[Rd] fairly OT: profiling

[A.J. Rossini]

Thomas Lumley <tlumley at u.washington.edu> writes:

> The following is from Eric Raymond's new book on Unix programming.
>
>   You'll get more insight from using profilers if you think of them less
>   as ways to collect individual performance numbers, and more as ways to
>   learn how performance varies as a function of interesting parameters
>   ... Try fitting those numbers to a model, using open-source software
>   like R or a good-quality proprietary tool like MATLAB.
>
> It's interesting to see the emphasis on statistical analysis of profiling
> data (as well as the fact that R is sufficiently well-known not to need
> explanation (or a reference or link)).

If you hang out with compiler writers more often (I do, about once a
year, thanks to an old friend who implements superscalars) you'd find
that they've felt the same way for years (and spend a good bit of time
with machine learning algorithms working on weak points, discovery,
and optimization).

(staying off topic, it's also amusing that some of the better groups
generally run on relatively ancient hardware, but that's another
story).

best,
-tony

-- 
rossini at u.washington.edu            http://www.analytics.washington.edu/ 
Biomedical and Health Informatics   University of Washington
Biostatistics, SCHARP/HVTN          Fred Hutchinson Cancer Research Center
UW (Tu/Th/F): 206-616-7630 FAX=206-543-3461 | Voicemail is unreliable
FHCRC  (M/W): 206-667-7025 FAX=206-667-4812 | use Email

CONFIDENTIALITY NOTICE: This e-mail message and any attachme...{{dropped}}