home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.arch:11002 comp.lang.misc:3832
- Newsgroups: comp.arch,comp.lang.misc
- Path: sparky!uunet!cs.utexas.edu!sun-barr!ames!purdue!mentor.cc.purdue.edu!pop.stat.purdue.edu!hrubin
- From: hrubin@pop.stat.purdue.edu (Herman Rubin)
- Subject: Re: how to advocate new software/hardware features (Re: Hardware Support for Numeric Algorithms)
- Message-ID: <By4Dwr.78r@mentor.cc.purdue.edu>
- Sender: news@mentor.cc.purdue.edu (USENET News)
- Organization: Purdue University Statistics Department
- References: <martin.722176263@bert> <TMB.92Nov20140542@arolla.idiap.ch> <7958@charon.cwi.nl>
- Date: Sun, 22 Nov 1992 13:38:51 GMT
- Lines: 42
-
- In article <7958@charon.cwi.nl> dik@cwi.nl (Dik T. Winter) writes:
- >In article <TMB.92Nov20140542@arolla.idiap.ch> tmb@idiap.ch writes:
- > > I think it might well be worth trying to standardize an interface to
- > > arithmetic that provides you with access to the low-level arithmetic
- > > primitives that you need. However, in order to make your code truly
- > > portable, you will need to come up with an interface that takes into
- > > account that there are machines with odd word sizes. Such a facility
- > > needs to have a much more abstract and much more task-specific than a
- > > general facility for accessing "low-level features". That is precisely
- > > why I think that Rubin and Silverman are barking up the wrong tree.
-
- [Excellent discussion by Dik Winter of some of the problems deleted.]
-
- >As I said, I do not know what the real solution is, it is certainly not
- >as clear-cut as Herman thinks.
-
- I never said that the solution is clear-cut. I do not think we are
- anywhere near a passable solution. What others and I, looking at this
- problem, have suggested is that improved access to the hardware, and
- improved means of communicating to the compilers, assemblers, etc.,
- be provided, with as much flexibility as possible.
-
- Since the problem is that complicated, the means of communication must be
- greatly enhanced. My suggestion for a versatile macro processor has not,
- as far as I can tell, been met by what has been suggested. The macro
- processor I have in mind is one where I can write out a transformation,
- describe it, and you can take my description, modify it slightly to your
- environment, and implement it in your production of code.
-
- Full efficiency of code can be attained by hand-optimization, with whatever
- aid one can get from computers, for a given machine. But there are aspects
- of a program which are at least partially portable, and certainly more
- comprehensible in some other form than assembler, especially the current
- assemblers. The improvements can only be made incrementally, and whenever
- some aspect of the problem is considered, complications are likely to arise.
-
-
- --
- Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
- Phone: (317)494-6054
- hrubin@snap.stat.purdue.edu (Internet, bitnet)
- {purdue,pur-ee}!snap.stat!hrubin(UUCP)
-