[SciPy-user] some benchmark data for numarray, Numeric and scipy-newcore

Ted Horst ted.horst at earthlink.net
Sun Dec 4 15:40:58 CST 2005


I took a quick look at this in numarray and scipy (I don't have 
Numeric).

It seems that numarray has separate lowest level functions for vector, 
vector multiplication and scalar, vector multiplication whereas scipy 
uses the same function for both.  This results in an extra load double 
instruction being executed in scipy for the scalar, vector case.

This is on a Powerbook G4, running OSX 10.3.9.

Ted

On Dec 3, 2005, at 21:07, Travis Oliphant wrote:

> David M. Cooke wrote:
>
>> These are my numbers. They don't support the above conclusions :-)
>> This is on an Athlon 64 3200+, running Debian. Numeric and numarray
>> are from Debian unstable.
>>
>> cookedm at arbutus$ py bench.py 12
>> Numeric-24.2
>> numarray-1.4.0
>> scipy-core-0.7.4.1550
>> benchmark size = 12  (vectors of length 16777216)
>> label            Numeric       numarray     scipy.base
>>     1             0.5642          1.086          0.541
>>     2              1.044          1.952         0.8472
>>     3             0.7367         0.7428         0.6201
>>     4              1.965         0.8811           2.45
>>     5              1.048         0.9061         0.9241
>>     6             0.7419         0.7496         0.5371
>>     7              1.972          1.293          2.412
>>     8              1.222           0.94         0.9529
>>     9              23.99          13.65          13.16
>>    10              14.96          13.65          11.42
>>    11              10.86          10.61          9.056
>> TOTAL               59.1          46.47          42.92
>
>
> I like your version better.... :-)   But, seriously,  it does still 
> look
> like something is going on for lines 4 and 7 that could be improved in
> scipy core....
>
> -Travis
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.net
> http://www.scipy.net/mailman/listinfo/scipy-user
>



More information about the SciPy-user mailing list