Re: Why are intepreters so slow today

Stefan Monnier (monnier@di.epfl.ch)
20 Apr 1994 04:46:01 +0200

In article <nagleCoInD3.4n3@netcom.com>, John Nagle <nagle@netcom.com> wrote:
> As I said, I could live with a 10x slowdown. But 1000x slowdowns
> indicate something is very wrong, like doing a storage allocation for every
> floating point operation.

#define NITPICK

storage allocation can be and often is much less than 1000
instructions. 10 would be more like it for a copying collector. So the
1000x must come from somewhere else !

#undef NITPICK

Stefan

-- 

----------------------------------------------------- -- On the average, people seem to be acting normal -- -----------------------------------------------------