[Openmcl-devel] single float support in OpenMCL

Gary Byers gb at clozure.com
Tue Aug 24 01:15:51 UTC 2004



On Mon, 23 Aug 2004, Gary King wrote:

> It's my understanding that unlike MCL, OpenMCL supports both IEEE
> single and double floats.

Both implementations support both IEEE single and double floats.

In MCL, a lisp SHORT-FLOAT denotes an IEEE single float, and the
other CL float types all denote an IEEE double.

In OpenMCL, lisp SHORT-FLOATs and SINGLE-FLOATs are IEEE singles,
while DOUBLE-FLOAT and LONG-FLOAT are IEEE doubles.

> If this is correct, is there some fast
> mechanism to test a number for single-float-hood (similar to
> ccl::double-float-p)?
>

(CCL::DOUBLE-FLOAT-P x)  is exactly the same as

(TYPEP x 'DOUBLE-FLOAT).

This doesn't tell you anything about whether a number is an IEEE single,
an IEEE double, or something else.

In MCL, the CL types SINGLE-FLOAT and DOUBLE-FLOAT are synonymous; in
most other implementations, they're disjoint.

> Thanks,
> --
> Gary Warren King, Lab Manager
> EKSL East, University of Massachusetts * 413 577 0176



More information about the Openmcl-devel mailing list