From: Bulat Ziganshin <bulatz@...>
Date: 2002-09-30T17:29:10+09:00
Subject: Re: R

Hello Yukihiro,

Monday, September 30, 2002, 10:19:15 AM, you wrote:

YM> Hi,

YM> In message "Re: R"
YM>     on 02/09/30, "Bulat Ziganshin" <bulatz@integ.ru> writes:

YM> |main ruby speed problem - dynamic method dispatching, which don't
YM> |needed for 99% of calls in low-level code (such as matrix
YM> |multiplication or string searching). so static typing of all variables
YM> |and methods must solve the problem
YM> |
YM> |anyone disagree with above?

YM> I'm not sure.  From my simple benchmarks, method cache hit rate is
YM> over 95%.  It suggests dynamic dispatch is not the main problem.

but hash itself need time to work :)  and other dispatching code too.
you say just that TREE SEARCHING is not the issue, i am right?

consider two examples i stated above. in C and in Ruby. so...?

-- 
Best regards,
 Bulat                            mailto:bulatz@integ.ru