RE: SAINT on the web

From: Giorgio Zoia (Giorgio.Zoia@epfl.ch)
Date: Sat Sep 18 1999 - 05:06:08 EDT


At 10:01 PM 9/17/99 -0400, Michael Gogins wrote:
>This sounds interesting. Would it be possible to take the opcode sources
>from sfront and paste them into the SAINT interpreter to make a complete
>SAOL runtime?
>

The SAINT virtual instruction set is built in a vectorial form, and opcodes
are fully related to this form. Opcodes then must be able to work on
a vector of samples or variables from "start" to "end"; essentially you
have to deal with block-by-block and sample-by-sample executions.
This in some way makes the instruction implementations a little more
complicated, but gives you the advantage to work on vectors when possible
and to reuse the complete software just changing this implementations (well,
theoretically, but you understand what I mean, just a few adjustments).
In the end what will be released is the portable part, plus some small
number of execution routines for PC. Perhaps math opcodes just to have
an example of how it works. What will not be released, at least in a near
future, is the specific vectorial instruction implementations for our
chosen platform (and the AudioBIFS, but this is not important for this
list, I guess).

I hope this makes things clearer, for now.

Giorgio

__________________________________________________________________
Giorgio ZOIA

Integrated Systems Laboratory - DE/LSI - EPFL
CH-1015 Lausanne - SWITZERLAND

Phone: + 41 21 693 69 79 E-mail: Giorgio.Zoia@epfl.ch
Fax: +41 21 693 46 63
__________________________________________________________________



This archive was generated by hypermail 2b29 : Wed May 10 2000 - 12:15:36 EDT