Re: Status of FLTK on MacOS X [test tolerances]

[Ben Abbott]

On Jan 6, 2010, at 2:24 PM, Michael D Godfrey wrote:

> Good.  It seems to me that the current tolerance values are working as they 
> should:
> on systems with stable software the tests pass, and (at least in the cases 
> that I have
> looked at) are consistent with IEEE floating point behavior.  The IEEE error 
> behavior,
> thanks to Kahan, is very well defined.  (And, these days, it is Kahan's 
> students who do 
> most of the floating point core designs.) So, the most likely cause of test 
> failures is software.
> Unfortunately, identifying where in the software things have gone wrong is 
> not easy.
> I suspect that in my Mac there may be some library confusion since I have 
> noticed
> that on some occasions I have found the following:  
> 1. Install some version of gfortran.  Get correct check results.
> 2. Try another gfortran. Failure.
> 3. Reinstall the first gfortran.  Failure...
> Or course there are other possibilities for error in this loop...  One clear 
> indication
> of problems is that fairly often ./configure fails on the CDOTU test.  
> Changes that
> should not affect this actually make the failure go away.  More needs to be 
> learned.
> 
> I recently updated sed to 4.2 and flex to 2.5.4.  This had noticeable 
> effects, not all good.
> But, some of the problems may be due to changes that should not have been 
> made.
> There are still way too many variables floating around in the Mac 10.6 
> systems.
> (Well, at least in mine: I have too much stuff in  /usr/local/bin and ../lib 
> much of it
> left over from when the system was running 10.5.) 
> 
> Michael

Hmmm ... I'd mentioned that the chol.cc test results appeared to me to be a bit 
random ... sometimes passing and other times not. After you comment, it 
occurred to me that that what was changing might have been my build process 
(I'd been switching between different versions of gcc). I'd venture a guess 
that the tests for chol.cc will reliably pass if Octave is built using Apple's 
gcc 4.2 (since I am using vecLib for my blas/lapack).

The two tests that have given me trouble end with the assertions below.

        %! assert(norm(A1(p,p) - A,Inf) < 1e1*eps('single'))

        %! assert(norm(A1(p,p) - single(Ac),Inf) < 1e1*eps('single'))

With my current build, the errors for each repeatable and are equal to 
7.0781e-08 and 1.4226e-07. The tolerance is 10*eps = 1.1921e-06.

I think it is safe to conclude your deduction is correct!

Thanks
Ben