Software/instructions/questions - Page 38

gd_barnes · 2023-05-27, 00:01

Quote:

Originally Posted by rogue

The bug with k=7000000 is also in fbncsieve. The code is to remove even k, but only when the base = 2. It was removing even k when the base is even. I will restart the range after I make this fix. Fortunately it hasn't cost me much time. I wanted to make a change to how I ran the range anyways.

The statement about it removing even k when the base is even is incorrect. It found many primes for even k for S586. It is only removing k's that are a multiple of the base when the -r switch is used. Of course for base 2 specifically, that's only the even k.

Therefore the removal of even k is not the problem with k=7000000.

Perhaps I misunderstood your intent here.

gd_barnes · 2023-05-27, 03:55

Serious issue for Riesel base 3:

srbsieve.ini:
base=3
mink=43300000
maxk=43400000
c=-1
maxNfbncsieve=3
maxKsrsieve2=100000
phase=60

Count of primes for n=1 output by fbncsieve and subsequently written to the pl_prime file: 2993

Actual primes for n=1 as proven by LLR: 1131

Number of composites in pl_prime file for n=1: 1862

The first "prime" in the file has a factor of 7.

Based on others comments in this thread, I think it's clear that fbncsieve has serious issues with small primes testing. It needs much stand-alone parallel testing comparing it to known good software done on its own before being incorporated into other software.

See attachments.

gd_barnes · 2023-05-27, 07:02

I wanted to see if the small primes issue popped up for S586 so I ran k = 1 to 10000 to n=60 with maxNfbncsieve=5. It did rear it's ugly head and an unrelated issue came up:

k=1 is showing up in the k's remaining file. It should be in the GFN file.

It's interesting that it went to the k's remaining instead of primes file. After all, 1*586^1+1 = 587 is prime. The small primes issue even bit us on a GFN.

For the small primes issue:

Primes missed:
3*586^1+1
1758*586^5+1
3516*586^3+1
9376*586^2+1

We were lucky enough in this case that all k's were quickly eliminated with somewhat larger primes.

Larger primes found:
3*586^6+1
1758*586^6+1
3516*586^7+1
9376*586^8+1

storm5510 · 2023-05-27, 23:25

I will step in this to ask a simple question: Within this environment, is there a way to get srsieve2 to use more than a single thread?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Useless SSE instructions	__HRB__	Programming	41	2012-07-07 17:43
Questions about software licenses...	WraithX	GMP-ECM	37	2011-10-28 01:04
Software/instructions/questions	gd_barnes	No Prime Left Behind	48	2009-07-31 01:44
Instructions to manual LLR?	OmbooHankvald	PSearch	3	2005-08-05 20:28
Instructions please?	jasong	Sierpinski/Riesel Base 5	10	2005-03-14 04:03

2023-05-27, 07:02	#410
gd_barnes "Gary" May 2007 Overland Park, KS 12,149 Posts	I wanted to see if the small primes issue popped up for S586 so I ran k = 1 to 10000 to n=60 with maxNfbncsieve=5. It did rear it's ugly head and an unrelated issue came up: k=1 is showing up in the k's remaining file. It should be in the GFN file. It's interesting that it went to the k's remaining instead of primes file. After all, 1586^1+1 = 587 is prime. The small primes issue even bit us on a GFN. For the small primes issue: Primes missed: 3586^1+1 1758586^5+1 3516586^3+1 9376586^2+1 We were lucky enough in this case that all k's were quickly eliminated with somewhat larger primes. Larger primes found: 3586^6+1 1758586^6+1 3516586^7+1 9376*586^8+1

2023-05-27, 23:25	#411
storm5510 Random Account Aug 2009 Not U. + S.A. 2⁴×3²×19 Posts	I will step in this to ask a simple question: Within this environment, is there a way to get srsieve2 to use more than a single thread?