McLaughlin, James on Mon, 08 Dec 2003 19:02:30 +0100

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

RE: bnrisprincipal to slow

Actually, I was running it on 2 machines. 
My home machine running version 2.2.6 completed the operation sometime during the night and took something over 3 days.
I was also running it on a computer at my old university, which has version 2.1.4 installed system wide and this is the one where the computations are in the 9th day.

	-----Original Message----- 
	From: McLaughlin, James 
	Sent: Mon 12/8/2003 5:34 AM 
	To: Karim BELABAS; pari-users list 
	Subject: RE: bnrisprincipal to slow

	How does
	 GP/PARI CALCULATOR Version 2.2.6 (development CHANGES-1.756)
	                i386 running cygwin (ix86 kernel) 32-bit version
	              compiled: May 28 2003, gcc-3.2 20020927 (prerelease)
	                (readline v4.3 enabled, extended help available)
	compare with version 2.2.7?
	I am trying to find a set of fundamental units for the field generated by a degree 37 polynomial
	 (discriminant 10^36*37^37) and it has been running for about 8 days now. A similar computation for a polynomial of degree 31 took about a day.  I am wondering if version 2.2.7 would speed things up considerably or perhaps the computations have run into some kind of loop?  My code for this is:
	a=(x + 2*I)^n;
	b=(x - 2*I)^n;
	f=     round ( real(  (a + b)/2 - (a - b)/(4*I)   ))   ;
	fus=    bnfinit( f,1).fu    ;
	Jimmy Mc Laughlin.
	First make absolutely sure you are using the CVS version of PARI [ headers
	should read something like
	        GP/PARI CALCULATOR Version 2.2.7 (development CHANGES-1.873)
	with possibly a slightly lower patchlevel. ]
	The development version is orders of magnitude faster than the stable ones
	(2.1.*) for algebraic number theory (and a few other things).