Hi Nuutti, ... You are quite right, you can either get rid of factors in the expression, or by sieving. But what I was trying to do there was find an...
594
Paul Jobling
Paul.Jobling@...
Mar 1, 2000 3:05 pm
All, Chris Caldwell has very kindly hosted my Proth presieving program at http://www.utm.edu/research/primes/programs/NewPGen/ Please feel free to have a look...
595
Paul Landon
plandon@...
Mar 1, 2000 6:45 pm
/** **\ In the name of GAUSS, The INFINITELY COUNTABLE, The FACTORISABLE. By those who range...
596
Dirk Augustin
augustd@...
Mar 1, 2000 7:12 pm
Paul, I have made a first performance test of your new sieving program. You've done a really good job! Congratulations! I sieved for Cun. Chains of the 2nd...
597
Chris Nash
chris_nash@...
Mar 1, 2000 10:34 pm
Hi folks ... Wow :) Great spoof on the comment lines that caused a little bit of a stir on primes-l for a routine that did indeed find all the primes, but...
598
Nuutti Kuosa
nuutti.kuosa@...
Mar 2, 2000 12:01 am
... I have tested it a little also. I am trying to find a 1000 digit sophie germain prime :) Lets see how long it will take. One note : program asks k min and...
599
Dirk Augustin
augustd@...
Mar 2, 2000 1:07 am
... Yes, kmin seems to be not working. Even if I start with a large kmin, for example kmin=500,000,000 the program starts at k=1. ... I would guess that...
600
Paul Jobling
Paul.Jobling@...
Mar 2, 2000 9:59 am
... Many thanks, Dirk. Yves Gallot has very kindly had a look at the code and almost immediately he came back with some conversions into assembler to make it...
601
Bouk de
bdewater@...
Mar 2, 2000 10:02 am
This program looks great! I am currenty running a seive for k*2^300000-1 to see how it works out on my lowly P 166. I will see how many I can weed out in a ...
602
Paul Jobling
Paul.Jobling@...
Mar 2, 2000 10:06 am
Hi Nuuti, ... No, you found the bug too. D'oh! That sort of slackness over testing and adding additional features at the last minute would cost me dear in my...
603
Paul Jobling
Paul.Jobling@...
Mar 2, 2000 10:15 am
Hi Bouk, ... I think you did - the maximum is 2.1 billion but really as Proth only goes to 1 billion there isn't much point using larger values....especially...
604
Bouk de
bdewater@...
Mar 2, 2000 11:06 am
Don't worry: I won't try n = -400 ;-) I have just one more question. Would it be possible to implement a range for n just like k? So one could, for instance,...
605
Paul Jobling
Paul.Jobling@...
Mar 2, 2000 12:25 pm
... I have given some thought to constant k, n varying testing, but it isn't as easy as constant n, k varying testing. Unless somebody knows the answer to the...
606
Yves Gallot
galloty@...
Mar 2, 2000 4:33 pm
... Development of Proth 5.5 will be finished next weekend and should be available in about 10 days, after a week of tests. The maximum for k is now 2^31 ~=...
607
Nuutti Kuosa
nuutti.kuosa@...
Mar 3, 2000 12:28 pm
Because there have been some talk about NewPGen I will introduce my test results. My goal was to find one Sophie germain prime size about 1,000 digit. I sieved...
608
Paul Jobling
Paul.Jobling@...
Mar 3, 2000 1:04 pm
Hi all, ... It is good to have some independent verification that this is a useful technique for speeding up searches. ... Are you are saying that it should...
609
Dirk Augustin
augustd@...
Mar 3, 2000 1:20 pm
I have tested NewPGen in the BiTwin mode and it is really fast. With Proth I would need more than 20 hours to test a single range (for example n=800,...
610
Paul Jobling
Paul.Jobling@...
Mar 3, 2000 2:07 pm
Hi Dirk, ... Well, I have just made the modifications (hey, it is Friday lunchtime) to allow CCs and BiTwin of arbitrary lengths to be specified. The limit on...
611
Dirk Augustin
augustd@...
Mar 3, 2000 3:10 pm
... Yes, it would be very surprising for me if anyone would find a Cunningham chain of length 10 with Proth. ... I believe (or more exactly: I hope) that at...
612
Nuutti Kuosa
nuutti.kuosa@...
Mar 3, 2000 3:38 pm
... prime, ... Are you are saying that it should allow longer chains to be sieved? I have recently added the option to search for length 3 Cunningham Chains...
613
Paul Landon
plandon@...
Mar 3, 2000 4:37 pm
Hiya, it was worth the attempt. I still haven't thrown the idea into the DEAD&FORGOTTEN bin yet. On my earlier question about an approximation for the number...
614
Yves Gallot
galloty@...
Mar 3, 2000 6:27 pm
... Be careful, if you use NewPGen in the BiTwin mode, you will not find all the twins of the form k.2^n+/-1 because those for which k.2^(n+1)+/-1 has a small...
615
Chris Nash
chris_nash@...
Mar 4, 2000 1:16 am
Hi folks ... Certainly do not do that - if there's ever the ability to efficiently optimize out calculations that are repeated, then it's worth pursuing. ...
616
Paul Jobling
Paul.Jobling@...
Mar 6, 2000 9:41 am
Hi Yves, all, ... I'll add that mode today - I guess it is mixed Twin/SG mode. Would there be any interest in a mixed Twin/CC mode as well? I might as well add...
617
Chris Nash
chris_nash@...
Mar 7, 2000 5:36 am
Hi folks, Over the past few weeks many people have written to me about bugs in PrimeForm v0.4.3, which no matter how I tried, I was unable to replicate. This...
618
Chris Nash
chris_nash@...
Mar 7, 2000 5:59 am
Hi folks Just a quick note - I have made a 'rush release' of PrimeForm v0.4.4, fixing the memory corruption error identified by Christ van Willegen. Once the...
619
Paul Jobling
Paul.Jobling@...
Mar 7, 2000 9:29 am
All, NewPGen version 1.7 is now available for download from http://www.utm.edu/research/primes/programs/NewPGen/ New features in this release: * A much wider...
620
Dick Boland
richard042@...
Mar 7, 2000 8:28 pm
I followed your advice Chris and used the pform.ini as provided, and had no problems, primeform seemed to behave just as it should. ... -Dick Boland ...
621
Dirk Augustin
augustd@...
Mar 8, 2000 1:25 am
Using NewPGen 1.7 I have just improved two of my records: Cunningham chain length 6, 2nd kind: 413022435*2^226+1 (previous record: 667630455*2^135+1) And...
622
Bouk de
bdewater@...
Mar 8, 2000 1:21 pm
It is more than twice as fast! Good work! Bouk de Water. ... http://click.egroups.com/1/937/4/_/619868/_/952421428/ ... ...
