% it will either start doing P-1 testing to find even larger factors, or starts doing a Lucas Lehmer test.

That is what the second cpu is doing. A LL test is used to determine if a number is prime or composite.
smh is offline   Reply With Quote
Old 2003-10-04, 13:33   #6
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

29·167 Posts
Default

The search for a Mersenne prime walks through different steps.

From Readme.txt:

Quote:
Factoring M400037 to 2^54 is 3.02% complete. Time: 0.121 sec.
This means prime95 is trying to find a small factor of 2^400037-1.
It is 3.02% of the way though looking at factors below 2^54. When
this completes it may start looking for factors less than 2^55.
Iteration: 941400 / 1667747 [56.45%]. Per iteration time: 0.109 sec.
This means prime95 just finished the 941400th iteration of a
Lucas-Lehmer primality test. The program must execute 1667747
iterations to complete the primality test. The average iteration
took 0.109 seconds.
Your two instances of prime95 are doing different jobs: the first is trial-factoring the number up to 2^68, while the second has completed the factoring step and went through the so-called Lucas Lehmer test.

HTH

Luigi
ET_ is offline   Reply With Quote
Old 2003-10-04, 14:21   #7
markr
 
markr's Avatar
 
"Mark"
Feb 2003
Sydney

10758 Posts
Default

It is quite possible that, for the exponent that you are LL testing, the factoring stages were already done before it was assigned to you, saving you some work! That could explain why your two copies are a different stages, if you started them at the same time.

The Math section of the help in the program is good, if you want to know more about the stages each exponent goes through.

markr (in Sydney)
markr is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Spot the Difference a1call Miscellaneous Math 34 2017-12-11 00:17
Unexpected speed difference cybergreg Hardware 13 2016-12-23 09:31
The US is not a democracy, and the difference is important jasong jasong 19 2014-06-28 18:19
V4 and V5 TF Reporting Difference RMAC9.5 PrimeNet 2 2010-05-23 02:20
difference between WZ1 and Wc1 in results.txt? ixfd64 Lounge 2 2005-11-19 18:13

All times are UTC. The time now is 16:32.


Fri Oct 7 16:32:03 UTC 2022 up 50 days, 14 hrs, 0 users, load averages: 1.73, 1.76, 1.60

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