[PRL] Cache and I/O Efficient Functional Algorithms

[johan at ccs.neu.edu]

That reminded me of a usenet posting, more that 15 years ago, about the vagaries of cache interaction on wall-clock performance.  IIRC, this is what convinced SimonPJ to never measure wall-clock time, but instead instrument binaries to count instructions executed.

http://compilers.iecc.com/comparch/article/96-02-165


----- Original Message -----
From: "Mitchell Wand" <wand at ccs.neu.edu>
To: "PRL" <prl at lists.ccs.neu.edu>, "Rajmohan Rajaraman" <rraj at ccs.neu.edu>, "Gene Cooperman" <gene at ccs.neu.edu>, "Peter J. Desnoyers" <pjd at ccs.neu.edu>
Sent: Tuesday, August 28, 2012 7:52:49 AM GMT -05:00 US/Canada Eastern
Subject: [PRL] Cache and I/O Efficient Functional Algorithms



Interesting-looking paper from Bob Harper (not involving type theory!) --Mitch 

http://www.cs.cmu.edu/~rwh/papers/iolambda/short.pdf 
_______________________________________________
PRL mailing list
PRL at lists.ccs.neu.edu
https://lists.ccs.neu.edu/bin/listinfo/prl