mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Riesel Prime Search (https://www.mersenneforum.org/forumdisplay.php?f=59)
-   -   Team Sieve for 2995125705 (https://www.mersenneforum.org/showthread.php?t=1932)

SlashDude 2004-01-14 19:27

Team Sieve for 2995125705
 
2995125705 will be our 2nd 15k to team test!

Because of it's size, we need to use a different sieve program. Phil Carmody's 'K' sieve will work with this 15k!

To sieve with ksieve, you don't need any starting data files, just the program, and the command line options!

1. To start, download ksieve from [url]http://fatphil.org/maths/ksieve/[/url] (If you can't download it there, PM me) Extract and save ksieve2s.exe (ksieve2s for unix) from the downloaded file.

2. Next, select you range: (All ranges in billions)
(Reply to this thread when you have selected a range)

Done (n=230k-1M):
-p=1000000000000-1250000000000 - Kosmaj
-p=1250000000000-3250000000000 - SlashDude

*** 2nd sieving phase ***
Done (n=560k-1M):
-p=3250000000000-3500000000000 - lsoule
-p=3500000000000-3750000000000 - Kosmaj
-p=3750000000000-4000000000000 - lsoule
-p=4000000000000-4250000000000 - Kosmaj
-p=4250000000000-4500000000000 - lsoule
-p=4500000000000-4750000000000 - segmtfault
-p=4750000000000-5000000000000 - CedricVonck
-p=5000000000000-5250000000000 - lsoule
-p=5250000000000-5500000000000 - Kosmaj
-p=5500000000000-5750000000000 - lsoule

Done (n=588k-1M):
-p=5750000000000-6000000000000 - segmtfault
-p=6000000000000-6250000000000 - Kosmaj
-p=6250000000000-6500000000000 - lsoule
-p=6500000000000-6750000000000 - segmtfault
-p=6750000000000-7000000000000 - lsoule
-p=7000000000000-7250000000000 - segmtfault
-p=7250000000000-7500000000000 - lsoule

Recent progress (variuous ranges, p in bn, n_min in k, n_max=1M)
[code]
p_min p_max n_min Who Status
7500 7793 600 segmtfault Complete
7793 8000 620 segmtfault Complete
8000 8500 620 lsoule Complete
8500 9000 620 segmtfault Complete
9000 9500 620 lsoule Complete
9500 9610 620 segmtfault Complete
9610 19628 650 segmtfault Complete
19628 21500 840 grobie Complete
21500 23500 890 amphoria Complete
23500 25000 910 Joshua2 Complete
=== Sieving in the n<1M range stopped ===
[/code]
[B]Sieving in the n<1M range stopped. Moving to n=1-2M.[/B]
Current progress, [I]p[/I] in bn.
[CODE]
to 50 Kosmaj Complete
50-500 Joshua2 Complete
500-1000 Joshua2 In Progress
[/CODE]

Next command (updated on April 14, 2006)
ksieve2s -m -k=2995125705 -n=1000000-2000000 -p=1000000000000-1100000000000 -f=01000.del

To run [I]ksieve[/I] at low priority on Windows, prepend
[I]start /B /LOW [/I]to the above command line. To stop it use Ctrl-Break (not Ctrl-C).

3. From a command prompt, "CD" (Change Directory) to the location you saved ksieve, and enter the "Command Line" for your selected range!

4. When ksieve completes the range, you should have a file that lists your range, and has a .del extension. Please send this file to [EMAIL=kosmaj@yahoo.com]Kosmaj[/EMAIL] (Or attach it to this thread - zip it first if possible!)

Thanks again everyone for your help!!
-Mark & Kosmaj

SlashDude 2004-01-23 23:16

FYI-
Sieving is complete through 1 trillion!

I am waiting until we have a large range sieved before I merge all the ranges. I am looking at starting LLR testing in mid to late February.

ET_ 2004-01-23 23:35

[QUOTE=SlashDude]FYI-
Sieving is complete through 1 trillion!

I am waiting until we have a large range sieved before I merge all the ranges. I am looking at starting LLR testing in mid to late February.[/QUOTE]

can't wait...

Luigi :banana:

SlashDude 2004-02-06 23:56

Just a quick update.
The sieve range up to p=3,250,000,000,000 should be complete in about 10 days. (That's what I get for running it on 6 500MHz machines :) When this range is complete, I will compile the llr test .txt files for 15k=2995125705 for the group test!

SlashDude 2004-02-27 23:03

The sieving is complete to P=3,250,000,000,000. I am working on combining the sieve files, and creating the llr test files! We should have some test files soon!

SlashDude 2004-02-29 03:47

I've ran into a small problem merging the NewPGen files and the Ksieve files. I think the fastest way is to fix this is to resieve from p=1 to 309 billion using ksieve. This will take a few days, but I am working on it.

Sorry for the delay.

Kosmaj 2004-02-29 14:39

You don't have to sieve again. Just use the abcmerge.pl perl script. It is included in the distribution of ksieve. Before using it, on the first line of the ABC file produced by NewPgen replace the product you used to represent 2995125705 in NewPgen, most likely you used 15*199675047, into 2995125705. Then concatenate del files from all segments produced by ksieve into one huge del file. And finally run the script on these two files.

Kosmaj 2004-12-28 07:50

We'll reuse this thread to coordinate 2nd sieving phase of k=2995125705. There are only 3 ranges now but I can easily add more, for example to 5 trillion. We can also set n_min to a lower value. Since ranges to n=536,000 are reserved we can set n_min=540000. segmtfault is on vacation and I will not need another range for about 10 days, so it mostly depends wheter lsoule plans to reserve more ranges now or not. I'll mail ksieve to lsoule and pepe tonight.

em99010pepe 2004-12-28 18:52

I'm pepe.

Just testing and a few questions:

1º - I need to reserve a sieve range.

2º - I need to create a shortcut to ksieve2m.exe or ksieve2s.exe and put the following text on the command line: -m -k=2995125705 -n=560000-1000000 -p=3250000000000-3500000000000 -f=32503500.del

3º - Which one should I ran: ksieve2m.exe or ksieve2s.exe? My processor is an AMD 64 3000+.

4º - How do I change the program priority without going to the task manager? It's running in normal. It should be idle.

5º - When stopping the client why does it starts again the sieve range? Am I doing something wrong? When I stop PSP range the program restarts from the last point not from the beginning!

Carlos

Kosmaj 2004-12-29 08:55

Pepe/Carlos,
Some answers:
4) I don't know how to change the priority without going to the Task Manager. Maybe by some "launcher" style application, but I don' t use any.

5) When you stop and restart the client, the lower limit of factors has to be adjusted manually. For example if you work on 3250-3500 and you stop the client at p=3330... the "-p" switch has be adjusted to read
-p=3330000000000-3500000000000
It's the best to let the client run non-stop, or as long as possible.

If the behaviour above can cause you problems it's the best to reconsider running it. If you want you can test a range using LLR instead of sieving.

Answers to first 3 questions are easy. We use ksieve2s and you can reserve the first range, 3250-3500. The short cut can be created in the usual way, but whenever you restart the client the "-p" switch has to be modified.

SlashDude 2004-12-29 21:03

[QUOTE=em99010pepe]I'm pepe.

Just testing and a few questions:


4º - How do I change the program priority without going to the task manager? It's running in normal. It should be idle.

Carlos[/QUOTE]

To start it at low priority, use the following command line:
start /LOW ksieve2s.exe -m -k=2995125705 -n=560000-1000000 -p=3250000000000-3500000000000 -f=32503500.del

-Mark


All times are UTC. The time now is 12:38.

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