On Tue, Jun 1, 2010 at 5:04 PM, Bill Hart <[email protected]> wrote:
> Which is curiously precisely how long the system GMP takes to compute
> mpz_fac_ui(1000000) on sage.math.
>
> Somehow you aren't picking up MPIR.
>
> Can you check that sage/local has libgmp and gmp.h coming from MPIR.
>
> You still need to build with the --enable-gmpcompat option to
> configure and you also need to do make install (unless you rename
> libmpir and mpir.h yourself).

I just did a full rebuild of the sage library, and now things are *awesome*

sage: time n=factorial(10^6)
CPU times: user 0.37 s, sys: 0.00 s, total: 0.37 s
Wall time: 0.36 s
sage: time m=n*n
CPU times: user 0.21 s, sys: 0.02 s, total: 0.23 s
Wall time: 0.23 s

This is solidly better than MPIR-1.2 in Sage.

Now I'll do a complete clean build of sage-4.4.3.alpha1 from source
with the new mpir, with tests, and see if anything comes up.

William

>
> Bill.
>
> On 1 June 2010 23:08, William Stein <[email protected]> wrote:
>> On Tue, Jun 1, 2010 at 2:43 PM, Bill Hart <[email protected]> 
>> wrote:
>>> Sorry, I mean, you don't need make install-gmpcompat any more, only
>>> make install.
>>
>> Thanks.
>>
>> I built and then started Sage, but even my first bencharmk is 3 times
>> *slower* than the old, old mpir in Sage.
>>
>> With MPIR-1.2.x (in current Sage):
>>
>> sage: time n=factorial(10^6)
>> CPU times: user 0.52 s, sys: 0.02 s, total: 0.54 s
>> Wall time: 0.54 s
>> sage: time k=n*n
>> CPU times: user 0.24 s, sys: 0.01 s, total: 0.25 s
>> Wall time: 0.25 s
>>
>>
>>
>> With the new MPIR-2.1.0:
>>
>> sage: time n=factorial(10^6)
>> CPU times: user 1.53 s, sys: 0.00 s, total: 1.53 s
>> Wall time: 1.54 s
>> sage: time k=n*n
>> CPU times: user 0.63 s, sys: 0.01 s, total: 0.64 s
>> Wall time: 0.64 s
>>
>>
>> This is on sage.math.washington.edu.    For this platform Sage doesn't
>> do anything nonstandard (just a very straightforward make, make
>> install...).
>>
>> Maybe the tuning or architecture detection code is totally broken.  I
>> don't know.
>>
>>
>> William
>>
>>
>>>
>>> On 1 June 2010 22:43, Bill Hart <[email protected]> wrote:
>>>> You don't need make install --enable-gmpcompat any more. Hasn't been
>>>> in MPIR for ages.
>>>>
>>>> Just do make install.
>>>>
>>>> You still have to pass --enable-gmpcompat to configure of course.
>>>>
>>>> Bill.
>>>>
>>>> On 1 June 2010 22:30, William Stein <[email protected]> wrote:
>>>>> On Tue, Jun 1, 2010 at 2:14 PM, Minh Nguyen <[email protected]> wrote:
>>>>>> Hi folks,
>>>>>>
>>>>>> MPIR 2.1.0-rc2 was released on 01st June 2010.
>>>>>>
>>>>>> Source:
>>>>>> http://www.mpir.org/mpir-2.1.0-rc2.tar.gz
>>>>>>
>>>>>> Documentation:
>>>>>> http://www.mpir.org/mpir-2.1.0.pdf
>>>>>>
>>>>>> A number of issues reported with the 2.1.0-rc1 release [1] are fixed.
>>>>>> Please test and report all issues.
>>>>>>
>>>>>> If you are having failures when building or tuning, try setting the
>>>>>> following prior to configuring and build:
>>>>>>
>>>>>> $ export MAKE='make'
>>>>>>
>>>>>> If you have failures when running the test suite, try edit the file
>>>>>> mpirtest to use only one CPU. You can use/tweak this script [2] to
>>>>>> automate these steps: configure, build, check, tune, run test suite.
>>>>>>
>>>>>> [1] 
>>>>>> http://groups.google.com/group/mpir-devel/browse_thread/thread/4550cb9a8aae0719/
>>>>>>
>>>>>> [2] http://sage.math.washington.edu/home/mvngu/apps/mpir/go-build.sh
>>>>>
>>>>> I made an spkg from this for Sage, and it doesn't work, since the
>>>>> install-gmpcompat target is suddenly gone.  What am I supposed to use
>>>>> instead to install gmpcompat'ability?
>>>>>
>>>>> make[4]: Leaving directory
>>>>> `/mnt/usb1/scratch/wstein/build/sage-4.4.2.rc0/spkg/build/mpir-2.1.0-rc2/src'
>>>>> make[3]: Leaving directory
>>>>> `/mnt/usb1/scratch/wstein/build/sage-4.4.2.rc0/spkg/build/mpir-2.1.0-rc2/src'
>>>>> make[2]: Leaving directory
>>>>> `/mnt/usb1/scratch/wstein/build/sage-4.4.2.rc0/spkg/build/mpir-2.1.0-rc2/src'
>>>>> make[1]: Leaving directory
>>>>> `/mnt/usb1/scratch/wstein/build/sage-4.4.2.rc0/spkg/build/mpir-2.1.0-rc2/src'
>>>>> make: *** No rule to make target `install-gmpcompat'.  Stop.
>>>>> Error installing MPIR.
>>>>>
>>>>> real    2m6.588s
>>>>> user    1m11.950s
>>>>> sys     0m51.550s
>>>>>
>>>>>
>>>>> --
>>>>> William Stein
>>>>> Professor of Mathematics
>>>>> University of Washington
>>>>> http://wstein.org
>>>>>
>>>>> --
>>>>> You received this message because you are subscribed to the Google Groups 
>>>>> "mpir-devel" group.
>>>>> To post to this group, send email to [email protected].
>>>>> To unsubscribe from this group, send email to 
>>>>> [email protected].
>>>>> For more options, visit this group at 
>>>>> http://groups.google.com/group/mpir-devel?hl=en.
>>>>>
>>>>>
>>>>
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups 
>>> "mpir-devel" group.
>>> To post to this group, send email to [email protected].
>>> To unsubscribe from this group, send email to 
>>> [email protected].
>>> For more options, visit this group at 
>>> http://groups.google.com/group/mpir-devel?hl=en.
>>>
>>>
>>
>>
>>
>> --
>> William Stein
>> Professor of Mathematics
>> University of Washington
>> http://wstein.org
>>
>> --
>> You received this message because you are subscribed to the Google Groups 
>> "mpir-devel" group.
>> To post to this group, send email to [email protected].
>> To unsubscribe from this group, send email to 
>> [email protected].
>> For more options, visit this group at 
>> http://groups.google.com/group/mpir-devel?hl=en.
>>
>>
>
> --
> You received this message because you are subscribed to the Google Groups 
> "mpir-devel" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to 
> [email protected].
> For more options, visit this group at 
> http://groups.google.com/group/mpir-devel?hl=en.
>
>



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

-- 
You received this message because you are subscribed to the Google Groups 
"mpir-devel" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/mpir-devel?hl=en.

Reply via email to