PC386 BSP vs. paranoia

[Joel Sherrill]

Eric Norum wrote:
> 
> I'm new to the RTEMS/PC world and I'm having troubles.  The latest
> problem is that paranoia does not seem to run properly on my 486-DX2/66
> target (pc386 BSP).  I get a couple of `FAILURE' messages (they scroll
> by awfully quickly) and then it locks up after printing:
> 
> Testing X^((X + 1) / (X - 1)) vs. exp(2) = 7.38905609893065041e+00 as X
> -> 1.

>From running this under Linux...

Testing X^((X + 1) / (X - 1)) vs. exp(2) = 7.38905609893065218e+00 as X
-> 1.
ERROR: Severity: DEFECT:  Calculated 7.38905609548934539e+00 for
        (1 + (-1.11022302462515654e-16) ^ (-1.80143985094819840e+16);
        differs from correct value by -3.44130679508225512e-09 .
        This much error may spoil financial
        calculations involving tiny interest rates.
Testing powers Z^Q at four nearly extreme values.
 ... no discrepancies found.

> 
> Is there some code to set up the floating point mode that's missing from
> the BSP?

This is probably just the way the i387 is.  You can compile
paranoia natively like this:

cd rtems-XXX/c/src/tests/samples/paranoia
gcc -o paranoia -DCYGNUS paranoia.c -lm

and compare the output yourself.  As I recall, the i387 plays it
fast and loose. :)
 
> Thanks,
> --
> Eric Norum                                 eric.norum at usask.ca
> Department of Electrical Engineering       Phone: (306) 966-5394
> University of Saskatchewan                 FAX:   (306) 966-5407
> Saskatoon, Canada.

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel at OARcorp.com                 On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
   Support Available             (256) 722-9985