[gmpi] Re: Reqs final draft

Paul Davis wrote:

In all cases it should define a FULL FPU control state, since the only goal here is to allow both host & plugins to be able to round to nearest (the most common) without having to change the FPU state.

I think we all agree that all exceptions should be masked, rounding to nearest (standard in all languages I guess). Now as for the precision.. 64bit or 80bits? It still matters for divisions.



i don't think we all agree that exceptions should be masked, but yes,
its definitely very important that the GMPI makes it clear what the
FPU state is going to be, and how it can be set (if at all) and reset
by plugins.

--p


One approach which might be possible is to define a sensible default state for the host to provide AND a preferences list (in a non-processor specific format) with which the plugin could use indicate to the host what it needed, with the host being responsible for dealing with the processor specific bits. This would be of real benefit to plugin developers, particularly those producing cross-platform code, since they wouldn't then need to delve into the murky world buried under their hi-level language of choice (unless they really wanted to).

--
These are my own views and may or may not be shared by my employer
/*********************************************************************/
/* Dave Malham   http://music.york.ac.uk/staff/research/dave_malham/ */
/* Department of Music    "http://www.york.ac.uk/depts/music/";       */
/* The University of York  Phone 01904 432448                        */
/* Heslington              Fax   01904 432450                        */
/* York YO10 5DD                                                     */
/* UK                   'Ambisonics - Component Imaging for Audio'   */
/*                    "http://www.york.ac.uk/inst/mustech/3d_audio/"; */
/*********************************************************************/

Other related posts: