Factoring progress has really improved!

eepiccolo · 2003-04-11, 15:05

I like to keep track of George's status report, and have been since December. I've noticed an interesting trend in the factoring progress.

http://home.earthlink.net/~eepiccolo...20progress.jpg

As you can see looking at the trend lines, in late February, the throughput of factored numbers nearly doubled! The Lone Mersenne Hunters started their forum on mersenneforum.com earlier in February, so my guess would be that having the forum here got people more aware and interested in factoring. Way to go, LMHs!

smh · 2003-04-11, 21:37

Yes, and factors only get sparser as we factor deeper, so the total effort put in it has gone up even more

hbock · 2003-04-11, 22:43

I also track the status page since Feb 2002, here some facts for the last year (02 Feb 2002 to 03 Feb 2003) :

Factored : +54055 (12077 below 20.4M)
Status Unknown : -130412 (-86706 below 20.4M)
Expected New Primes : -0.59
P-90 CPU Years : -839511 (-116786 below 20.4M)

So we got an average of about 148.5 factors per day (or 1040 per week).
Within the last 5 weeks we had about 180 factors per day, consistent to your calculated slope of 177 factors.
Also worth mentioning that about 78% CPU Years were saved due to factoring in the range 35.1-79.3M (whole factoring contribution >86 %).

Indeed, the forum has a great impact on factoring and the GIMPS performance at all!

trif · 2003-04-12, 02:04

I do wonder what impact that will have on factoring progress overall. I've been keeping track of Primenet tests completed over the past two weeks, and over my observation period 2,612 trial factoring tests have been completed, and 3,427 first time LL's (does include 33M's, but those are not a significant fraction, also includes tests that ended due to finding a factor in P1). This only includes results reported to Primenet, not anything being done by the LMH. In order to keep up with LL's, trial factoring needs to complete about 10% more tests than LL's completed (due to factors found). Instead, TF completion is only about 76% of LL completion.

2003-04-11, 15:05	#1
eepiccolo Dec 2002 Frederick County, MD 2×5×37 Posts	Factoring progress has really improved! I like to keep track of George's status report, and have been since December. I've noticed an interesting trend in the factoring progress. http://home.earthlink.net/~eepiccolo...20progress.jpg As you can see looking at the trend lines, in late February, the throughput of factored numbers nearly doubled! The Lone Mersenne Hunters started their forum on mersenneforum.com earlier in February, so my guess would be that having the forum here got people more aware and interested in factoring. Way to go, LMHs!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Improved NFS polynomial selection	jasonp	Operation Kibibit	5	2014-09-07 11:02
Progress	R.D. Silverman	Factoring	0	2012-05-22 14:03
In Progress?	R.D. Silverman	Cunningham Tables	33	2010-05-07 14:02
could oddperfect's ecm progress page be improved?	jasong	GMP-ECM	11	2007-05-30 03:08
Henk's graphs of factoring progress	trif	Lone Mersenne Hunters	4	2003-02-18 16:51

2003-04-11, 21:37	#2
smh "Sander" Oct 2002 52.345322,5.52471 29×41 Posts	Yes, and factors only get sparser as we factor deeper, so the total effort put in it has gone up even more

2003-04-11, 22:43	#3
hbock Feb 2003 163₁₀ Posts	I also track the status page since Feb 2002, here some facts for the last year (02 Feb 2002 to 03 Feb 2003) : Factored : +54055 (12077 below 20.4M) Status Unknown : -130412 (-86706 below 20.4M) Expected New Primes : -0.59 P-90 CPU Years : -839511 (-116786 below 20.4M) So we got an average of about 148.5 factors per day (or 1040 per week). Within the last 5 weeks we had about 180 factors per day, consistent to your calculated slope of 177 factors. Also worth mentioning that about 78% CPU Years were saved due to factoring in the range 35.1-79.3M (whole factoring contribution >86 %). Indeed, the forum has a great impact on factoring and the GIMPS performance at all!

2003-04-12, 02:04	#4
trif Aug 2002 312₈ Posts	I do wonder what impact that will have on factoring progress overall. I've been keeping track of Primenet tests completed over the past two weeks, and over my observation period 2,612 trial factoring tests have been completed, and 3,427 first time LL's (does include 33M's, but those are not a significant fraction, also includes tests that ended due to finding a factor in P1). This only includes results reported to Primenet, not anything being done by the LMH. In order to keep up with LL's, trial factoring needs to complete about 10% more tests than LL's completed (due to factors found). Instead, TF completion is only about 76% of LL completion.