[daip] Re: compiler gnu gcc 3.4.2
Michael Rupen
mrupen at milkyway.gsfc.nasa.goV
Mon Oct 18 10:26:51 EDT 2004
Hi Eric,
I can't believe you were able to track down something so subtle, so
quickly...many thanks. I've re-installed AIPS here using gcc 2.95.3,
and just for fun re-ran the Y2K test. It's a little slower (AIPS mark of 100)
and more accurate (all accuracies are better, with mode 'T' being up to 1.4
bits better [on IMCLN]).
IMEAN now works as expected.
Cheers,
Michael
On Fri, 15 Oct 2004, Eric Greisen wrote:
> When in debug or in optimize
>
> $APLSUB/ENORM.FOR
>
> C
> INTEGER I
> DOUBLE PRECISION AGIANT, FLOATN, RDWARF, RGIANT, S1, S2,
> * S3, XABS, X1MAX, X3MAX, ZERO, ONE
> DATA ONE, ZERO, RDWARF, RGIANT /1.0D0, 0.0D0, 3.834D-20, 1.304D19/
> C-----------------------------------------------------------------------
>
> has a value for ONE of 0
>
> $APLSUB/LMSTR1.FOR
>
> C
> INTEGER MAXFEV, MODE, NFEV, NJEV, NPRINT
> DOUBLE PRECISION FACTOR, FTOL, GTOL, XTOL, ZERO
> DATA FACTOR, ZERO /1.0D2, 0.0D0/
> C-----------------------------------------------------------------------
>
> has a value for FACTOR of 0
>
> God knows where else this will hit. The use of DATA here is not
> essential - the latter works when I put a FACTOR = 1.D2 in the regular
> code. The former worked when I used gdb to set ONE to 1.D0.
>
> Can you forward this to your gnu friends? Their parser is failing.
>
> Eric Greisen
>
More information about the Daip
mailing list