Distributed sieving discussions - Page 2

axn · 2006-06-01, 18:42

Quote:

Originally Posted by michaf

I'd say the ABCD format would be nice for space-savings, but then,
20Mb is doable these days (not to mention that the file will be getting
smaller every 5 seconds :>)

Unfortunately, the files are hosted under a free geocities account and the d/l limit is about 4mb / hr

Quote:

Originally Posted by michaf

How much candidates are left in total? as in how much k/n pairs?

About 10 million k/n pairs are left currently (351 k's).

Quote:

Originally Posted by michaf

Is the human-readability of any relevance? The formats are easily converted to one another after all.

I would say, not much! Ability to edit the file to remove a k will be valuable. But that could be done by a simple utility.

michaf · 2006-06-01, 18:53

Iirc I have a 20Mb homepage to my avail :>
I'd be happy to host it
If the file is about 10Mb I would be fine, 10M pairs will be 20Mbyte uncompressed, 4-5 Mb compressed, that should do it

No idea on the data-transfer limits though... I will need to check on that first, before I promise anything...

axn · 2006-06-01, 19:02

Quote:

Originally Posted by michaf

Iirc I have a 20Mb homepage to my avail :>
I'd be happy to host it
If the file is about 10Mb I would be fine, 10M pairs will be 20Mbyte uncompressed, 4-5 Mb compressed, that should do it

No idea on the data-transfer limits though... I will need to check on that first, before I promise anything...

Any and all help much appreciated

michaf · 2006-06-01, 20:12

Just found out that I have a data-limit of 50Mb per 24 hours.
If I surpass that, it will just become unvisible to the public.
(No idea if you can surpass that with a direct link to the file though :>)
if it is 5 Mb for the file, it is worth it, no more then 10 downloads a day, but it should suffice; I'd be able to put factor-found files with it, so one doesn't need to dl the whole bunch all the time.

Now you try and get the starting size a tad smaller :P

axn · 2006-06-01, 20:39

Quote:

Originally Posted by michaf

Now you try and get the starting size a tad smaller :P

michaf · 2006-06-01, 21:07

But seriously, when we get to 70k for the riesels, we should start the sieving effort.

In the meantime, there should come
* the ABCD-format option on srsieve and srfile to reduce file-size
* speedups on srsieve if at all possible :>
* more riesels/sierpinskies eliminated

On the less-serious side (though hopefully not so in the future):
erm... I am testing with sieving n=2M-3M right now...
Sieved to 6M and ''only'' 7.09M candidates left
Now if we only got them tested upto 2M already :>

jasong · 2006-06-02, 00:47

Quote:

Originally Posted by michaf

Just found out that I have a data-limit of 50Mb per 24 hours.
If I surpass that, it will just become unvisible to the public.
(No idea if you can surpass that with a direct link to the file though :>)
if it is 5 Mb for the file, it is worth it, no more then 10 downloads a day, but it should suffice; I'd be able to put factor-found files with it, so one doesn't need to dl the whole bunch all the time.

Now you try and get the starting size a tad smaller :P

10 downloads is a little low, in my opinion. Perhaps we should consider other possibilities. I've come up with a few:

An irc bot that gives you the file(or manual transfer by contacting someone with the file and asking for a transfer)

aol IM file ransfer, and lastly

a torrent file. This last one is my favorite idea, since people could download it whenever. It's only a 2-3 minute thing to install the BitTorrent program, and you never know when it might be useful.

What do you guys think?

Citrix · 2006-06-02, 01:58

http://www.dropload.com/

This might be useful for transfering large files.

geoff · 2006-06-02, 03:48

Version 0.1.12 of srsieve will write an ABCD file when the --abcd switch is used. I sieved 353 sr5 candidates to p=10 million which left about 14 million terms remaining, and the ABCD file came to 42MB uncompressed, 10.5MB compressed with gzip -9, and 9.3MB compressed with bzip2 -9.

I also added a --checkpoint switch which when used in conjunction with the --factors switch will write just a one line checkpoint.txt file instead of a full sieve file. If you have the original sieve file plus the found factors plus the checkpoint file then you have all the information to recreate the new sieve file with srfile if necessary.

The srfile utility should get built along with srsieve, it can convert between any format srsieve writes, and can remove the factors in srfactors.txt from the sieve.

geoff · 2006-06-02, 04:30

Will the distributed sieve start from n=0? Is 2 million the best upper bound?

Do we have the tools to record who found the factors, perhaps for use in a future scoring system?

Should we be collecting prp/llr residues now?

michaf · 2006-06-02, 17:28

Torrent would be practical if someone has his computer connected 24/7, a lot of people do these days though :>

Dropload hasn't worked perfectly for me on some occasions, but it did do the job on most.

irc isn't my favorite, but certainly a nice possibility

Best would be to have someone who is able to host some bigger files and have the bandwith to go with it :>

The sieve would start from where the prp-ing left off, and eliminating the newly prp-d numbers every now and then. Same as when a new prime is found --> eliminate that one from the sieve

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Soapbox Discussions	only_human	Soap Box	41	2019-11-16 15:46
Our discussions here: how can we improve things?	Brian-E	Soap Box	105	2013-11-10 12:26
n=390000 discussions and old reservations	Mini-Geek	Twin Prime Search	31	2010-05-22 23:13
Primegrid discussions	pacionet	Twin Prime Search	17	2007-01-20 11:22
Automated PRP discussions	ltd	Sierpinski/Riesel Base 5	20	2006-09-02 22:19

2006-06-01, 18:53	#13
michaf Jan 2005 111011111₂ Posts	Iirc I have a 20Mb homepage to my avail :> I'd be happy to host it If the file is about 10Mb I would be fine, 10M pairs will be 20Mbyte uncompressed, 4-5 Mb compressed, that should do it No idea on the data-transfer limits though... I will need to check on that first, before I promise anything...

2006-06-01, 20:12	#15
michaf Jan 2005 1DF₁₆ Posts	Just found out that I have a data-limit of 50Mb per 24 hours. If I surpass that, it will just become unvisible to the public. (No idea if you can surpass that with a direct link to the file though :>) if it is 5 Mb for the file, it is worth it, no more then 10 downloads a day, but it should suffice; I'd be able to put factor-found files with it, so one doesn't need to dl the whole bunch all the time. Now you try and get the starting size a tad smaller :P

2006-06-01, 21:07	#17
michaf Jan 2005 479 Posts	But seriously, when we get to 70k for the riesels, we should start the sieving effort. In the meantime, there should come * the ABCD-format option on srsieve and srfile to reduce file-size * speedups on srsieve if at all possible :> * more riesels/sierpinskies eliminated On the less-serious side (though hopefully not so in the future): erm... I am testing with sieving n=2M-3M right now... Sieved to 6M and ''only'' 7.09M candidates left Now if we only got them tested upto 2M already :>

2006-06-02, 01:58	#19
Citrix Jun 2003 62E₁₆ Posts	http://www.dropload.com/ This might be useful for transfering large files.

2006-06-02, 03:48	#20
geoff Mar 2003 New Zealand 13·89 Posts	Version 0.1.12 of srsieve will write an ABCD file when the --abcd switch is used. I sieved 353 sr5 candidates to p=10 million which left about 14 million terms remaining, and the ABCD file came to 42MB uncompressed, 10.5MB compressed with gzip -9, and 9.3MB compressed with bzip2 -9. I also added a --checkpoint switch which when used in conjunction with the --factors switch will write just a one line checkpoint.txt file instead of a full sieve file. If you have the original sieve file plus the found factors plus the checkpoint file then you have all the information to recreate the new sieve file with srfile if necessary. The srfile utility should get built along with srsieve, it can convert between any format srsieve writes, and can remove the factors in srfactors.txt from the sieve.

2006-06-02, 04:30	#21
geoff Mar 2003 New Zealand 13×89 Posts	Will the distributed sieve start from n=0? Is 2 million the best upper bound? Do we have the tools to record who found the factors, perhaps for use in a future scoring system? Should we be collecting prp/llr residues now?

2006-06-02, 17:28	#22
michaf Jan 2005 479 Posts	Torrent would be practical if someone has his computer connected 24/7, a lot of people do these days though :> Dropload hasn't worked perfectly for me on some occasions, but it did do the job on most. irc isn't my favorite, but certainly a nice possibility Best would be to have someone who is able to host some bigger files and have the bandwith to go with it :> The sieve would start from where the prp-ing left off, and eliminating the newly prp-d numbers every now and then. Same as when a new prime is found --> eliminate that one from the sieve