Re: stats on unresolved dependencies

Beatrice Apvrille <apvrille@cri.ensmp.fr>
Mon, 31 Jan 1994 16:20:30 GMT

          From comp.compilers

Related articles
stats on unresolved dependencies gxj@dcs.ed.ac.uk (Graham Jones) (1994-01-24)
Re: stats on unresolved dependencies petersen@sp90.csrd.uiuc.edu (1994-01-28)
Re: stats on unresolved dependencies apvrille@cri.ensmp.fr (Beatrice Apvrille) (1994-01-31)
| List of all articles for this month |
Newsgroups: comp.compilers
From: Beatrice Apvrille <apvrille@cri.ensmp.fr>
Keywords: optimize, bibliography
Organization: Compilers Central
References: 94-01-096
Date: Mon, 31 Jan 1994 16:20:30 GMT

|> Dependencies that can not be resolved at compile time force compilers to
|> generate serial code. I am interested in finding out how often (in
|> scientific applications), and in what programs they tend to occur.




I suggest that you read the following article :


@INPROCEEDINGS{Bane:93,
AUTHOR = "Banerjee, Uptal
and Eigenmann, Rudolf
and Nicolau, Alexandru
and Padua, David A.",
TITLE = {Automatic Program Parallelisation},
BOOKTITLE = {Proceedings of the IEEE},
YEAR = 1993,
MONTH = feb,
VOLUME = 81,
NUMBER = 2
}


which is an overwiew of automatic program parallelization. The last
section surveys experimental studies.


If you understand french, you could also read the following thesis, which
(among others) gives a comparison between several parallelizers (from
universities), on the number of independencies they are able to discover :


@PHDTHESIS{Yang:93,
AUTHOR = {Yang, Yi-Qing},
TITLE ={Tests des De'pendances et Transformations de Programme},
SCHOOL = "Universite' Paris VI ",
YEAR = 1993,
MONTH = {Novembre}
}


I hope this will help you. Don't hesitate to ask for more questions, I
will try to answer.


Be'atrice Apvrille, apvrille@cri.ensmp.fr


Centre de Recherche en Informatique, Ecole des Mines de Paris
35, rue Saint-Honore', 77305 Fontainebleau Cedex, France
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.