mersenneforum.org  

Go Back   mersenneforum.org > Other Stuff > Archived Projects > NFSNET Discussion

 
 
Thread Tools
Old 2009-10-08, 16:55   #12
bdodson
 
bdodson's Avatar
 
Jun 2005
lehigh.edu

210 Posts
Default

Quote:
Originally Posted by R.D. Silverman View Post
I can provide lattice siever code. GGNFS is available......
You once used my (line) siever code before switching to CWI.

Allow me to ask: how will you do 10,269- without participants?
This number is quite difficult.
Sieving is in progress here, with the 16e siever. Region has width
200M-20M, of which 50M or so is complete. I'm keeping the number
of cores between 200-300, with up to 100 small memory cores (1Gb)
running Batalov-Dodson numbers (3, 521+ due tomorrow). -Bruce

PS - page 112 looks to be full, at 30 entries; Serge reports that
the first-five are already updated. Sam's also upated the progress
on the wanted lists from page 111.

Last fiddled with by bdodson on 2009-10-08 at 17:02 Reason: wanteds
bdodson is offline  
Old 2009-10-08, 17:09   #13
R.D. Silverman
 
R.D. Silverman's Avatar
 
Nov 2003

22·5·373 Posts
Default

Quote:
Originally Posted by bdodson View Post
Sieving is in progress here, with the 16e siever. Region has width
200M-20M, of which 50M or so is complete. I'm keeping the number
of cores between 200-300, with up to 100 small memory cores (1Gb)
running Batalov-Dodson numbers (3, 521+ due tomorrow). -Bruce

PS - page 112 looks to be full, at 30 entries; Serge reports that
the first-five are already updated. Sam's also upated the progress
on the wanted lists from page 111.
I find it curious that over time the number of entries per page has shrunk...
There used to be near 60 entries/page.
R.D. Silverman is offline  
Old 2009-10-08, 18:11   #14
bdodson
 
bdodson's Avatar
 
Jun 2005
lehigh.edu

210 Posts
Default

Quote:
Originally Posted by R.D. Silverman View Post
I find it curious that over time the number of entries per page has shrunk...
There used to be near 60 entries/page.
Maybe due to smaller numbers? It's just the most recent
2 or 3 pages that seem to have stopped near 30 entries;
back at page 90 there were 40 entries. There has already
been a lot of activity on wanted and/or first-fives; so perhaps
I'm premature on the page closing, due to wishful thinking.

-Bruce
bdodson is offline  
Old 2009-10-08, 18:33   #15
R.D. Silverman
 
R.D. Silverman's Avatar
 
Nov 2003

22·5·373 Posts
Default

Quote:
Originally Posted by bdodson View Post
Maybe due to smaller numbers? It's just the most recent
2 or 3 pages that seem to have stopped near 30 entries;
back at page 90 there were 40 entries. There has already
been a lot of activity on wanted and/or first-fives; so perhaps
I'm premature on the page closing, due to wishful thinking.

-Bruce
I think it driven by the need to have everything fit on an 8.5" x 11"
*printed* page. Many of the numbers have to be broken into 2 lines.
R.D. Silverman is offline  
Old 2009-10-09, 14:12   #16
Wacky
 
Wacky's Avatar
 
Jun 2003
The Texas Hill Country

32×112 Posts
Default

Quote:
Originally Posted by R.D. Silverman View Post
I can provide lattice siever code. GGNFS is available......
You once used my (line) siever code before switching to CWI.
As I stated over in the NFS@Home thread, the issue is not access to source code for a lattice siever. The issue is access to development platforms and programmers who have the time, tools, and ability to debug the porting of a common protocol to the various platforms.

I have a "NFSNet" version of the siever siever running on Mac OSX. However, my largest contributor does not have a "compatible" version on 64-bit Linux. Nor do we have a version for any form of Windows.

Quote:
Allow me to ask: how will you do 10,269- without participants?
This number is quite difficult.
The same way that Tom does things -- a very non-NFSNet collaboration of a few contributors (with access to many machines).

I find it interesting that NFSNet is continually "faulted" because of shortcomings when competing efforts are applauded even though they have those same shortcomings.
Wacky is offline  
Old 2009-10-14, 11:22   #17
fivemack
(loop (#_fork))
 
fivemack's Avatar
 
Feb 2006
Cambridge, England

2·3·29·37 Posts
Default

I think the way that I do things gets a rather different set of collaborators than NFSnet or NFS@home can manage: in particular, I imagine that the administrators of large clusters with idle time and with batch-submission interfaces are generally much happier with users running scripts which call executables to do a fairly well-defined job than with users running clients that collect their own work over the Internet.

Certainly I would not be happy to run NFS@home on the machines here at the office on which I sometimes run gnfs-lasieve4I16e.

I know that my approach is much inferior in terms of getting really large amounts of compute time to fully-automated systems running on many home PCs, but the activation energy to doing it my way is much lower, and something like the way I do it is necessary to exploit the set of machines that I get to use.
fivemack is offline  
Old 2009-10-31, 01:26   #18
Wacky
 
Wacky's Avatar
 
Jun 2003
The Texas Hill Country

44116 Posts
Default

Tom,

I understand your comments about "effort", and the "constraints" on comfortable participation.

My only regret is that we cannot all come together and produce a protocol that provides a common "format" for the allocation and reporting of results.

This protocol would provide a uniform method of problem description, and a uniform format for the reporting the results. This reporting should be done in a manner that allows the easy extraction of a summary of the sieving without transmitting ALL of the details of the relations found.

NFS@Home has effectively replaced NFSNet because Greg (at least thinks that he) has the resources to handle thousands of participants on a single central server. NFSNet did not utilize that approach because we lacked the resources and also wished to have a "fall bacK" protocol that would compensate for a failure at any server node within the system.
Wacky is offline  
Old 2010-02-20, 00:54   #19
Wacky
 
Wacky's Avatar
 
Jun 2003
The Texas Hill Country

32·112 Posts
Default Calling Don Leclair

Does anyone know how to reach Don?

it has been some time since I have been on contact with him.

Richard
Wacky is offline  
Old 2010-02-20, 23:43   #20
dleclair
 
dleclair's Avatar
 
Mar 2003

7·11 Posts
Default

Hi Richard,

Very nice to hear from you. I'll send you my current e-mail address in a PM.

-Don
dleclair is offline  
 

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
strange thing happening in factordb firejuggler FactorDB 96 2019-12-06 15:32
Is there much happening on 1k-10k ? fivemack Aliquot Sequences 5 2010-02-05 15:47

All times are UTC. The time now is 06:11.


Sun Oct 17 06:11:30 UTC 2021 up 86 days, 40 mins, 0 users, load averages: 0.93, 0.92, 0.93

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.