Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!news.cs.indiana.edu!ux1.cso.uiuc.edu!news.iastate.edu!john
From: john@iastate.edu (Hascall John Paul)
Newsgroups: comp.arch
Subject: Re: Compilers and efficiency
Message-ID: <1991Apr17.032515.26901@news.iastate.edu>
Date: 17 Apr 91 03:25:15 GMT
References: <7117@auspex.auspex.com> <10095@mentor.cc.purdue.edu> <7207@auspex.auspex.com>
Sender: news@news.iastate.edu (USENET News System)
Organization: Iowa State University, Ames, IA
Lines: 25

In article <7207@auspex.auspex.com> guy@auspex.auspex.com (Guy Harris) writes:
}>If a hardware polynomial evaluation takes longer than an explicit loop,
}>it is not the fault of the instruction, but of the implementation.

}What if a hardware polynomial evaluation takes longer than an *unrolled*
}loop, with the zero-coefficient terms taken out?  Or do you have an
}implementation in mind that takes zero cycles to skip the terms with
}zero coefficients?

Doesn't seem so impossible:

           POLY   CMASK,CTABLE,X,RESULT

where the CMASK operand is a bitmask which indicates which elements
of the CTABLE array are non-zero.  [probably this instruction would
still be a waste of real estate for most]

John

--
John Hascall                        An ill-chosen word is the fool's messenger.
Project Vincent
Iowa State University Computation Center                       john@iastate.edu
Ames, IA  50011                                                  (515) 294-9551