mersenneforum.org 100M digits, how much trial factoring will it do?
 User Name Remember Me? Password
 Register FAQ Search Today's Posts Mark Forums Read

 2009-03-02, 08:20 #1 xorbe     Feb 2009 4916 Posts 100M digits, how much trial factoring will it do? How many bits can I expect it to do trial factoring up to, until it starts the P-1 step? (2^74? 2^77?) (just for reference, v25.8 build 4) I used Manual Assignment from the web page, and pasted it into worktodo.txt It reports the results automatically, right? I read through the forum, but it's not clear. The software and web page has been updated a lot since some of the forum posts were written. Last fiddled with by xorbe on 2009-03-02 at 08:26
2009-03-02, 08:54   #2

"Richard B. Woods"
Aug 2002
Wisconsin USA

769210 Posts

Quote:
 Originally Posted by xorbe How many bits can I expect it to do trial factoring up to, until it starts the P-1 step? (2^74? 2^77?)
There's a source module line

#define FAC78 337400000L

that means exponent 337,400,000 (101,567,520 decimal digits) is the dividing point between TF limits of 2^77 and 2^78. However, those are the ultimate total limits. Prime95 and PrimeNet assignments now stop TF 2 bits short of that limit, then do P-1, then do the final 2 bits of TF.

So, for an exponent slightly under 337,400,000, I'd expect TF to 2^75, then P-1, then TF from 2^75 to 2^77.

Quote:
 It reports the results automatically, right?
Yes, if you have the right settings in Test->PrimeNet and Advanced->Manual Communication. I don't think any of that has changed in quite a while.

Last fiddled with by cheesehead on 2009-03-02 at 09:02

 2009-03-02, 19:06 #3 xorbe     Feb 2009 7310 Posts Wow... 2^72: 1.4397 GHz·days 2^73: 2.8793 GHz·days 2^74: 61% done So, at 3.66 GHz, 75-77 are going to take 3.1, 6.3, 12.6 days. I notice the workload page says "1943 days to go" for the exponent, hrm that's a long time...
 2009-03-02, 20:31 #4 Batalov     "Serge" Mar 2008 Phi(4,2^7658614+1)/2 3×132×19 Posts That workload page aint lying. 5-6 core years? Sounds about right, give or take a few years. Please drop a note later, how much time the P-1 will have taken... Also, here's a bit forgotten thread -- useful to read... http://www.mersenneforum.org/showthread.php?t=11037
 2009-03-03, 02:42 #5 xorbe     Feb 2009 73 Posts According to Advanced -> Time, LL will take about 1 year, 2.6 months (444 days) -- with 8 threads anyways. That's actually doable. 1 thread was estimated 1725 days. Depending whether P-1 step is after 2^75 or 2^77, I'll report the P-1 time estimation in 4-13 days from now... Last fiddled with by xorbe on 2009-03-03 at 02:48
 2009-03-03, 10:35 #6 xorbe     Feb 2009 1118 Posts Huh, if I manually set the affinity of the thread doing TF, it runs 10% faster.
2009-03-07, 23:18   #7
xorbe

Feb 2009

73 Posts

Too bad we can't edit posts after a while. Otherwise, I would update the first post in this thread with the answer that it started P-1 stage 1 after TF to 2^75. Instead an interested person will have to scan down the thread...

Quote:
 Originally Posted by Batalov Please drop a note later, how much time the P-1 will have taken...
M(332197123), P-1 stage 1 was at 4.6% today, and looks like it's going to take 6.5 more days, and "Chance ... 6.05%".

Last fiddled with by xorbe on 2009-03-07 at 23:21

2009-03-08, 18:07   #8

"Richard B. Woods"
Aug 2002
Wisconsin USA

22×3×641 Posts

Quote:
 Originally Posted by xorbe Instead an interested person will have to scan down the thread...
Interested persons won't have any trouble -- this forum's software makes it easy to find, and go to, new unread posts.

Quote:
 M(332197123), P-1 stage 1 was at 4.6% today, and looks like it's going to take 6.5 more days, and "Chance ... 6.05%".
Be skeptical about prime95's time estimates for P-1. I find that it consistently underestimates P-1 stage 2 duration by a factor of about 2 on my Pentium D. YMMV

 2009-03-08, 22:00 #9 xorbe     Feb 2009 73 Posts I estimated the time manually with a calculator (from on-screen % and time info) only for Stage 1 -- no idea about Stage 2 until it gets there... Last fiddled with by xorbe on 2009-03-08 at 22:01
 2009-03-09, 01:27 #10 cheesehead     "Richard B. Woods" Aug 2002 Wisconsin USA 22×3×641 Posts So you're not clicking on Test->Status to see the estimated completion dates?
2009-03-09, 06:38   #11
xorbe

Feb 2009

73 Posts

Quote:
 Originally Posted by cheesehead So you're not clicking on Test->Status to see the estimated completion dates?
That only tells me its expected completion of LL, and nothing about P-1 Stage 1 or Stage 2, or TF 2^76 or 2^77. (Win64 client, v25.8b4)

About 21% now, on track to complete P-1 Stage 1 around next weekend I guess...

Last fiddled with by xorbe on 2009-03-09 at 06:41

 Similar Threads Thread Thread Starter Forum Replies Last Post joblack LMH > 100M 1 2009-10-08 12:31 davieddy Lounge 1 2008-10-18 10:40 Uncwilly LMH > 100M 351 2008-09-25 06:45 abc_temp Factoring 14 2007-12-23 20:25 harlee Software 3 2006-10-15 04:38

All times are UTC. The time now is 08:17.

Thu Dec 9 08:17:16 UTC 2021 up 139 days, 2:46, 0 users, load averages: 2.10, 2.02, 1.78