From bug-octave-request at bevo dot che dot wisc dot edu Thu Dec 11 16:24:16 2003
Subject: inconsistent [a b].^x vs [a^x b^x]
From: Paul Kienzle <pkienzle at jazz dot ncnr dot nist dot gov>
To: bug-octave at bevo dot che dot wisc dot edu
Date: Thu, 11 Dec 2003 17:23:52 -0500

I'm finding [a b].^x != [a^x b^x].  Is there anything we can do about it?

Here's the results for octave-2.1.52 on Debian with -llapack -lblas

	octave:2> [-2^-.5,0^-.5], [-2,0].^-.5, octave_config_info('BLAS_LIBS')
	ans = -0.70711       Inf
	ans = NaN            Inf
	ans = /usr/lib/liblapack2.so /usr/lib/libblas2.so

and octave-2.1.40 on IRIX with -lblas and octave's lapack:

	octave:1>  [-2^-.5,0^-.5], [-2,0].^-.5, octave_config_info('BLAS_LIBS')
	ans = -0.70711       Inf
	ans = 0.00000 - 0.70711i      NaN +     NaNi
	ans = -lblas

and octave-2.1.48 on IRIX with -lscs:

	octavePAK:2> [-2^-.5,0^-.5], [-2,0].^-.5, octave_config_info('BLAS_LIBS')
	ans = -0.70711       Inf
	ans = 0.00000 - 0.70711i      Inf +     NaNi
	ans = -lscs

and octave-2.1.52 on OS X 10.3 with -framework vecLib:

	octave:1>  [-2^-.5,0^-.5], [-2,0].^-.5, octave_config_info('BLAS_LIBS')
	ans = -0.70711       Inf
	ans = 0.00000 - 0.70711i      NaN +     NaNi
	ans = -framework vecLib

FWIW, matlab is also inconsistent on IRIX:

	>> [-2^-.5,0^-.5], [-2,0].^-.5, version
	ans = -0.7071       Inf
	ans = 0.0000 - 0.7071i      Inf          
	ans = 6.5.1.199709 (R13) Service Pack 1

Paul Kienzle
pkienzle at users dot sf dot net



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------