On Tue, 2007-07-31 at 17:35 -0500, Alec Mihailovs wrote:
> > Actually, even on my 32 bit core duo, the long double type seems to give
> > 64 bits of precision, so using it might help a little. Do you have any
> > idea how to check at run/compile time what the precision of a double or
> > a long do
On Tue, 2007-07-31 at 22:16 -0700, Justin C. Walker wrote:
[...]
> Checking partition count computation:
>
> On a Core 2 Duo 2.33 Mhz, computing the number of partitions of 10^9:
>Mathematica 5.2 (PartitionsP[10^9]:95.5115 s
>Sage 2.7.2.1 (number_of_partitions(10^9): 125.2 s
>
On Tue, 2007-07-31 at 22:16 -0700, Justin C. Walker wrote:
>
> On Jul 31, 2007, at 18:36 , William Stein wrote:
>
> >
> > Hi,
> >
> > Just to extend this thread some more, a few remarks.
>
> While we're extending this, here's my $0.02 canadian (:-})
[...]
> Checking partition count computatio
On 7/31/07, Justin C. Walker <[EMAIL PROTECTED]> wrote:
> While we're extending this, here's my $0.02 canadian (:-})
> > Yep -- Mathematica 5.2 interestingly totally sucks at
> > computing the number of partitions on an Intel OSX
> > machine... and SAGE rocks.
>
> Checking partition count computat
On Jul 31, 2007, at 18:36 , William Stein wrote:
>
> Hi,
>
> Just to extend this thread some more, a few remarks.
While we're extending this, here's my $0.02 canadian (:-})
> On sage.math:
> SAGE:
> sage: time n = number_of_partitions(10^7)
> CPU times: user 0.73 s, sys: 0.00 s, total: 0.73 s
Hi,
Just to extend this thread some more, a few remarks.
(1) quaddouble has been included in SAGE for several months
now, thanks to the hard work of Didier Deshommes and Robert
Bradshaw.
sage: RQDF
Real Quad Double Field
sage: RQDF(2).sin()
0.9092974268256816953960198659117448427022549714478902
Ok, it looks like I've tracked down the problem. When sage built on my
machine, it did not compile mpfr with -O2.
I've replicated this behavior in my own build of mpfr, and it seems to
account exactly for the slowdown. Some other options passed to gcc are
different, as well. They may be less impo
From: "Alec Mihailovs" <[EMAIL PROTECTED]>
> Being an assembler programmer, I can say definitely that all FPU registers
> on x86 are 80-bit and all compilers that I know compile long double as
> 80-bit numbers.
>From other point of view, 80-bit real gives 64-bit precision in usual sense
(mantis
> Actually, even on my 32 bit core duo, the long double type seems to give
> 64 bits of precision, so using it might help a little. Do you have any
> idea how to check at run/compile time what the precision of a double or
> a long double is?
Being an assembler programmer, I can say definitely tha
Here is a better way to tell exactly what libraries are being used:
Specify them exactly by file name and link statically.
Finally, I found something that narrows down the problem a little bit, I
think.
[EMAIL PROTECTED]:~/temp$ g++ partitions_c.cc -O3
/home/bober/sage-2.7.1/sage-2.7.1/local/li
Or probably 212 actually. :-)
On 31 Jul, 22:24, Bill Hart <[EMAIL PROTECTED]> wrote:
> I do highly recommend this quad double library by the way. And they've
> implemented all manor of transcendental functions too!! The quad-
> doubles would give you 206 bits, even on your machine.
>
> Bill.
>
>
I do highly recommend this quad double library by the way. And they've
implemented all manor of transcendental functions too!! The quad-
doubles would give you 206 bits, even on your machine.
Bill.
On 31 Jul, 21:33, Jonathan Bober <[EMAIL PROTECTED]> wrote:
> On Mon, 2007-07-30 at 17:24 -0700, B
Ah, that's better. Excellent. I feel much happier with this library
now.
Thanks.
Bill.
On 31 Jul, 21:38, "Mike Hansen" <[EMAIL PROTECTED]> wrote:
> From COPYING,
>
> "Redistribution and use in source and binary forms, with or without
> modification, are permitted provided that the following con
I believe that the IEEE standard guarantees you 80 bits (though it's
only 64 bits of mantissa or something like that). The trouble is, you
aren't guaranteed the IEEE standard.
I've spent much time researching this, but either I didn't look at the
right websites, or this stuff isn't documented wel
>From COPYING,
"Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met"
--Mike
On 7/31/07, Bill Hart <[EMAIL PROTECTED]> wrote:
>
> Oh well, I don't understand all this licensing stuff. So do you
> understand
On Mon, 2007-07-30 at 17:24 -0700, Bill Hart wrote:
> Wow!! Excellent work indeed.
>
> In fact on 64 bit X86 systems you could actually use the 128 bit long
> doubles to give you a little bit more precision (I believe it only
> gives you 80 bits including exponent and sign, so probably 64 bit
> m
Oh well, I don't understand all this licensing stuff. So do you
understand this license? What about derived works. Does that mean it
is not possible to modify this library and redistribute the modified
version?
In particular, as a C++ library it is no use to me unmodified. There
are also some fun
In this case, the license does not says that it is in the public domain
(but that it is a copyrighted work!), but you can use it as
"AS IS".
I think that the only condition that is imposed to us is
to include the declaimer.
Pablo
On 7/31/07, Pablo De Napoli <[EMAIL PROTECTED]> wrote:
> Certainly
Certainly no, not everything from a public institution is in the
public domain. This
should be analyzed case by case.
In case of doubt it would be better to ask the author.
Pablo
On 7/31/07, Bill Hart <[EMAIL PROTECTED]> wrote:
>
> Yes, I've ended up using Hida and Bailey's quad-double package.
ok, so I did misunderstand your response. :)
Regarding ModuleElements and AdditiveGroupElements, I guess I was
considering modules to be Abelian groups with the action of a ring. I guess
I can see advantage and disadvantage for both hierarchy. I think
AdditiveGroupElement being derived from Module
On Jul 31, 2007, at 11:16 AM, Soroosh Yazdani wrote:
> I just realized that I was using a seriously outdated version of
> sage (2.6.something). Using 2.7.2.1, scalar division on matrices
> does work. In case anybody was wondering.
>
> However, I am still confused on your response regarding Ma
Yes, I've ended up using Hida and Bailey's quad-double package. Very
cool.
But the license just says not to use the LB name to promote any
derived product. Am I right in assuming this is GPL compatible, i.e.
because they are a public institution everything is automatically
public domain?
Bill.
I just realized that I was using a seriously outdated version of sage (
2.6.something). Using 2.7.2.1, scalar division on matrices does work. In
case anybody was wondering.
However, I am still confused on your response regarding Matrix being derived
from AlgebraElement. Are you saying that Matrix
"William Stein" <[EMAIL PROTECTED]> writes:
> I also added f.polynomial(...) for f a multivariate polynomial, which
> sort of fits into the thread of this discussion. This is
> a very useful function for certain applications -- it allows you to
> view a multivariate polynomial as a single variab
On Jul 30, 2007, at 12:26 PM, didier deshommes wrote:
>
> 2007/7/30, Carl Witty <[EMAIL PROTECTED]>:
>> It seems pretty strange to me, mostly because you lose too much
>> information by eliding zeroes. As far as I can tell, given
>> MPolynomialRing(QQ,2,order='lex'), all of the following polyno
I think you do not need to specify stdc++ explicitly when calling the
compiler & linker via "g++" but you do if via "gcc". And it's the
library containing C++ specifics including cin, cout etc.
John
On 7/31/07, David Harvey <[EMAIL PROTECTED]> wrote:
>
> h okay there's a few pedantic things
On 7/31/07, Martin Albrecht <[EMAIL PROTECTED]> wrote:
>
> > Is this a bug, or am I not using this correctly?
>
> I'll add the option to construct an MPolynomial_libsingular from a PolyDict.
Too late, I just did it, since I needed it for something else I'm
doing (related to power series over polyn
> Is this a bug, or am I not using this correctly?
I'll add the option to construct an MPolynomial_libsingular from a PolyDict.
Martin
--
name: Martin Albrecht
_pgp: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x8EF0DC99
_www: http://www.informatik.uni-bremen.de/~malb
_jab: [EMAIL PROTE
28 matches
Mail list logo
