Additional sieving for manual testing

rogue · 2006-09-29, 18:24

Taking 217592. In case you are wondering I have a number of PCs without Internet access. This one appears to be beyond LLRNet limits for a while, so it shouldn't cause any problems.

For my current status, this is the current breakdown for my k:
390000 < 1
350000 < 2 < 370000
300000 < 2 < 320000
250000 < 1 < 270000
200000 < 1 < 210000
150000 < 2 < 180000
1 < 150000

The ones < 200000 are very dense, so it will be a while before they hit 200000, unless they find a prime before then. As for the rest of the k I have reserved, I have been completely stumped.

I think that I have spent over 6 Ghz/years on these k.

If I release any of these k before finding a prime, I will report the residues.

masser · 2006-09-29, 20:21

rogue,

Are you using srsieve to sieve just your reservations? I'm getting good sieving speeds (125000 p/sec) on my slowest machines (650 Mhz) using a file with only my 6 candidates. If you're not already doing this, it might be worthwhile to devote one or more of your slower machines to sieving just the candidates you have reserved.

masser

rogue · 2006-09-29, 20:58

Quote:

Originally Posted by masser

rogue,

Are you using srsieve to sieve just your reservations? I'm getting good sieving speeds (125000 p/sec) on my slowest machines (650 Mhz) using a file with only my 6 candidates. If you're not already doing this, it might be worthwhile to devote one or more of your slower machines to sieving just the candidates you have reserved.

masser

No, I hadn't considered that.

michaf · 2006-09-29, 21:47

Isn't our combined sieving speed faster then that? On _all_ candidates.
(I get 60000 p/sec alone, combined I have no idea what we are doing, but I'm sure it's more then double that)

So downloading the sr5data.txt file, and filtering out your candidates might be more efficient.

Unless of course, you already have sieved out far ahead of the combined sieve.
And unless I'm missing something here too :>

axn · 2006-10-03, 20:22

Moved out the posts from reservation thread.

axn · 2006-10-03, 20:23

Quote:

Originally Posted by rogue

If I release any of these k before finding a prime, I will report the residues.

I would suggest that you send the residues regardless. We'll setup the stats, prime or no prime

rogue · 2006-10-03, 23:15

Quote:

Originally Posted by axn1

I would suggest that you send the residues regardless. We'll setup the stats, prime or no prime

I can do that.

BTW, sieving was marginally better than PRP testing, but not enough for me to spend a lot of time on it. I estimated that it would remove at about twice the speed of PRP testing, although I didn't run it for a number of hours to verify that result. I only ran it for about 15 minutes.

One more thing, I do take the .dat file every week or so and extract a file for each k and update each LLR instance, so I am still taking advantage of every else's sieving efforts.

axn · 2006-10-03, 23:27

Oh, and one more thing. If you are doing your own sieving, don't forget to send in the factors also. Otherwise we'll have holes lying around

rogue · 2006-10-04, 20:37

To whom do I mail the residues?

axn · 2006-10-04, 21:41

Quote:

Originally Posted by rogue

To whom do I mail the residues?

Same as the one used for regular sieving.

2006-09-29, 20:21	#2
masser Jul 2003 wear a mask 11001111010₂ Posts	rogue, Are you using srsieve to sieve just your reservations? I'm getting good sieving speeds (125000 p/sec) on my slowest machines (650 Mhz) using a file with only my 6 candidates. If you're not already doing this, it might be worthwhile to devote one or more of your slower machines to sieving just the candidates you have reserved. masser Last fiddled with by masser on 2006-09-29 at 20:22

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Manual Testing CPU	Unregistered	Information & Answers	6	2019-12-17 23:35
Manual Testing ECM on cofactors?	yih117	PrimeNet	24	2018-02-03 15:46
Manual testing on CPU list	Fred	PrimeNet	3	2016-02-12 02:49
Using Prime95 for manual testing	opyrt	Prime Sierpinski Project	10	2009-08-04 22:32
Manual Testing	Unregistered	Information & Answers	2	2009-01-17 14:13

2006-09-29, 18:24	#1
rogue "Mark" Apr 2003 Between here and the 11·577 Posts	Taking 217592. In case you are wondering I have a number of PCs without Internet access. This one appears to be beyond LLRNet limits for a while, so it shouldn't cause any problems. For my current status, this is the current breakdown for my k: 390000 < 1 350000 < 2 < 370000 300000 < 2 < 320000 250000 < 1 < 270000 200000 < 1 < 210000 150000 < 2 < 180000 1 < 150000 The ones < 200000 are very dense, so it will be a while before they hit 200000, unless they find a prime before then. As for the rest of the k I have reserved, I have been completely stumped. I think that I have spent over 6 Ghz/years on these k. If I release any of these k before finding a prime, I will report the residues.

2006-09-29, 21:47	#4
michaf Jan 2005 479 Posts	Isn't our combined sieving speed faster then that? On _all_ candidates. (I get 60000 p/sec alone, combined I have no idea what we are doing, but I'm sure it's more then double that) So downloading the sr5data.txt file, and filtering out your candidates might be more efficient. Unless of course, you already have sieved out far ahead of the combined sieve. And unless I'm missing something here too :>

2006-10-03, 20:22	#5
axn Jun 2003 11675₈ Posts	Moved out the posts from reservation thread.

2006-10-03, 23:27	#8
axn Jun 2003 31·163 Posts	Oh, and one more thing. If you are doing your own sieving, don't forget to send in the factors also. Otherwise we'll have holes lying around

2006-10-04, 20:37	#9
rogue "Mark" Apr 2003 Between here and the 11×577 Posts	To whom do I mail the residues?