mersenneforum.org  

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

Reply
 
Thread Tools
Old 2011-02-28, 18:35   #1
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

7·13·47 Posts
Default M2629093 has a factor: 63721413359381377

Quote:
M2629093 has a factor: 63721413359381377
I found this, unexpectedly, in P-1 stage2. The question is, why wasn't it found before? As far as PrimeNet knew, it was TF'd up to 2^61 (unfortunately, PrimeNet erases, or at least doesn't display, other details of an exponent's history once a factor is found), and this should've been found in TF to 2^56 (mfaktc also found the factor no problem when I asked it to look in that range). So how was it skipped on PrimeNet? Faked results?
James Heinrich is offline   Reply With Quote
Old 2011-02-28, 18:40   #2
lycorn
 
lycorn's Avatar
 
"GIMFS"
Sep 2002
Oeiras, Portugal

30568 Posts
Default

TF ro 61 bits performed by a faulty machine?
lycorn is offline   Reply With Quote
Old 2011-02-28, 20:59   #3
cheesehead
 
cheesehead's Avatar
 
"Richard B. Woods"
Aug 2002
Wisconsin USA

22·3·641 Posts
Default

How long would it take for a 64-bit system to do a TF to 2^56 on all unfactored exponents up to, say, 50M?
cheesehead is offline   Reply With Quote
Old 2011-02-28, 21:37   #4
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

7·13·47 Posts
Default

Quote:
Originally Posted by cheesehead View Post
How long would it take for a 64-bit system to do a TF to 2^56 on all unfactored exponents up to, say, 50M?
In terms of actual work (note that effort numbers are culmulative, not per-range):
Code:
up to M1,xxx,xxx = 508GHz-days
up to M2,xxx,xxx = 638GHz-days
up to M3,xxx,xxx = 726GHz-days
up to M4,xxx,xxx = 792GHz-days
up to M5,xxx,xxx = 842GHz-days
up to M6,xxx,xxx = 882GHz-days
up to M7,xxx,xxx = 914GHz-days
up to M8,xxx,xxx = 944GHz-days
up to M9,xxx,xxx = 970GHz-days
up to M10,xxx,xxx = 993GHz-days
up to M11,xxx,xxx = 1,014GHz-days
up to M12,xxx,xxx = 1,033GHz-days
up to M13,xxx,xxx = 1,051GHz-days
up to M14,xxx,xxx = 1,067GHz-days
up to M15,xxx,xxx = 1,082GHz-days
up to M16,xxx,xxx = 1,096GHz-days
up to M17,xxx,xxx = 1,109GHz-days
up to M18,xxx,xxx = 1,121GHz-days
up to M19,xxx,xxx = 1,133GHz-days
up to M20,xxx,xxx = 1,144GHz-days
up to M21,xxx,xxx = 1,154GHz-days
up to M22,xxx,xxx = 1,164GHz-days
up to M23,xxx,xxx = 1,173GHz-days
up to M24,xxx,xxx = 1,182GHz-days
up to M25,xxx,xxx = 1,191GHz-days
up to M26,xxx,xxx = 1,199GHz-days
up to M27,xxx,xxx = 1,207GHz-days
up to M28,xxx,xxx = 1,214GHz-days
up to M29,xxx,xxx = 1,222GHz-days
up to M30,xxx,xxx = 1,229GHz-days
up to M31,xxx,xxx = 1,236GHz-days
up to M32,xxx,xxx = 1,242GHz-days
up to M33,xxx,xxx = 1,248GHz-days
up to M34,xxx,xxx = 1,255GHz-days
up to M35,xxx,xxx = 1,261GHz-days
up to M36,xxx,xxx = 1,266GHz-days
up to M37,xxx,xxx = 1,272GHz-days
up to M38,xxx,xxx = 1,277GHz-days
up to M39,xxx,xxx = 1,283GHz-days
up to M40,xxx,xxx = 1,288GHz-days
up to M41,xxx,xxx = 1,293GHz-days
up to M42,xxx,xxx = 1,298GHz-days
up to M43,xxx,xxx = 1,303GHz-days
up to M44,xxx,xxx = 1,307GHz-days
up to M45,xxx,xxx = 1,312GHz-days
up to M46,xxx,xxx = 1,317GHz-days
up to M47,xxx,xxx = 1,321GHz-days
up to M48,xxx,xxx = 1,325GHz-days
up to M49,xxx,xxx = 1,329GHz-days
Note that the amount of effort to factor to a fixed bitdepth varies greatly on the exponent size (smaller exponents, lots more effort).

A variant approach would be to TF to 4 bitlevels below the default level for that range, which yields very different numbers:
Code:
up to M1,xxx,xxx = 0GHz-days (~2^52)
up to M2,xxx,xxx = 13GHz-days (~2^54)
up to M3,xxx,xxx = 50GHz-days (~2^56)
up to M4,xxx,xxx = 118GHz-days (~2^57)
up to M5,xxx,xxx = 218GHz-days (~2^57)
up to M6,xxx,xxx = 361GHz-days (~2^58)
up to M7,xxx,xxx = 490GHz-days (~2^59)
up to M8,xxx,xxx = 608GHz-days (~2^59)
up to M9,xxx,xxx = 789GHz-days (~2^60)
up to M10,xxx,xxx = 975GHz-days (~2^60)
up to M11,xxx,xxx = 1,144GHz-days (~2^60)
up to M12,xxx,xxx = 1,297GHz-days (~2^60)
up to M13,xxx,xxx = 1,439GHz-days (~2^60)
up to M14,xxx,xxx = 1,647GHz-days (~2^61)
up to M15,xxx,xxx = 1,886GHz-days (~2^61)
up to M16,xxx,xxx = 2,112GHz-days (~2^61)
up to M17,xxx,xxx = 2,322GHz-days (~2^61)
up to M18,xxx,xxx = 2,518GHz-days (~2^61)
up to M19,xxx,xxx = 2,703GHz-days (~2^61)
up to M20,xxx,xxx = 2,876GHz-days (~2^61)
up to M21,xxx,xxx = 3,041GHz-days (~2^61)
up to M22,xxx,xxx = 3,199GHz-days (~2^61)
up to M23,xxx,xxx = 3,347GHz-days (~2^61)
up to M24,xxx,xxx = 3,574GHz-days (~2^62)
up to M25,xxx,xxx = 3,848GHz-days (~2^62)
up to M26,xxx,xxx = 4,109GHz-days (~2^62)
up to M27,xxx,xxx = 4,364GHz-days (~2^62)
up to M28,xxx,xxx = 4,608GHz-days (~2^62)
up to M29,xxx,xxx = 4,841GHz-days (~2^62)
up to M30,xxx,xxx = 5,219GHz-days (~2^63)
up to M31,xxx,xxx = 5,918GHz-days (~2^63)
up to M32,xxx,xxx = 6,593GHz-days (~2^63)
up to M33,xxx,xxx = 7,244GHz-days (~2^63)
up to M34,xxx,xxx = 7,867GHz-days (~2^63)
up to M35,xxx,xxx = 8,481GHz-days (~2^63)
up to M36,xxx,xxx = 9,075GHz-days (~2^63)
up to M37,xxx,xxx = 9,647GHz-days (~2^63)
up to M38,xxx,xxx = 10,313GHz-days (~2^64)
up to M39,xxx,xxx = 11,403GHz-days (~2^64)
up to M40,xxx,xxx = 12,462GHz-days (~2^64)
up to M41,xxx,xxx = 13,500GHz-days (~2^64)
up to M42,xxx,xxx = 14,510GHz-days (~2^64)
up to M43,xxx,xxx = 15,492GHz-days (~2^64)
up to M44,xxx,xxx = 16,460GHz-days (~2^64)
up to M45,xxx,xxx = 17,387GHz-days (~2^64)
up to M46,xxx,xxx = 18,304GHz-days (~2^64)
up to M47,xxx,xxx = 19,201GHz-days (~2^64)
up to M48,xxx,xxx = 20,507GHz-days (~2^65)
up to M49,xxx,xxx = 22,139GHz-days (~2^65)
up to M50,xxx,xxx = 23,737GHz-days (~2^65)
Or for 8 levels below standard:
Code:
up to M1,xxx,xxx = 0GHz-days (~2^48)
up to M2,xxx,xxx = 1GHz-days (~2^50)
up to M3,xxx,xxx = 3GHz-days (~2^52)
up to M4,xxx,xxx = 7GHz-days (~2^53)
up to M5,xxx,xxx = 14GHz-days (~2^53)
up to M6,xxx,xxx = 23GHz-days (~2^54)
up to M7,xxx,xxx = 34GHz-days (~2^55)
up to M8,xxx,xxx = 49GHz-days (~2^55)
up to M9,xxx,xxx = 71GHz-days (~2^56)
up to M10,xxx,xxx = 95GHz-days (~2^56)
up to M11,xxx,xxx = 116GHz-days (~2^56)
up to M12,xxx,xxx = 135GHz-days (~2^56)
up to M13,xxx,xxx = 153GHz-days (~2^56)
up to M14,xxx,xxx = 179GHz-days (~2^57)
up to M15,xxx,xxx = 209GHz-days (~2^57)
up to M16,xxx,xxx = 237GHz-days (~2^57)
up to M17,xxx,xxx = 263GHz-days (~2^57)
up to M18,xxx,xxx = 288GHz-days (~2^57)
up to M19,xxx,xxx = 311GHz-days (~2^57)
up to M20,xxx,xxx = 332GHz-days (~2^57)
up to M21,xxx,xxx = 353GHz-days (~2^57)
up to M22,xxx,xxx = 373GHz-days (~2^57)
up to M23,xxx,xxx = 391GHz-days (~2^57)
up to M24,xxx,xxx = 420GHz-days (~2^58)
up to M25,xxx,xxx = 454GHz-days (~2^58)
up to M26,xxx,xxx = 486GHz-days (~2^58)
up to M27,xxx,xxx = 518GHz-days (~2^58)
up to M28,xxx,xxx = 549GHz-days (~2^58)
up to M29,xxx,xxx = 578GHz-days (~2^58)
up to M30,xxx,xxx = 606GHz-days (~2^59)
up to M31,xxx,xxx = 633GHz-days (~2^59)
up to M32,xxx,xxx = 660GHz-days (~2^59)
up to M33,xxx,xxx = 685GHz-days (~2^59)
up to M34,xxx,xxx = 710GHz-days (~2^59)
up to M35,xxx,xxx = 734GHz-days (~2^59)
up to M36,xxx,xxx = 757GHz-days (~2^59)
up to M37,xxx,xxx = 779GHz-days (~2^59)
up to M38,xxx,xxx = 805GHz-days (~2^60)
up to M39,xxx,xxx = 848GHz-days (~2^60)
up to M40,xxx,xxx = 889GHz-days (~2^60)
up to M41,xxx,xxx = 930GHz-days (~2^60)
up to M42,xxx,xxx = 969GHz-days (~2^60)
up to M43,xxx,xxx = 1,008GHz-days (~2^60)
up to M44,xxx,xxx = 1,046GHz-days (~2^60)
up to M45,xxx,xxx = 1,082GHz-days (~2^60)
up to M46,xxx,xxx = 1,118GHz-days (~2^60)
up to M47,xxx,xxx = 1,153GHz-days (~2^60)
up to M48,xxx,xxx = 1,206GHz-days (~2^61)
up to M49,xxx,xxx = 1,273GHz-days (~2^61)
up to M50,xxx,xxx = 1,339GHz-days (~2^61)
Or for the lesser of (default TF level -or- 2^56):
Code:
up to M1,xxx,xxx = 0GHz-days (~2^52)
up to M2,xxx,xxx = 13GHz-days (~2^54)
up to M3,xxx,xxx = 50GHz-days (~2^56)
up to M4,xxx,xxx = 116GHz-days (~2^56)
up to M5,xxx,xxx = 166GHz-days (~2^56)
up to M6,xxx,xxx = 205GHz-days (~2^56)
up to M7,xxx,xxx = 237GHz-days (~2^56)
up to M8,xxx,xxx = 267GHz-days (~2^56)
up to M9,xxx,xxx = 293GHz-days (~2^56)
up to M10,xxx,xxx = 316GHz-days (~2^56)
up to M11,xxx,xxx = 337GHz-days (~2^56)
up to M12,xxx,xxx = 356GHz-days (~2^56)
up to M13,xxx,xxx = 374GHz-days (~2^56)
up to M14,xxx,xxx = 390GHz-days (~2^56)
up to M15,xxx,xxx = 405GHz-days (~2^56)
up to M16,xxx,xxx = 419GHz-days (~2^56)
up to M17,xxx,xxx = 432GHz-days (~2^56)
up to M18,xxx,xxx = 445GHz-days (~2^56)
up to M19,xxx,xxx = 456GHz-days (~2^56)
up to M20,xxx,xxx = 467GHz-days (~2^56)
up to M21,xxx,xxx = 477GHz-days (~2^56)
up to M22,xxx,xxx = 487GHz-days (~2^56)
up to M23,xxx,xxx = 497GHz-days (~2^56)
up to M24,xxx,xxx = 505GHz-days (~2^56)
up to M25,xxx,xxx = 514GHz-days (~2^56)
up to M26,xxx,xxx = 522GHz-days (~2^56)
up to M27,xxx,xxx = 530GHz-days (~2^56)
up to M28,xxx,xxx = 538GHz-days (~2^56)
up to M29,xxx,xxx = 545GHz-days (~2^56)
up to M30,xxx,xxx = 552GHz-days (~2^56)
up to M31,xxx,xxx = 559GHz-days (~2^56)
up to M32,xxx,xxx = 565GHz-days (~2^56)
up to M33,xxx,xxx = 572GHz-days (~2^56)
up to M34,xxx,xxx = 578GHz-days (~2^56)
up to M35,xxx,xxx = 584GHz-days (~2^56)
up to M36,xxx,xxx = 590GHz-days (~2^56)
up to M37,xxx,xxx = 595GHz-days (~2^56)
up to M38,xxx,xxx = 601GHz-days (~2^56)
up to M39,xxx,xxx = 606GHz-days (~2^56)
up to M40,xxx,xxx = 611GHz-days (~2^56)
up to M41,xxx,xxx = 616GHz-days (~2^56)
up to M42,xxx,xxx = 621GHz-days (~2^56)
up to M43,xxx,xxx = 626GHz-days (~2^56)
up to M44,xxx,xxx = 631GHz-days (~2^56)
up to M45,xxx,xxx = 635GHz-days (~2^56)
up to M46,xxx,xxx = 640GHz-days (~2^56)
up to M47,xxx,xxx = 644GHz-days (~2^56)
up to M48,xxx,xxx = 648GHz-days (~2^56)
up to M49,xxx,xxx = 653GHz-days (~2^56)
up to M50,xxx,xxx = 657GHz-days (~2^65)

Assuming a 4-core i7-920 @ 3.0GHz can do 4.6GHz-days per day per core, it can do approx 560GHz-days per month of TF. You can figure out how long it would take to do any/all of the above.
edit: of course, a system with a good CUDA GPU could potentially get it done much faster...

Last fiddled with by James Heinrich on 2011-02-28 at 21:46
James Heinrich is offline   Reply With Quote
Old 2011-02-28, 23:04   #5
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

3·7·17·31 Posts
Default

Quote:
Originally Posted by cheesehead View Post
How long would it take for a 64-bit system to do a TF to 2^56 on all unfactored exponents up to, say, 50M?
Maybe someone with a hot GPU and mfaktc could do this as a project (at least in part.)
Uncwilly is online now   Reply With Quote
Old 2011-03-01, 01:00   #6
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

102658 Posts
Default

Quote:
Originally Posted by Uncwilly View Post
Maybe someone with a hot GPU and mfaktc could do this as a project (at least in part.)
I could generate the big worktodo if anyone wants to take it on.
<a few minutes pass>
Actually, I did, so people can grab it and work on it if they want. But someone should co-ordinate the effort so there's not too much overlap of work. There's just over 1 million assignment lines :surprised
http://mersenne-aries.sili.net/temp/worktodo-TF56.7z (since attachments here are limited to 256kB)
James Heinrich is offline   Reply With Quote
Old 2011-03-01, 02:06   #7
firejuggler
 
firejuggler's Avatar
 
"Vincent"
Apr 2010
Over the rainbow

23·5·73 Posts
Default

don't forget that mfaktc ignore the exponent below 1 000 000
firejuggler is offline   Reply With Quote
Old 2011-03-01, 02:19   #8
KingKurly
 
KingKurly's Avatar
 
Sep 2010
Annapolis, MD, USA

19810 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
I could generate the big worktodo if anyone wants to take it on.
<a few minutes pass>
Actually, I did, so people can grab it and work on it if they want. But someone should co-ordinate the effort so there's not too much overlap of work. There's just over 1 million assignment lines :surprised
http://mersenne-aries.sili.net/temp/worktodo-TF56.7z (since attachments here are limited to 256kB)
I did M4999XXXX and found nothing (...which, I suppose, is exactly what I expected to find!) Even though it took barely any time at all, I don't plan on continuing unless I come across some spare CPU/GPU time. I've already got enough assignments, and every time I look at your P1small page I always find more...

I wonder how many factors we should expect to find, if someone decides to continue (start?) the effort. I mean, sure, we shouldn't find any, but I think the point of the exercise is that we may well find some missing ones.
KingKurly is offline   Reply With Quote
Old 2011-03-01, 02:30   #9
firejuggler
 
firejuggler's Avatar
 
"Vincent"
Apr 2010
Over the rainbow

23·5·73 Posts
Default

ok so...
gtx 460 and core 2 duo E8300.
one assignement take about 18 seconds. but use 'only' 60 M/sec . It goes up to 120 M/s on billion digit assignment under full load.
but since I only have 2 core, it slow down a bit.... so with 2 instance, i get one assignment in 21 seconds, plus half of one between the end of one and the start of the following.

Considering 1 M assignment, 21 500 000 seconds/2=10 750 000, 10750000/3600,2986.1 hours or a bit more than 124 days of work.

Last fiddled with by firejuggler on 2011-03-01 at 03:11 Reason: edit : more reasonable timing
firejuggler is offline   Reply With Quote
Old 2011-03-01, 02:30   #10
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

10000101101012 Posts
Default

Quote:
Originally Posted by firejuggler View Post
don't forget that mfaktc ignore the exponent below 1 000 000
And likewise Prime95 refuses TF on exponents < 20,000
James Heinrich is offline   Reply With Quote
Old 2011-03-01, 02:35   #11
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

7·13·47 Posts
Default

Since I don't have a particularly good GPU to throw at it, I'll just run the ~14000 exponents <1M overnight (on Prime95) and see what it comes up with. I'm churning through them at about 1/second in that range, so shouldn't take long. Hopefully no factors ;)

Last fiddled with by James Heinrich on 2011-03-01 at 02:35
James Heinrich is offline   Reply With Quote
Reply



Similar Threads
Thread Thread Starter Forum Replies Last Post
A new factor of F11?! siegert81 FermatSearch 2 2018-01-24 04:35
1M-9M GPU TF vs. CPU P-1/ECM (KWh/factor) VictordeHolland Lone Mersenne Hunters 4 2015-03-07 09:29
New factor for F17 Buckle Factoring 15 2011-03-15 12:05
use of factor? (just to be sure) Ivan Semenov Data 2 2004-05-29 14:30
Shortest time to complete a 2^67 trial factor (no factor) dsouza123 Software 12 2003-08-21 18:38

All times are UTC. The time now is 13:24.


Fri Jul 7 13:24:34 UTC 2023 up 323 days, 10:53, 0 users, load averages: 1.50, 1.25, 1.18

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2023, 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