mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   And now for something completely different (https://www.mersenneforum.org/forumdisplay.php?f=119)
-   -   Carol / Kynea Coordinated Search - Reservations/Status (https://www.mersenneforum.org/showthread.php?t=21216)

gd_barnes 2018-01-12 05:22

[QUOTE=wombatman;477317]Lucky for you I'm taking tomorrow off from work, so here's a Windows version compiled with the generic x86-64 GCC flags. Try it out and see if it works for you. :smile:[/QUOTE]

Excellent! It works! Thank you! :grin:

Perhaps Mark could post this more generic version on his site. :smile:

Mark, I will reserve bases 56 and 58 to n=10K. I'll post primes in the other thread.

wombatman 2018-01-12 06:35

[QUOTE=gd_barnes;477318]Excellent! It works! Thank you! :grin:

Perhaps Mark could post this more generic version on his site. :smile:

Mark, I will reserve bases 56 and 58 to n=10K. I'll post primes in the other thread.[/QUOTE]

Very welcome! :tu:

rogue 2018-01-12 14:09

[QUOTE=gd_barnes;477299]Make or Makefile are not executables so trying to execute them does not work. When I type "make" or "makefile" at the command prompt it just says:
[code]
'make' is not recognized as an internal or external command, operable program or batch file.
[/code]

What does "mingw64" mean? I have not compiled C programs in Windows/Linux before so you are going to have to be very basic here.[/QUOTE]

mingw64 is an application which give you access to many of the unix command line programs that you are familiar with from Linux, such as grep and diff. It also has gcc/g++, a compiler for C/C++ code. make is another program that is bundled with mingw64 that will build libraries or executables using a makefile, which has the switches and other settings needed to build said library/executable.

mingw64 is free to d/l and install and everything that comes with it is free to use. You just need to add mingw64's bin directory to the PATH environment variable.

For your Linux distro you would need to install the packages that include make and gcc/g++ then you can build.

rogue 2018-01-12 14:13

[QUOTE=wombatman;477317]Lucky for you I'm taking tomorrow off from work, so here's a Windows version compiled with the generic x86-64 GCC flags. Try it out and see if it works for you. :smile:[/QUOTE]

I assume you modified the makefile. Let me know what compile and link switches you used I would guess that you removed "-march=native". In most of my software I avoid using -march as much as possible as it reduces cross-platform compatibility, but not in cksieve as I started with a makefile used by another program. I will eventually make that multi-threaded and when I do it will not use that switch.

wombatman 2018-01-12 16:11

[QUOTE=rogue;477327]I assume you modified the makefile. Let me know what compile and link switches you used I would guess that you removed "-march=native". In most of my software I avoid using -march as much as possible as it reduces cross-platform compatibility, but not in cksieve as I started with a makefile used by another program. I will eventually make that multi-threaded and when I do it will not use that switch.[/QUOTE]

I changed from "native" to "x86-64" as that appeared to be the most generalized 64-bit flag I could use. That was the only change and it compiled under MinGW and the Win10 Ubuntu bash shell without any other modification needed.

The GLIBC error Gary got appears to be due to using the older version of Ubuntu. From a quick search around, it seems that one simply can't get the appropriate version of GLIBC unless you compile it yourself for the older distros.

Edit: And I set the ARCH flag to x86-64, not sse.

gd_barnes 2018-01-12 22:31

Base 56 is complete to n=10K; 3 new primes were found for n=1000-10K; base released.

Base 58 is complete to n=10K; 3 new primes were found for n=1000-10K; base released.

Reserving base 202 to n=10K.

gd_barnes 2018-01-14 10:04

Base 202 is complete to n=10K; 11 new primes were found for n<=10K; base released.

All bases <= 204 are now complete to n=10K. :-)

gd_barnes 2018-01-17 01:54

For historical reference:

All bases shown on the primes page have been doublechecked to n=2500.

Base 22 has been doublechecked to n=10K.

PRP tests have been run on primes n>2500 for all bases.

All corrections have been made.

This includes all bases <= 204, 214, 218, 220, 222, 228, 252, 278, 290, 316, 326, 640, 688, 720, 1656, 1968, 2010, 2026, 2264, 2482, and 2634.

gd_barnes 2018-01-17 03:43

Reserving base 34 to n=30K.

I will doublecheck it to n=10K.

gd_barnes 2018-01-24 06:52

1 Attachment(s)
Base 34 is complete to n=30K. No primes were found for n=10K-30K.

This was unexpected. It is the largest primeless gap for b<38 and n<=30K. It is a fairly heavy weight base and the last prime was at n=8093. For that reason I'm attaching results.

Base 34 was also doublechecked to n=10K. No problems found.

sweety439 2018-01-24 20:07

[QUOTE=gd_barnes;478215]Base 34 is complete to n=30K. No primes were found for n=10K-30K.

This was unexpected. It is the largest primeless gap for b<38 and n<=30K. It is a fairly heavy weight base and the last prime was at n=8093. For that reason I'm attaching results.

Base 34 was also doublechecked to n=10K. No problems found.[/QUOTE]

Don'e forget, the bases b<38 include bases 4, 8, 16, 32 and 36, although they are not shown in the webpage.


All times are UTC. The time now is 22:03.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.