[Ecls-list] Immediate short-float for ECL

[Juan Jose Garcia-Ripoll]

2006/9/30, Lars Brinkhoff <lars at nocrew.org>:
> [Resend.  Note: this is just for review and discussion.  I will keep
> working to remove test failures, etc.  There are also some stray fixes
> unrelated to floats in this patch.]
>
> Here's a preview of my patch which adds an immediate short-float type,
> temporarily called SI::IMMEDIATE-SHORT-FLOAT, to ECL.  It's not 100%
> complete yet, and there are test regressions, but I believe most
> things work.

I had a look at your code. You are clearing some bits in the floating
point number to replace them with the type tag. That does not seem
very safe in 32-bit computers. Isn't the last bits where the exponent
is stored?

I think that your code is much better suited for 64-bit environments,
where one has plenty of bits to spare, but this code wouldn't make use
of those extra bits:
+       ({                                                      \
+               union { float flo; cl_fixnum fix; } _tmp__val_; \
+               _tmp__val_.flo = (n);                           \
+               _tmp__val_.fix &= ~IMMEDIATE_TAG;               \
+               _tmp__val_.fix |= IMM_SHORTFLOAT_TAG;           \
+               (cl_object)_tmp__val_.fix;                      \

Perhaps something like
                 struct { int32_t fix; float flo } _tmp__val_;
would be better there.

I would also replace the name immediate-floats with short-float and
single-float (t_imm_shortfloat -> t_shortfloat). In 32-bit
environments, when configured with --enable-short-floats they would be
different and have different tags. In 64-bits there would not be
short-floats, all single-floats would be immediate and the tags would
be the same (t_shortfloat == t_singlefloat), so that the compiler
collapses the code for short-floats and single-floats.

What do you think?

Regards,

Juanjo

-- 
Max-Planck-Institut für Quantenoptik
Hans-Kopfermann-Str. 1, Garching, D-85748, Germany
Phone: +49 89 32905 345   Fax: +49 89 32905 336
http://www.mpq.mpg.de/Theorygroup/CIRAC/