Hi,
Continuing this thread, I think that building Sage shouldn't require
X11. E.g., on t2, the new R png tests fail:
File "/scratch/wstein/build/sage-4.4.1.alpha2/devel/sage/sage/interfaces/r.py",
line 993:
sage: r.png(filename='"%s"'%filename) # filename not needed in
notebook, used for doc
William,
I am subscribed to sage-release. OK, I'll cc there and then move over
to there.
The upgrade of polybori to
http://sage.math.washington.edu/home/mvngu/spkg/standard/polybori/polybori-0.6.4.p0.spkg
does not quite help in my case.
Did you have any problems on ia64 with gcc-4.5?
The bug see
On Fri, Apr 30, 2010 at 9:39 PM, Mike Hansen wrote:
> On Fri, Apr 30, 2010 at 9:27 PM, Dima Pasechnik wrote:
>> Seems that this is connected to the reported
>> *** glibc detected *** python: corrupted double-linked list:
>> 0x6140b350 ***
>> troubles.
>
> My guess is that it is indeed
> h
On Fri, Apr 30, 2010 at 9:27 PM, Dima Pasechnik wrote:
> Seems that this is connected to the reported
> *** glibc detected *** python: corrupted double-linked list:
> 0x6140b350 ***
> troubles.
My guess is that it is indeed
http://trac.sagemath.org/sage_trac/ticket/8830 . You can try
ins
I've built sage-4.4.1.alpha2 on ia64 (Skynet's iras), with gcc-4.5.0,
to investigate recently made GAP patches.
It builds, but during the process, as well as during 'make test' I see
lots of
Unhandled SIGSEGV: A segmentation fault occure
On Apr 30, 2010, at 8:00 PM, Bill Hart wrote:
Finally, the figures for karatsuba with signed coefficients that vary
wildly. Hardly distinguishable from the figures for classical
multiplication. With that I finish my little study. And in the
inimitable words of Adam and Jamie (and with about as m
Finally, the figures for karatsuba with signed coefficients that vary
wildly. Hardly distinguishable from the figures for classical
multiplication. With that I finish my little study. And in the
inimitable words of Adam and Jamie (and with about as much "science"
to back it up): MYTH BUSTED!!
So w
>
> > Maybe we should just *always* build it with X support except on Mac?
> > But presumably it needs some X library to connect to to do that...?
>
> sphg-install in R has numerous problems. Someone updated it which
> caused major hassles on Solaris, with more fallout on Linux. They
> obviously
Now karatsuba for signed coefficients:
len = 1, min = 0, av = 0, max = 0, prec = 106
len = 2, min = 0, av = 1, max = 5, prec = 106
len = 3, min = 0, av = 1, max = 4, prec = 106
len = 4, min = 0, av = 0, max = 3, prec = 106
len = 6, min = 0, av = 0, max = 6, prec = 106
len = 8, min = 0, av = 1, max
And now for karatsuba. First for unsigned coefficients all about the
same magnitude:
len = 1, min = 0, av = 0, max = 0, prec = 106
len = 2, min = 0, av = 1, max = 2, prec = 106
len = 3, min = 0, av = 1, max = 3, prec = 106
len = 4, min = 0, av = 1, max = 3, prec = 106
len = 6, min = 0, av = 0, max
Now for toom cook with wildly varying signed coefficients. A little
over twice the precision loss of the classical algorithm. Hardly what
I'd call an unmitigated disaster. Certainly very usable.
len = 1, min = 0, av = 18, max = 79, prec = 106
len = 2, min = 0, av = 27, max = 101, prec = 106
len =
Now toom cook with signed coefficients:
len = 1, min = 0, av = 0, max = 0, prec = 106
len = 2, min = 0, av = 1, max = 5, prec = 106
len = 3, min = 0, av = 2, max = 7, prec = 106
len = 4, min = 0, av = 0, max = 3, prec = 106
len = 6, min = 0, av = 2, max = 6, prec = 106
len = 8, min = 0, av = 2, ma
As promised, here with toom cook figures, first for unsigned
coefficients all about the same magnitude:
len = 1, min = 0, av = 0, max = 0, prec = 106
len = 2, min = 0, av = 1, max = 2, prec = 106
len = 3, min = 0, av = 2, max = 6, prec = 106
len = 4, min = 0, av = 1, max = 3, prec = 106
len = 6, m
On Apr 30, 12:23 am, Robert Bradshaw
wrote:
>
> > (RJF) Maxima was designed around exact arithmetic, and generally offers to
> > convert floats to their corresponding exact rationals before doing
> > anything
> > that requires arithmetic. It makes no claims about floats per se,
> > partly
> > b
On Apr 30, 2:17 am, Bill Hart wrote:
> Actually, I lie, slightly. I did find one instance of `numerical
> stability' used in reference to the FFT, and that is on wikipedia (so
> now we all know it must be true).
Again,
Accuracy and stability of numerical algorithms
By Nicholas J. Higham
I thi
On Apr 30, 1:57 am, Bill Hart wrote:
> On Apr 30, 6:58 am, rjf wrote:
concept. What is it used for? I
> can't imagine defining a GCD in this context as divisibility is an
> exact phenomenon.
Google for "approximate GCD".
>
> I hear the term numerical stability used quite a lot. The two context
On 30 April 2010 01:43, kcrisman wrote:
> Moved from
> http://groups.google.com/group/sage-release/browse_thread/thread/e7d692cb7859162e
> on sage-release:
>
> On Apr 29, 6:32 pm, Dan Drake wrote:
>> On Thu, 29 Apr 2010 at 08:46AM -0700, kcrisman wrote:
>> > It would be interesting to see what h
And finally, the figures for Rader-Brennan for signed coefficients
whose magnitudes also vary wildly: as with FHT, on average, no usable
information results. No FFT algorithm will be of use in that
situation. Joris vdH's algorithm is your only hope.
After dinner, figures for Karatsuba and Toom Coo
Here are the figures for Rader-Brennan for signed coefficients. On
average, the precision loss is radically worse in this case than Fast
Hartley, but still quite usable:
len = 1, min = 0, av = 0, max = 1, prec = 106
len = 2, min = 0, av = 1, max = 5, prec = 106
len = 3, min = 0, av = 2, max = 12,
I installed Andreas Enge's mpfrcx package:
http://www.multiprecision.org/index.php?prog=mpfrcx&page=html
which just worked for me.
I took a closer look and the Rader-Brennan FFT is a complex FFT taking
mpcx polynomials as input. When multiplying polynomials over the
reals, it simply converts to
cch schrieb:
To bb;
1. Is there anywhere a md5 available?
2. The downlaod has a speed of about 13 KB/s, not a breakneck speed, one
might get it in about 20 hours. Is it possible to copy the iso to
another place with large pipes?
1. Yes, you can also find md5 file in the same director
On Fri, Apr 30, 2010 at 10:11:33AM -0700, Pierre wrote:
> i'm on 4.3 (and said so already, look again :-) )
Oups
> so there's little hope.
I think some variant of
sage -upgrade http://sage.math.washington.edu/home/release/sage-4.3.5/
should upgrade your sage to the 4.3.5 version which is
i'm on 4.3 (and said so already, look again :-) )
so there's little hope. It seems to be possible to multiply rational
fractions by passing them to pari, which returns a string, then parse
the string with regular expressions to get the numerator and
denominator, and then try :
result=
sage.rings.
Hi Pierre,
> ok, so it failed with :
You didn't answer John's crucial question :
> What version of sage are you running? If it was 4.3.5 then you could
> do the following. Start in the directory where sage is installed
> (usually called SAGE_ROOT). I assume that this sage is in your pat
> Try looking at:
>
> http://www.sagemath.org/doc/developer/walk_through.html
hi rob,
i quite like the documentation there, but i don't seem to be able to
commit any of it to memory -- lack of practice, surely. Every time
somebody mentions applying a patch, i think "oh no, don't want to read
all
Hi Pierre
On Fri, Apr 30, 2010 at 09:37:43AM -0700, Pierre wrote:
> > Just to let you now: This is supposed to be fixed by Ticket #8296 since
> > sage-4.3.4. Moreover it actually works on my machine and some other. Please
> > tell us if it's still broken after upgrading (if you dare to ;-)
>
ok, so it failed with :
m-guillot:sage-ratfunc pedro$ hg qpush
applying trac4000_433_combined.patch
patching file sage/libs/flint/fmpz_poly.pxi
Hunk #1 FAILED at 2
Hunk #3 FAILED at 109
2 out of 3 hunks FAILED -- saving rejects to file sage/libs/flint/
fmpz_poly.pxi.rej
patching file sage/rings/fr
> Just to let you now: This is supposed to be fixed by Ticket #8296 since
> sage-4.3.4. Moreover it actually works on my machine and some other. Please
> tell us if it's still broken after upgrading (if you dare to ;-)
hi florent,
i knew you had a bot checking for the keyword 'emacs' on the sag
On Fri, Apr 30, 2010 at 8:58 AM, Minh Nguyen wrote:
> Hi Robert,
>
> On Sat, May 1, 2010 at 12:39 AM, Robert Miller wrote:
>> #7304 just got hit...
>>
>> I've noticed an increasing amount of this lately...
>
> And I have been removing spammer accounts lately. The most insidious
> one so far is a
Hi Pierre,
On Fri, Apr 30, 2010 at 08:09:44AM -0700, Pierre wrote:
> hey thanks, i'll try that when i have time. I'm on sage 4.3 -- i
> didn't want to upgrade because the emacs completion is broken in some
> recent version prior to 4.4, i noticed this on my other computer; i
> guess upgradin
Hi Robert,
On Sat, May 1, 2010 at 12:39 AM, Robert Miller wrote:
> #7304 just got hit...
>
> I've noticed an increasing amount of this lately...
And I have been removing spammer accounts lately. The most insidious
one so far is a spammer going by the username of bascorp. I have
removed that acco
On Apr 30, 7:25 am, Pierre wrote:
> > Would it be an option for you to apply the patch at
>
> > http://trac.sagemath.org/sage_trac/ticket/4000
>
> well, i went to sage-days in marseille and i'm supposed to know about
> these things... but mostly i went home feeling that it was just a bit
> too
hey thanks, i'll try that when i have time. I'm on sage 4.3 -- i
didn't want to upgrade because the emacs completion is broken in some
recent version prior to 4.4, i noticed this on my other computer; i
guess upgrading now would take me to 4.4, so i won't.
On 30 avr, 16:59, John Cremona wrote:
>
What version of sage are you running? If it was 4.3.5 then you could
do the following. Start in the directory where sage is installed
(usually called SAGE_ROOT). I assume that this sage is in your path.
sage -clone ratfunc ## wait a while
cd devel/sage-ratfunc
sage -hg qinit
sage -hg qimport
h
Though on the plus side, this spammer is actually advertising what he
says he is. In Russian.
On Apr 30, 10:44 am, kcrisman wrote:
> On Apr 30, 10:39 am, Robert Miller wrote:
>
> > #7304 just got hit...
>
> And # 7765.
>
>
>
> > I've noticed an increasing amount of this lately... How are these
On Apr 30, 10:39 am, Robert Miller wrote:
> #7304 just got hit...
And # 7765.
>
> I've noticed an increasing amount of this lately... How are these
> spammers getting accounts? We still have to give out accounts one at a
> time, right?
Nope.
http://trac.sagemath.org/sage_trac/register
So ju
#7304 just got hit...
I've noticed an increasing amount of this lately... How are these
spammers getting accounts? We still have to give out accounts one at a
time, right?
--
Robert L. Miller
http://www.rlmiller.org/
--
To post to this group, send an email to sage-devel@googlegroups.com
To uns
That patch applied fine to 4.3.5, but not quite to 4.4 (see the ticket
for details): rebase needed.
John
On 30 April 2010 15:09, Sebastian Pancratz wrote:
> Dear Pierre,
>
>> I'm trying to find a workaround for my particular example, can someone
>> help ?
>
> Would it be an option for you to app
> Would it be an option for you to apply the patch at
>
> http://trac.sagemath.org/sage_trac/ticket/4000
well, i went to sage-days in marseille and i'm supposed to know about
these things... but mostly i went home feeling that it was just a bit
too complicated... i barely remember that step 1
Dear Pierre,
> I'm trying to find a workaround for my particular example, can someone
> help ?
Would it be an option for you to apply the patch at
http://trac.sagemath.org/sage_trac/ticket/4000
? If I remember correctly, everything should be contained in the last
patch file on that ticket,
I'm trying to find a workaround for my particular example, can someone
help ?
Somehow i've managed to compute separately a polynomial A and another
one B, such that A/B is what I'm trying to compute. However asking
sage
sage: A/B
just takes ages (exactly how long i was not patient enough to find
>
> Approximate GCD? That's a curious concept. What is it used for? I
> can't imagine defining a GCD in this context as divisibility is an
> exact phenomenon.
For example, in an inverse parametrization problem. Suppose that you
have a rational curve given by a parametrization with float
cofficien
There is the following paper on numerically stable polynomial
multiplication by Joris van der Hoeven:
http://www.texmacs.org/joris/stablemult/stablemult-abs.html
But please note carefully the note on that page about an error in
bound 3.
Joris writes:
"Basically, when doing
the multiplication (
On Friday 30 April 2010, Sebastian Pancratz wrote:
> Indeed. Unfortunately, the patch file is rather large, which makes it
> difficult to have it tested and reviewed, and to keep it alive as
> other Sage code changes. A while ago I spoke to Martin Albrecht about
> this and we are both still inter
Indeed. Unfortunately, the patch file is rather large, which makes it
difficult to have it tested and reviewed, and to keep it alive as
other Sage code changes. A while ago I spoke to Martin Albrecht about
this and we are both still interested in trying to push #4000 a little
harder again very so
Actually, I lie, slightly. I did find one instance of `numerical
stability' used in reference to the FFT, and that is on wikipedia (so
now we all know it must be true). There I presume the context is
signal processing rather than polynomial multiplication. I did track
down one article which specifi
On Apr 30, 6:58 am, rjf wrote:
> On Apr 29, 10:58 am, Robert Bradshaw
> wrote:
>
> > On Apr 29, 2010, at 8:30 AM, rjf wrote:
>
> > > (RJF)Again, I see no definition of what you mean by accuracy in the result
> > > of polynomial multiplication.The easiest position to take is that of
> > > MPFR-
On Apr 30, 8:33 am, bb wrote:
> I am on the third day of the download. It stalls many times and without
> a downloadmanager it would be practically impossible to get it at all.
>
> Is there an alternative download site, that you announced? If so - which
> address?
I haven't done anything, it was
On Apr 29, 2010, at 10:58 PM, rjf wrote:
On Apr 29, 10:58 am, Robert Bradshaw
wrote:
On Apr 29, 2010, at 8:30 AM, rjf wrote:
(RJF)Again, I see no definition of what you mean by accuracy in
the result
of polynomial multiplication.The easiest position to take is that
of MPFR--
considering
49 matches
Mail list logo
