36M to 37M to 2^60 is done

eepiccolo · 2003-04-14, 18:16

Like the subject says, 36M to 37M to 2^59. I'm only going to 2^59 so I can get some idea of how long factoring takes. If I went to 2^60, it would take me three times as long :D .

Edit: 2^59 was only going to take a month, so I've bumped it up to 2^60, as is now reflected in the subject line.

eepiccolo · 2003-04-14, 19:00

Wow, found a factor on the second exponent!

eepiccolo · 2003-04-15, 12:01

Things are chugging right along, but I have a question, and I really hope someone can answer it. In the results.txt, I noticed that sometimes there is a timestamp for when a factor is found, but sometimes there is not. Is this normal?

S80780 · 2003-04-15, 14:14

This is just a guess. Did you find the factors in question nearly instantly after factoring started? If so, the reason for the lack of a timestamp could be that the last entry in results.txt was done under one sec. ago, so that both timestamps match. I noticed that benchmark results are sometimes split by a timestamp with a one sec. difference to the innicial one.

Benjamin

eepiccolo · 2003-04-15, 14:48

Quote:

Originally Posted by S80780

This is just a guess. Did you find the factors in question nearly instantly after factoring started? If so, the reason for the lack of a timestamp could be that the last entry in results.txt was done under one sec. ago, so that both timestamps match. I noticed that benchmark results are sometimes split by a timestamp with a one sec. difference to the innicial one.

Benjamin

Posibly that is it, but the main reason I'm concerned is I don't want there to be a problem when George loads the results into the database. I want to make sure these missing timestamps won't cause a problem.

S80780 · 2003-04-15, 15:36

I don't think so. I sent George my results without the timestamps from the very beginning of my LMH-work. He never complaint.

eepiccolo · 2003-06-11, 13:48

I'm halfway done :) . Here are my results so far.

[code:1] range # found 58-59 59-60 % factored % expected
36.0M-36.1M 93 40 53 3.582% 3.419%
36.1M-36.2M 76 44 32 2.986% 3.419%
36.2M-36.3M 83 46 37 3.148% 3.419%
36.3M-36.4M 91 42 49 3.501% 3.419%
36.4M-36.5M 100 49 51 3.885% 3.419%

36.0M-36.5M 443 221 222 3.411% 3.419%[/code:1]
It's kind of interesting that the number of factors found per 100,000 block was more stable for 2^58 - 2^59 than for 2^59 - 2^60. Of course, this is just a small sample. We'll have to see what happens for 36.5M to 37M

eepiccolo · 2003-08-01, 17:03

Alright, I'm done :) :) :) . Here are the rest of my results.

[code:1] range # found 58-59 59-60 % factored % expected
36.5M-36.6M 71 42 29 2.706% 3.419%
36.6M-36.7M 70 35 35 2.693% 3.419%
36.7M-36.8M 77 50 27 3.003% 3.419%
36.8M-36.9M 85 43 42 3.307% 3.419%
36.9M-37.0M 69 38 31 2.704% 3.419%

36.0M-37.0M 815 429 386 3.146% 3.419%[/code:1]

So, that's that. There sure were a lot less factors in the second half. Oh well.

2003-04-14, 18:16	#1
eepiccolo Dec 2002 Frederick County, MD 2×5×37 Posts	36M to 37M to 2^60 is done Like the subject says, 36M to 37M to 2^59. I'm only going to 2^59 so I can get some idea of how long factoring takes. If I went to 2^60, it would take me three times as long :D . Edit: 2^59 was only going to take a month, so I've bumped it up to 2^60, as is now reflected in the subject line.

2003-04-14, 19:00	#2
eepiccolo Dec 2002 Frederick County, MD 2×5×37 Posts	Wow, found a factor on the second exponent!

2003-04-15, 12:01	#3
eepiccolo Dec 2002 Frederick County, MD 2×5×37 Posts	Things are chugging right along, but I have a question, and I really hope someone can answer it. In the results.txt, I noticed that sometimes there is a timestamp for when a factor is found, but sometimes there is not. Is this normal?

2003-04-15, 14:14	#4
S80780 Jan 2003 far from M40 5³ Posts	This is just a guess. Did you find the factors in question nearly instantly after factoring started? If so, the reason for the lack of a timestamp could be that the last entry in results.txt was done under one sec. ago, so that both timestamps match. I noticed that benchmark results are sometimes split by a timestamp with a one sec. difference to the innicial one. Benjamin

2003-04-15, 15:36	#6
S80780 Jan 2003 far from M40 5³ Posts	I don't think so. I sent George my results without the timestamps from the very beginning of my LMH-work. He never complaint.

2003-06-11, 13:48	#7
eepiccolo Dec 2002 Frederick County, MD 2·5·37 Posts	I'm halfway done :) . Here are my results so far. [code:1] range # found 58-59 59-60 % factored % expected 36.0M-36.1M 93 40 53 3.582% 3.419% 36.1M-36.2M 76 44 32 2.986% 3.419% 36.2M-36.3M 83 46 37 3.148% 3.419% 36.3M-36.4M 91 42 49 3.501% 3.419% 36.4M-36.5M 100 49 51 3.885% 3.419% 36.0M-36.5M 443 221 222 3.411% 3.419%[/code:1] It's kind of interesting that the number of factors found per 100,000 block was more stable for 2^58 - 2^59 than for 2^59 - 2^60. Of course, this is just a small sample. We'll have to see what happens for 36.5M to 37M

2003-08-01, 17:03	#8
eepiccolo Dec 2002 Frederick County, MD 2×5×37 Posts	Alright, I'm done :) :) :) . Here are the rest of my results. [code:1] range # found 58-59 59-60 % factored % expected 36.5M-36.6M 71 42 29 2.706% 3.419% 36.6M-36.7M 70 35 35 2.693% 3.419% 36.7M-36.8M 77 50 27 3.003% 3.419% 36.8M-36.9M 85 43 42 3.307% 3.419% 36.9M-37.0M 69 38 31 2.704% 3.419% 36.0M-37.0M 815 429 386 3.146% 3.419%[/code:1] So, that's that. There sure were a lot less factors in the second half. Oh well.