Old sr1sieve
Old sr1sieve
If you look at stat from sieving process you will find that sr1sieve version is older then new one ( latest is 1.45)
And in 1.43 ( you use 1.41) is written
Version 1.4.3: (Posted September 12, 2012)
Rebuilt Win64 with latest MinGW. Previous builds could report invalid
factors (which are rejected by the code) due to a miscompile by MinGW.
So maybe you need to use latest version od sr1sieve.
And in 1.43 ( you use 1.41) is written
Version 1.4.3: (Posted September 12, 2012)
Rebuilt Win64 with latest MinGW. Previous builds could report invalid
factors (which are rejected by the code) due to a miscompile by MinGW.
So maybe you need to use latest version od sr1sieve.
Re: Old sr1sieve
I´m using an script to check for false factors.
I got those bins from geoffreys Google site (outdated), just got the binaries from an other site (v1.4.5). Anyway, we only got the bins for Win.
I got those bins from geoffreys Google site (outdated), just got the binaries from an other site (v1.4.5). Anyway, we only got the bins for Win.
Re: Old sr1sieve
http://www.bc-team.org/downloads.php?vi ... ece6bbca12 : link for sr1sieve 1.4.5
If you need sr1sieve for linux just run Makefile inside this file without any modification: it will be built without any problems
If you need sr1sieve for linux just run Makefile inside this file without any modification: it will be built without any problems
Re: Old sr1sieve
is there any valid reason why you left 1.41 for Windows and make 1.45 for Linux? In source code is valid 1.45 for Windows app, so you need just to copy to your server. It is not so hard work?
Re: Old sr1sieve
It's not just copy it to the server.
We did testing before, we checked if results of Windows are identical to results of Linux, so both can be validated against each other.
So releasing a new version has to be very solid to be sure everything is working, not only computing, also validation, assimilation and the autmatic handling on the backend.
We did testing before, we checked if results of Windows are identical to results of Linux, so both can be validated against each other.
So releasing a new version has to be very solid to be sure everything is working, not only computing, also validation, assimilation and the autmatic handling on the backend.
Re: Old sr1sieve
For me it is logical to use same version from same source code.. There is no need to test anything since that code and that exe file for windows are in use for many years. I am sure that error will be found in meantime,and since there is no error there is no reason for not use that version . You use 1.41 for Windows and 1.45 for Linux, use 1.45 for both OS. It was tested all those years that code is online. 1.41 has some errors , 1.45 has no errors.yoyo hat geschrieben:It's not just copy it to the server.
We did testing before, we checked if results of Windows are identical to results of Linux, so both can be validated against each other.
So releasing a new version has to be very solid to be sure everything is working, not only computing, also validation, assimilation and the autmatic handling on the backend.
But it is your server and your project.
Re: Old sr1sieve
As said, it is not only the exe, it ist also the Server side and Handling on the Backend which has to fit. And this nobody tested, this is unique on my Server.