Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Matlab Benchmarks Update

10 views
Skip to first unread message

Harmonic Software

unread,
May 3, 2006, 12:13:12 PM5/3/06
to

We have updated the Matlab benchmarks at:

http://www.omatrix.com/bench.html
http://www.omatrix.com/benchstats.html

Source code for the benchmarks is available in the O-Matrix download:
http://www.omatrix.com/demodown.html

user....@yahoo.com

unread,
May 3, 2006, 1:32:34 PM5/3/06
to
Scilab 4.0 vs Matlab 7.1 SP3

http://www.sciviews.org/benchmark/index.html

Windows XP SP2 PIV HT 3.2 Ghz (GenuineIntel x86 Family 15 Model 2
Stepping 9)

Scilab-4.0

___________________________________________
scilab-4.0

Copyright (c) 1989-2006
Consortium Scilab (INRIA, ENPC)
___________________________________________


Startup execution:
loading initial environment

Scilab Benchmark 2

==================

Number of times each test is run__________________________: 3

I. Matrix calculation

---------------------

Creation, transp., deformation of a 1500x1500 matrix (sec): 0.4322917

800x800 normal distributed random matrix ^1000______ (sec): 0.203125

Sorting of 2,000,000 random values__________________ (sec): 0.828125

700x700 cross-product matrix (b = a' * a)___________ (sec): 0.1979167

Linear regression over a 600x600 matrix (c = a \ b') (sec): 0.1145833

--------------------------------------------------

Trimmed geom. mean (2 extremes eliminated): 0.2590245

II. Matrix functions

--------------------

FFT over 800,000 random values______________________ (sec): 0.984375

Eigenvalues of a 320x320 random matrix______________ (sec): 0.375

Determinant of a 650x650 random matrix______________ (sec): 0.1197917

Cholesky decomposition of a 900x900 matrix__________ (sec): 0.125

Inverse of a 400x400 random matrix__________________ (sec): 0.0885417

--------------------------------------------------

Trimmed geom. mean (2 extremes eliminated): 0.1777417

III. Programmation

------------------

750,000 Fibonacci numbers calculation (vector calc)_ (sec): 0.25

Creation of a 2250x2250 Hilbert matrix (matrix calc) (sec): 0.3854167

Grand common divisors of 70,000 pairs (recursion)___ (sec): 0.4739583

Creation of a 220x220 Toeplitz matrix (loops)_______ (sec): 0.5625

Escoufier's method on a 37x37 matrix (mixed)________ (sec): 0.625

--------------------------------------------------

Trimmed geom. mean (2 extremes eliminated): 0.4683793

Total time for all 15 tests_________________________ (sec): 5.765625

Overall mean (sum of I, II and III trimmed means/3)_ (sec): 0.2783402

--- End of test ---


Matlab 7.1.0.246 (R14) Service Pack 3
Matlab Benchmark 2
==================
Number of times each test is run__________________________: 3

I. Matrix calculation
---------------------
Creation, transp., deformation of a 1500x1500 matrix (sec): 1.9189
800x800 normal distributed random matrix ^1000______ (sec): 0.28557
Sorting of 2,000,000 random values__________________ (sec): 0.45095
700x700 cross-product matrix (b = a' * a)___________ (sec): 0.20764
Linear regression over a 600x600 matrix (c = a \ b') (sec): 0.14014

------------------------------------------------------
Trimmed geom. mean (2 extremes eliminated): 0.29903

II. Matrix functions
--------------------
FFT over 800,000 random values______________________ (sec): 0.27978
Eigenvalues of a 320x320 random matrix______________ (sec): 0.46314
Determinant of a 650x650 random matrix______________ (sec): 0.081383
Cholesky decomposition of a 900x900 matrix__________ (sec): 0.11033
Inverse of a 400x400 random matrix__________________ (sec): 0.089263

------------------------------------------------------
Trimmed geom. mean (2 extremes eliminated): 0.14019

III. Programmation
------------------
750,000 Fibonacci numbers calculation (vector calc)_ (sec): 0.77142
Creation of a 2250x2250 Hilbert matrix (matrix calc) (sec): 4.2263
Grand common divisors of 70,000 pairs (recursion)___ (sec): 1.066
Creation of a 220x220 Toeplitz matrix (loops)_______ (sec): 0.006017
Escoufier's method on a 37x37 matrix (mixed)________ (sec): 0.56908

------------------------------------------------------
Trimmed geom. mean (2 extremes eliminated): 0.77638


Total time for all 15 tests_________________________ (sec): 10.6659
Overall mean (sum of I, II and III trimmed means/3)_ (sec): 0.31928
--- End of test ---

Harmonic Software

unread,
May 4, 2006, 11:19:30 AM5/4/06
to
That page actually uses dated versions of Matlab (v6), O-Matrix, (v5), and
Scilab (v2.7).
The performance of O-Matrix has increased dramatically since that review and
I believe that Matlab and SciLab
performance has changed in recent version also.

<user....@yahoo.com> wrote in message
news:1146677554.5...@j33g2000cwa.googlegroups.com...

Carl

unread,
May 9, 2006, 2:31:18 PM5/9/06
to
Harmonic Software wrote:

Is this the end of this thread? What is the overall conclusion of the
benchmark update? To me it seems that Scilab is an incredible performer as
compared with Matlab. If this is the case, Scilab has travelled a very long
way since version 2.7!!!

Carl

Harmonic Software

unread,
May 10, 2006, 12:14:18 PM5/10/06
to
If SciLab performance has improved since 2.7 we would appreciate the update.
Please forward the SciLab implementation of the following benchmark and
results to in...@omatrix.com

"Carl" <phleum...@chello.se> wrote in message
news:Q758g.46$YI3.2@amstwist00...

0 new messages