matrix decompositions fails on Max os X 10.2

matrix decompositions fails on Max os X 10.2

[Erik Ø. Sørensen]

Matrix decompositions seem to fail:

PASS:   bidiag_decomp m(5,3)
(  9,  7)[0,0]:    1.00000095367431574                        1
(  9,  7)[0,1]:   0.500000476837157981                      0.5
(  9,  7)[0,3]:   0.250000238418579046                     0.25
(  9,  7)[2,1]:   0.250000238418578935                     0.25
(  9,  7)[3,4]:    0.12500011920928944                    0.125
(  9,  7)[4,3]:   0.125000119209289468                    0.125
(  9,  7)[5,2]:   0.125000119209289468                    0.125
(  9,  7)[6,1]:    0.12500011920928944                    0.125
FAIL:   bidiag_decomp m(9,7)
(  2,  2)[0,1]:   0.500000476837157759                      0.5
(  2,  2)[1,0]:   0.500000476837157759                      0.5
FAIL:   bidiag_decomp hilbert(2)
(  3,  3)[0,0]:    1.00000095367431574                        1

and it goes on, some of the decompositions passing, many of them failing,
seems to be mainly in the 7th  to  9th significant digits. Would be very
grateful for any help with what causes this -- I have never had any
problems with gsl on the linux laptop which has suffered massive hardware
failure.

Thanks,
Erik

-- 
Erik Øiolf Sørensen  <mailto:erik.sorensen@nhh.no>
phd student <http://www.geocities.com/erik_oiolf/>
Department of Economics, Norwegian School of Economics
Visiting Queen's University, Kingston, Ontario.
phone, office: (+1) 613-533-6000, ext. 74030.

Re: matrix decompositions fails on Max os X 10.2

[Brian Gough]

Erik Ø. Sørensen writes:
 > Matrix decompositions seem to fail:
 > 
 > PASS:   bidiag_decomp m(5,3)
 > (  9,  7)[0,0]:    1.00000095367431574                        1

Try compiling that directory without optimisation, to see if it is
compiler-related.  I think it is likely to be, since the code works on
other platforms.

Re: matrix decompositions fails on Max os X 10.2

[Erik Ø. Sørensen]

[-- Attachment #1: Type: text/plain, Size: 1305 bytes --]

On Friday, September 6, 2002, at 02:01  PM, Brian Gough wrote:
> Try compiling that directory without optimisation, to see if it is
> compiler-related.  I think it is likely to be, since the code works on
> other platforms.

I have recompiled with less optimization, and with -O0 the matrix 
decompositions work. This, however, allowed me to find another mistake. 
Compiling the tests with no optimization, they stop further down, in 
the specfunc directory. Only this failure is reported:

FAIL: gsl_sf_conicalP_0_e(0.0, 100.0, &r)
   expected:   0.3009174858819926
   obtained:    0.300917485881992   7.027199282265925e-16  2.33526e-15
   fracdiff: 1.014601497407329e-15
   value not within tolerance of expected value
      0.30091748588199202  7.02719928226592486e-16

This is the only reported failure when the full make check is compiled 
with no optimization. And it is awfully close to machine epsilon.

I must retract a statement I made in an earlier posting about the 
beneficial effect of Apple's upgrade of gcc. It seems I must 
inadvertently have compiled the checks with less than -O2 when the 
internal compiler error first disappeared. I am sorry for that.

--
Erik Ø. Sørensen,
phd student, Economics
Norwegian School of Economics and Business Administration

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: Erik Ø. Sørensen.vcf --]
[-- Type: text/directory; x-unix-mode=0644; name="Erik Ø. Sørensen.vcf", Size: 1030 bytes --]

\0B\0E\0G\0I\0N\0:\0V\0C\0A\0R\0D\0
\0V\0E\0R\0S\0I\0O\0N\0:\03\0.\00\0
\0N\0:\0S\0ø\0r\0e\0n\0s\0e\0n\0;\0E\0r\0i\0k\0 \0Ø\0.\0;\0;\0;\0
\0F\0N\0:\0E\0r\0i\0k\0 \0Ø\0.\0 \0S\0ø\0r\0e\0n\0s\0e\0n\0
\0O\0R\0G\0:\0N\0o\0r\0w\0e\0g\0i\0a\0n\0 \0S\0c\0h\0o\0o\0l\0 \0o\0f\0 \0E\0c\0o\0n\0o\0m\0i\0c\0s\0;\0
\0T\0I\0T\0L\0E\0:\0p\0h\0d\0 \0s\0t\0u\0d\0e\0n\0t\0
\0E\0M\0A\0I\0L\0;\0t\0y\0p\0e\0=\0W\0O\0R\0K\0;\0t\0y\0p\0e\0=\0p\0r\0e\0f\0:\0e\0r\0i\0k\0.\0s\0o\0r\0e\0n\0s\0e\0n\0@\0n\0h\0h\0.\0n\0o\0
\0T\0E\0L\0;\0t\0y\0p\0e\0=\0W\0O\0R\0K\0;\0t\0y\0p\0e\0=\0p\0r\0e\0f\0:\01\0 \06\01\03\0 \05\03\03\0 \06\00\00\00\0x\07\04\00\03\00\0
\0T\0E\0L\0;\0t\0y\0p\0e\0=\0H\0O\0M\0E\0:\01\0 \06\01\03\0 \05\04\05\0 \03\03\09\07\0
\0T\0E\0L\0;\0t\0y\0p\0e\0=\0F\0A\0X\0:\01\0 \06\01\03\0 \05\03\03\0 \02\02\05\07\0
\0A\0D\0R\0;\0t\0y\0p\0e\0=\0H\0O\0M\0E\0;\0t\0y\0p\0e\0=\0p\0r\0e\0f\0:\0;\0;\02\03\03\0 \0J\0o\0h\0n\0s\0o\0n\0 \0S\0t\0.\0;\0K\0i\0n\0g\0s\0t\0o\0n\0;\0O\0n\0t\0a\0r\0i\0o\0;\0K\07\0L\0 \01\0Y\02\0;\0C\0a\0n\0a\0d\0a\0
\0A\0D\0R\0;\0t\0y\0p\0e\0=\0W\0O\0R\0K\0:\0;\0;\0Q\0u\0e\0e\0n\0'\0s\0 \0U\0n\0i\0v\0e\0r\0s\0i\0t\0y\0\\0n\0E\0c\0o\0n\0o\0m\0i\0c\0s\0 \0D\0e\0p\0a\0r\0t\0m\0e\0n\0t\0\\0n\0D\0u\0n\0n\0i\0n\0g\0 \0H\0a\0l\0l\0\\0,\0 \09\04\0 \0U\0n\0i\0v\0.\0 \0a\0v\0e\0n\0u\0e\0;\0K\0i\0n\0g\0s\0t\0o\0n\0;\0O\0n\0t\0a\0r\0i\0o\0;\0K\07\0L\0 \03\0N\06\0;\0C\0a\0n\0a\0d\0a\0
\0U\0R\0L\0:\0h\0t\0t\0p\0:\0/\0/\0w\0w\0w\0.\0g\0e\0o\0c\0i\0t\0i\0e\0s\0.\0c\0o\0m\0/\0e\0r\0i\0k\0_\0o\0i\0o\0i\0l\0f\0/\0
\0E\0N\0D\0:\0V\0C\0A\0R\0D\0

[-- Attachment #3: Type: text/plain, Size: 1 bytes --]