[QUOTE=R.D. Silverman;111562]We should not be missing factors under 40 digits......[/QUOTE]Agreed, but it cuts both ways. We're finding p48-p50 factors by ECM these days and we "shouldn't" be finding them with the amount of effort currently be applied.


Paul

On the other hand, on the hardware I've got a 143-digit-difficulty SNFS takes about ten hours and the 722 curves at B1=1e7 that ECM expects to need for a 40-digit factor take about sixteen hours, so I'm not sure a 40-digit factor is really an ECM miss at this level.

[I suppose saying that obliges me to set my machines to stomping through small SNFS runs; my present matrix for F1009 is 3.8M square of weight 350M, so I'll have time to do several SNFS runs on the other machine while the one with lots of memory sorts out the linear algebra]

I don't know how many ECM machines are out there: a thousand curves at 1e7 on each number in the table takes I guess one GHz-year on Opteron or Core2 machines. I suppose there are likely to be more than six ECM machines, and the tables came out last month, but I don't know how fast the CPUs are ...

8,7,173+ factored
 

c103 | 8,7,173+ is factored
[code]N=1403497557008862485716200846271790437037865255087497577561289699815630000925225345697367235455588464771
( 103 digits)
Divisors found:
r1=237311114983809397535134259633210257429757260095283 (pp51)
r2=5914166966450839911062940955586958322702325213217137 (pp52)
Version: GGNFS-0.77.1-20060513-pentium4
Total time: 12.11 hours.[/code]Divisors's primality checked with alpertron's java applet.

6,7,199+ factored
 

6,7,199+ is factored
[code]
Number: 6,7,199+
N=1069275425374024679398106207379298409471523458422913713168954653803374362245966831096700701840128052618609
( 106 digits)
Divisors found:
r1=1283581022644393510093595111271584179681 (pp40)
r2=833040849397365554383825941330945835699449509984493688972862067089 (pp66)
Version: GGNFS-0.77.1-20060513-athlon-xp
Total time: 55.42 hours.
Scaled time: 34.36 units (timescale=0.620).
[/code]

Divisors's primality checked with alpertron's java applet.

9,2,164+ factored
 

c106 | 9,2,164+ is factored
[code]
Number: 9,2,164+

N=8259711588703901915372111526231510631759849720341953289934989487214501192871460979310656906083491771818449
( 106 digits)
Divisors found:
r1=32380079681975934172208389377488163475425997516219873 (pp53)
r2=255086203302383855326202667402434251591014912331913713 (pp54)
Version: GGNFS-0.77.1-20060513-pentium4
Total time: 19.84 hours.
Scaled time: 25.10 units (timescale=1.265).

[/code]Divisors's primality checked with alpertron's java applet.

7,3,215+ factored
 

c109 | 7,3,215+ is factored

[code]Number: 7,3,215+
N=7167187165139253948528231197317484352651552082870862016466597207279657083359132520576460023863868125592165731
( 109 digits)
Divisors found:
r1=891947626988383924420586354082244878980948026969621 (pp51)
r2=8035434983261179939892656374850368655702152654266884821911 (pp58)
Version: GGNFS-0.77.1-20060513-pentium4
Total time: 27.34 hours.
Scaled time: 32.46 units (timescale=1.187).[/code]Divisors's primality checked with alpertron's java applet.

New update
 

The latest tables are uploading as I type this. There were 42 factorizations reported in the last two weeks and there are now 140 unfactored composites.

Despite Bob finding a p38 "ECM miss", there have been some impressive factors reported to the ECMNET server, including a p48. Sometimes you get unlucky and sometimes you get lucky. That's the nature of ECM. In fact, as the server is roughly half-way through handing out p45-optimal tasks, I'd expect the chance of missing a p38 to be comparable with that of finding a p48.


Paul

Homogeneous Cunninghams
 

[QUOTE=xilman;112148]The latest tables are uploading as I type this. There were 42 factorizations reported in the last two weeks and there are now 140 unfactored composites.

Despite Bob finding a p38 "ECM miss", there have been some impressive factors reported to the ECMNET server, including a p48. Sometimes you get unlucky and sometimes you get lucky. That's the nature of ECM. In fact, as the server is roughly half-way through handing out p45-optimal tasks, I'd expect the chance of missing a p38 to be comparable with that of finding a p48.


Paul[/QUOTE]

Here are two more:

9,5,152+

152 (8) 62856944357055017462865341921.168689762101167471281199842404712441111865889.24074438232147042150741880807856242419700149106977335675336039969

9,7,158+
158 (2) 1681400381924539640417.17132234421007860902639755863585221792849328761922653.1573576963765394424274462498769215987583358257960378202339980509363712629909


9,8,139+ seems to be "missing" from the tables.

[QUOTE=R.D. Silverman;112581]9,8,139+ seems to be "missing" from the tables.[/QUOTE]It seems to be in my table on the web ...


Paul

Another update is in progress. There are now 128 unfactored composites in the tables.


Paul

[QUOTE=xilman;112591]Another update is in progress. There are now 128 unfactored composites in the tables.


Paul[/QUOTE]

Here are a few lines from the table at:


[url]http://www.leyland.vispa.com/numth/factorization/anbn/9+8.txt[/url]

135 P69
136 1862583595234721.27577451610480123780025759297. P79
137 13303523.5814743609.89534726097264584137.202629831798581235224170114127947674113633. P52
138 277.6738541.3956998753.268392054481.38073249796033.1083142670013493. P26
139 1950449.64973847674507. C112
140 9521.104586721.9722288339825561675068321. P55
141 283.113535739.26828459910941672000233146121. P49


The entry for 139 lists a C112. Yet this number has not appeared in
any of the update files, nor is it on the reservations list (that I can find)