Final 3.4.1 source released

[mabshoff]

Hello folks,

as expected changes over 3.4.1.rc4 are minimal:

Merged in Sage 3.4.1.final:

#5284: Michael Abshoff: Set sage-flags.txt up to SSE2 only when
building Sage in SSE2 only mode/remove SSSE3 and SSE4 flags (followup
to #5219) [Reviewed by Gonzalo Tornaria]
#5829: Minh Van Nguyen: copyright on standard documentation [Reviewed
by Dan Drake]

To build an SSE2 only binary do the following:

SAGE_SIMD_MODE=SSE2; export SAGE_SIMD_MODE
make

This will obviously have an impact on performance and I am afraid it
could be significant for certain problems. If anyone bothers to do two
builds please let us know. Good candidates for performance regressions
would be linear algebra, i.e. anything that involves ATLAS.

All the bits are as usual in

sage.math.washington.edu/home/mabshoff/release-cycles-3.4.1/

I am building my 3.4.2.alpha0 merge tree and should merge in a couple
hours. Hopefully 3.4.2.alpha0 will drop tomorrow.

Cheers,

Michael

[David M. Monarres]

On an upgrade from 3.4 on Mac OS X 10.5.6 (intel) I get the following
doctest errors:

The following tests failed:


sage -t "devel/sage/sage/algebras/quaternion_algebra_element.py"
sage -t "devel/sage/sage/schemes/elliptic_curves/
ell_rational_field.py"
sage -t "devel/sage/sage/structure/sage_object.pyx"

The errors are a bit long so here is a link to the test.log:

http://www-rohan.sdsu.edu/~monarres/test.log




--
David Monarres
dmmon...@gmail.com

"There... I've run rings 'round you logically"
-- Monty Python's Flying Circus

[mabshoff]

On Apr 22, 12:54 pm, "David M. Monarres" <dmmonar...@gmail.com> wrote:

Hi David,

> On an upgrade from 3.4 on Mac OS X 10.5.6 (intel) I get the following  
> doctest errors:

Thanks for the build report.

> The following tests failed:
>
>         sage -t  "devel/sage/sage/algebras/quaternion_algebra_element.py"
>         sage -t  "devel/sage/sage/schemes/elliptic_curves/
> ell_rational_field.py"
>         sage -t  "devel/sage/sage/structure/sage_object.pyx"
>
> The errors are a bit long so here is a link to the test.log:
>
> http://www-rohan.sdsu.edu/~monarres/test.log

The problems:

sage -t "devel/sage/sage/schemes/elliptic_curves/
ell_rational_field.py"
**********************************************************************
File "/Users/ayeq/sage/devel/sage/sage/schemes/elliptic_curves/
ell_rational_field.py", line 4941:
sage: P = G(E.0) + G(E.1) + G(phi(F.0)); P
Expected:
(-867/3872*a - 3615/3872 : -18003/170368*a - 374575/170368 : 1)
Got:
(-51/1058*a + 141/1058 : -1581/12167*a - 9912/12167 : 1)
**********************************************************************
File "/Users/ayeq/sage/devel/sage/sage/schemes/elliptic_curves/
ell_rational_field.py", line 4943:
sage: P.division_points(2)
Expected:
[(1/8*a + 5/8 : -5/16*a - 9/16 : 1)]
Got:
[(1/2*a - 1/2 : 1/2*a - 5/2 : 1)]
**********************************************************************
File "/Users/ayeq/sage/devel/sage/sage/schemes/elliptic_curves/
ell_rational_field.py", line 3185:
sage: E.cremona_label()
Expected:
Traceback (most recent call last):
...
RuntimeError: Cremona label not known for Elliptic Curve defined
by y^2 + x*y + 3*y = x^3 + 2*x^2 + 4*x + 5 over Rational Field.
Got:
'10351a1'
**********************************************************************

This is caused by you having some optional Cremona database installed
and it is #5346 which has a trivial fix suggested by John - someone
just has to put a patch up and unless someone beats me to it I will do
it tonight :)

The other two failures are upgrade specific and happen when when left
over pyc and so files are around due to the Quaternion classes being
moved. One brutal fix is to do delete devel/sage* and to do a -ba.
The more fine tuned fix is to delete some py files and touch a couple
others followed by a -b, but I would need to look up the details.

> --
> David Monarres
> dmmonar...@gmail.com

Cheers,

Michael

[mabshoff]

On Apr 22, 1:50 pm, mabshoff <mabsh...@googlemail.com> wrote:
> On Apr 22, 12:54 pm, "David M. Monarres" <dmmonar...@gmail.com> wrote:
>
> Hi David,

<SNIP>

> The other two failures are upgrade specific and happen when when left
> over pyc and so files are around due to the Quaternion classes being
> moved. One brutal fix is to do  delete devel/sage* and to do a -ba.

Ehh, devel/sage/build obviously - or if you nuked devel/sage* just
force a rebuild of the sage.spkg via

sage -f spkg/standard/sage-3.4.1.spkg

Sorry for the double post, but I didn't want to kill people's install
by incorrect advice.

Cheers,

Michael

[Marshall Hampton]

I am having an error in building 3.4.1 on an intel mac running 10.5.
The error is in mpir:

Deleting assembly files which depend on PIC assembly working or 32 bit
OSX on Intel hardware
checking build system type... Invalid configuration `penryn-apple-
darwin9.6.0': machine `penryn-apple' not recognized
configure: error: /bin/sh ./config.sub penryn-apple-darwin9.6.0 failed
Failed to configure.


The chip is an intel Core 2 Duo, 2.4 Ghz, on a Macbook. This was on a
fresh source build, not an upgrade.

-M. Hampton

[Bill Hart]

Hi Marshall. I think I have a fix for this. But I've started a thread
on the MPIR development list for this:

http://groups.google.co.uk/group/mpir-devel/browse_thread/thread/34a4e60f556c8ba3

Bill.

[John Cremona]

2009/4/22 mabshoff <mabs...@googlemail.com>:

I am fixing this. In fact all three only arise when the optional
database is installed (reason: you then get generators for more
curves from the database instead of computing them). I will post
patch which addresses all three at #5346.

John

[John Cremona]

2009/4/23 John Cremona <john.c...@gmail.com>:

Done and ready for review.

[Bill Hart]

The latest RC of MPIR 1.1.1 should fix this issue.

Bill.

On 23 Apr, 01:43, Bill Hart <goodwillh...@googlemail.com> wrote:
> Hi Marshall. I think I have a fix for this. But I've started a thread
> on the MPIR development list for this:
>

> http://groups.google.co.uk/group/mpir-devel/browse_thread/thread/34a4...

[Harald Schilly]

I just want to add that 3.4.1 compiles on my hp 2140 netbook (3.4
didn't) on a dual core Intel(R) Atom(TM) CPU N270 / model 28

All tests pass, except for a timeout in schemes/elliptic_curves/
ell_rational_field.py

h

[evan foss]

How long did that take to build?

--
http://www.coe.neu.edu/~efoss/
http://evanfoss.googlepages.com/

[William Stein]

On Sun, Apr 26, 2009 at 8:25 AM, evan foss <evan...@gmail.com> wrote:
>
> How long did that take to build?

I did a similar build and it took about 10 hours I think. Most of the
time (maybe 2-4 hours?) is spent running the FLINT *test suite*, not
building anything (so, I think running the FLINT test suite should be
*off* in the release!). ATLAS takes < 1 hour.

William

--
William Stein
Associate Professor of Mathematics
University of Washington
http://wstein.org

[Harald Schilly]

On Apr 26, 5:25 pm, evan foss <evanf...@gmail.com> wrote:
> How long did that take to build?

don't know, some hours. .. started in the night before going to sleep
and in the morning everything was done. i did a parallel make and i
have 2 gb of ram. maybe 10 hours, but i think it was faster.

h

[evan foss]

Ok thanks, I have been wondering about the performance of netbooks.

[Minh Nguyen]

Hi folks,

On Wed, Apr 22, 2009 at 5:49 AM, mabshoff <mabs...@googlemail.com> wrote:
>
> Hello folks,
>
> as expected changes over 3.4.1.rc4 are minimal:
>
> Merged in Sage 3.4.1.final:
>
> #5284: Michael Abshoff: Set sage-flags.txt up to SSE2 only when
> building Sage in SSE2 only mode/remove SSSE3 and SSE4 flags (followup
> to #5219) [Reviewed by Gonzalo Tornaria]
> #5829: Minh Van Nguyen: copyright on standard documentation [Reviewed
> by Dan Drake]
>
> To build an SSE2 only binary do the following:
>
> SAGE_SIMD_MODE=SSE2; export SAGE_SIMD_MODE
> make
>
> This will obviously have an impact on performance and I am afraid it
> could be significant for certain problems. If anyone bothers to do two
> builds please let us know. Good candidates for performance regressions
> would be linear algebra, i.e. anything that involves ATLAS.
>
> All the bits are as usual in
>
> sage.math.washington.edu/home/mabshoff/release-cycles-3.4.1/

And you can find a nicely formatted version of the release tour at

http://mvngu.wordpress.com/2009/04/27/sage-341-released/

<SNIP>
--
Regards
Minh Van Nguyen

[Harald Schilly]

On Apr 27, 4:48 am, evan foss <evanf...@gmail.com> wrote:
> Ok thanks, I have been wondering about the performance of netbooks.
>
>

well, i'm surprised how good it is. you can give me some calculations
and i can benchmark it.

h