[QUOTE=R.D. Silverman;96043]I see that Paul has already added some extensions.

Please note that I have already done

5,2,249-, 238+
5,3,243-, 245-, 233+, 250+
5,4,229+, 230+, 232+, 235+, 237+, 238+

I will send them to Paul when I return to my desk on Monday.[/QUOTE]I see that my announcement here isn't here, as it were. Very strange. I've never had posts disappear before now.

The summary is that I've added the 3,x tables to index 500 (was 400) and the 5,x tables to index 300 (was 250). There were 318 composites when I uploaded them to the web at [url]http://www.leyland.vispa.com/numth/factorization/anbn/main.htm[/url] but with Bob's results noted above and a few more received by email that total is clearly somewhat lower.


Paul

4,3,317- finished
 

Burned my hands on the stove, but finished cooking:

4,3,317- C187 = p59.p129

p59 = 13227646225937724272944520392629756157407220989538665766393
p129 = 229733952495204266560150125511063162346121749657389065703671671375649856380911884807442405883053996933310581233395088060111079783

That felt good -- good to have it done, that is. :yucky: :sick:

The matrix phase of ggnfs goes much faster with a surfeit of relations. After an 8-day linear algebra failure, I sieved for 10 more GHz days, at which point the matrix phase only took 4.5 days. Lighter-weight relations due to more large prime collisions, courtesy of the birthday problem?

4,3,317+ should be finished sieving this week. :ermm:

I think I'll stick to smaller SNFS factorizations until I know more about how all this works, either by rolling my own code (ambitious!) or perusing existing source code.

In spite of my whining, it has been quite satisfying so far. Much more to learn. Several threads on this site have been quite helpful.

[QUOTE=xilman;96049]I see that my announcement here isn't here, as it were. Very strange. I've never had posts disappear before now.

The summary is that I've added the 3,x tables to index 500 (was 400) and the 5,x tables to index 300 (was 250). There were 318 composites when I uploaded them to the web at [url]http://www.leyland.vispa.com/numth/factorization/anbn/main.htm[/url] but with Bob's results noted above and a few more received by email that total is clearly somewhat lower.


Paul[/QUOTE]

Here are some results. I seem to have misplaced 5,4,237+. I know
that I finished it, but apparantly I failed to enter it into my tables. It
is only C84. I will leave it for someone else.

5,2,+
238 (2,14,34) 2381.1444395820001.4336038313376141.4439578280596869874391771515377998261369.2791368726798872593181323551511317827490945751289933906573793829

5,2,-

249 (1,3,83) 15917886275881.379430828291524666867076641282481705401.453876922288597963738663664053822048234397490891868738699529071

5,3,+

233 (1) 467.89106850515391.1996654534559163949122235547837.10899087021695763579340228873062117328743323844282256033472834756815303436756182621670152328487607685765105185178929
250 (2,10,50) 158979001.965937945808775007660407557715337851473677001.405239598504855217046848271281602728150616805348116924007073914771169751721290567888001

5,3,-
243 (1,3,9,27,81) 487.26056392823.623867150599052585000639930886115726298821.21607540195055409527183393360153534365222491285911136156829
245 (1,5,7,35,49) 16349831.97313051590429992644712681280741.1632847030872667105139479820336035339329305406298020005914484956706257917787691

5,4,+

229 (1) 2749.49892748442261.585147188477895871619353787.16047640711154399322814570918737295587643837824388017428287515213232423187248869680631675082628938577965145697934487
230 (2,10,46) 161921.747853602083024437091752734195197341023807300747481.12768365831991328536090735994043188199147035794685837570680741534521
232 (8) 413211914881.62285999377990247848847904185329249.158409917558520102395633095053459169.779071313063788786549816161168514109921026740721319746256257121435412246721
235 (1,5,47) 8461.26321.365135367719155937792335574794960327203027544787401.6799387644953849047254068936200110328608776038611306711294900208586321
238 (2,14,34) 5237.86874408534929.189561975568415899740770700173.7494954654531852764906216431020778987934213.386970162750752438589407520566732166034344733

[QUOTE=FactorEyes;96097]Burned my hands on the stove, but finished cooking:

4,3,317- C187 = p59.p129

p59 = 13227646225937724272944520392629756157407220989538665766393
p129 = 229733952495204266560150125511063162346121749657389065703671671375649856380911884807442405883053996933310581233395088060111079783

That felt good -- good to have it done, that is. :yucky: :sick:

The matrix phase of ggnfs goes much faster with a surfeit of relations. After an 8-day linear algebra failure, I sieved for 10 more GHz days, at which point the matrix phase only took 4.5 days. Lighter-weight relations due to more large prime collisions, courtesy of the birthday problem?

4,3,317+ should be finished sieving this week. :ermm:

I think I'll stick to smaller SNFS factorizations until I know more about how all this works, either by rolling my own code (ambitious!) or perusing existing source code.

In spite of my whining, it has been quite satisfying so far. Much more to learn. Several threads on this site have been quite helpful.[/QUOTE]

4,3,298+ is smaller and would finish base 4,3 to index 300.

[QUOTE=R.D. Silverman;96141]Here are some results. I seem to have misplaced 5,4,237+. I know
that I finished it, but apparantly I failed to enter it into my tables. It
is only C84. I will leave it for someone else.

<snip>

[/QUOTE]


I have 5,4,250+ as well:

5,4, +

250 (2,10,50) 6333038413045313451168451001.9826130927089133719737940360239268546931049671089505795660172422324337629613660567647491588938331961909196433001

[QUOTE=R.D. Silverman;96159]I have 5,4,250+ as well:

5,4, +

250 (2,10,50) 6333038413045313451168451001.9826130927089133719737940360239268546931049671089505795660172422324337629613660567647491588938331961909196433001[/QUOTE]

And a partial result for 5,3,229+

9011132130494074891377113961034819.C106

[QUOTE=FactorEyes;96097]
The matrix phase of ggnfs goes much faster with a surfeit of relations. After an 8-day linear algebra failure, I sieved for 10 more GHz days, at which point the matrix phase only took 4.5 days. Lighter-weight relations due to more large prime collisions, courtesy of the birthday problem?
[/QUOTE]
More relations = more lighter-weight relations, plus if you have many large primes you can pick and choose for the matrix the ones that form cycles with fewer relations, then throw away the rest of the relations. It's a rule of thumb that once you have about 10-15% more relations than you need to form a matrix at all, the matrix you get can be a little smaller but have half the nonzero entries.

The birthday problem only applies to the quadratic sieve with one large prime. The behavior of QS with two large primes and NFS with multiple large primes is much more complicated.

jasonp

5,2,227+
 

[QUOTE=R.D. Silverman;96160]And a partial result for 5,3,229+

9011132130494074891377113961034819.C106[/QUOTE]

Here is 5,2,227+ C158 = p57.p102

363208303748573630593692837773585407342096316553868817401
182363187834133751692583924920591372799212683840516955111705468803190905200219367431513070793759851379

5,3,226+ is in progress. 5,2,232+ will be next.

I am a bit surprised about the lack of results from the extensions.

[QUOTE=R.D. Silverman;96503]I am a bit surprised about the lack of results from the extensions.[/QUOTE]I have received a few more, but only a few, by personal email (which I prefer, by the way).

Unfortunately I've not yet found time to update the web pages and post a summary here. I hope to be able to do so this weekend.

Paul

[QUOTE=xilman;96506]I have received a few more, but only a few, by personal email (which I prefer, by the way).

Unfortunately I've not yet found time to update the web pages and post a summary here. I hope to be able to do so this weekend.

Paul[/QUOTE]

I post in public to let others know what I am doing.

Hopefully, this prevents duplication of effort.

[QUOTE=R.D. Silverman;96509]I post in public to let others know what I am doing.[/QUOTE]In that spirit, from 3,2,495+, C90 = p37.p54:

p37 = 2689510889073404152260600172103673361
p54 = 260210445078307092674098506182878025080405580334147001

[QUOTE]4,3,298+ is smaller and would finish base 4,3 to index 300[/QUOTE]

4,3,317+ has finished sieving. I'll start 4,3,298+ this evening if nobody else has -- I'll check back here before then.

There are a handful of C12x and C13x composites in the 4,3 and 6,5 tables. Maybe I'll show them some love.