Hi all, I've tried to just get everyone once, hope that's how it's worked out. I graduated today and am losing my Cambridge email address, from now on please ...
For the best of all possible reasons, Mike. (My younger son just checked out of his course, with first class honours, and must now work out whether his...
Is there something special about Phi(3673,8185) that causes false composite results in bases 5, 11, 17, 31, 37 and 67 (and maybe others) unless -a1 or -a2 is...
My output: Microsoft(R) Windows NT(TM) (C) Copyright 1985-1996 Microsoft Corp. U:\>p:\pfgw\pfgw -qPhi(3673,8185) -b5 PFGW Version 20021217.Win_Dev (Beta...
In the last three days I have had three false composite results out of pfgw -a1 while verifying submission to the prime pages. All submission go through a...
I would certainly NOT use the v22.12 version for primality proving at ALL on a P4 system. There was a serious bug in the code when it came to the proving...
Prompted by Jim's comments that pfgw is not "safe" on Linux I'm trying to run the Windows version using wine. I'm now running the command line version in PRP...
... I think you misinterpreted my comments. It is not that it is not "safe", it is that I have not tested it much (I do have one Linux system running proth...
I hope soon we should see new processor Athlon 64. After a while, approximately in the beginning of 2004, I think all of us shall have some results of his...
Hi all, If you have a small number (less than 2^31, I would say from a quick look at the source) and you test it using -tm, -tp, or -tc, even if it is prime ...
There is one more problem I found recently. On 2 machines I run the command line version of pfgw using the Task Scheduler (Win-2k), I set it to start at a...
I will look at this, but it is VERY possible, that there is nothing to do (other than find a different "scheduling" mechanism). The buildin scheduler is not...
Jim - over the weekend I found another little bug in the 'Lucky Minus' file handling code. I'll fix that and send the corrected file to you. Regards, Paul. ...
FYI, I have fixed this bug and the one that I saw last week (-t[pmc] with small candidates). I hope that didn't duplicate work, but it was only a three line...
Thanks for your reply. Note that there are two issues here: 1) Command line client doesn't catch the scheduller's kill signal and suddenly stops without...
I have not had a chance to look at this problem (yet). The scheduler handles command line apps by killing the CMD console, and not the app. There is no...
An example of numbers I'm working on right now: PRP: 121088*13^121088-1 85000/448095 requires between 90 minutes of fast P-4's and 200 minutes on slow P-3's....
... I'm just looking at the code, and although it pulls the power of 2 out (as if it was going to use it for a SPRP test), if you run it it does a fermat PRP ...
I have had a look at the code and can see no good reason why b is limited to the range 2 to 255. To allow values up to 2^31 is a simple two line change: ...
... The prime95 fft routines allow multiplying by a small constant in the rounding and carry propagation step. This is much cheaper than a separate step after...
Thanks George. I was pretty sure this was the case, but was not exactly sure (in the PRP since). Jim. ... limited to ... the ... separate ... Anyway, Chris ...
This is just a passing thought, but since the PFGW executable is so *huge*, I routinely compress it using the UPX utility (see http://upx.sourceforge.net/). ...
Changes were made in the code that chooses the FFT size due to just this type of problem occuring with certain numbers in the 1.1 release. Choosing the correct...
This error has nothting to do with those changes. This error was "added" since 1.1 In the code for PRP, this code "existed". It is a "quick-check" to see if...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
