Re: Virtual procesor
- Posted by Mike The Spike <mtsreborn at yahoo.com> Mar 30, 2001
- 476 views
Well, one thing is for sure... If what David says was true, then Intel is in the wrong bussiness and should be coding virtual Pentium IV's that run on 486s instead of spending a decade engineering faster chips... See what I mean? You can't just do post-link optimisations on machine code, interpret it, and expect it to run anywhere near the speed of non-interpreted unoptimised code being executed by the processor. A virtual machine is fun for many things. For one, imagine a 'pcall()' for Euphoria, wich interprets each machine code. That way, with future support for non-intel platforms, your machine code still works. And that day might be here sooner than you expect... Mike The Spike --- jzeitlin at cyburban.com wrote: > > > > > > > > > > On Wed, 28 Mar 2001 17:07:16 -0800, David Cuny > <euphoria_coder at HOTMAIL.COM> > wrote: > > >Thomas Parslow wrote: > > >>You can't make your computer faster by using a > virtual machine > > >I recall that, in certain circumstances, the > virtual machine code actually > >runs faster than native. Typically, there's some > clever optimizer that's > >looking at the instruction blocks, and figuring out > optimal branching and > >other stuff that's beyond my comprehension. > > I've heard the same thing - but I also seem to > recall that this can only > happen in cases where you're 'emulating' a slow > processor on a fast one, > which is just the opposite of what the person who > started this thread wants > to do. > > -- > Jeff Zeitlin > jzeitlin at cyburban.com > (ILink: news without the abuse. Ask via email.) > > > > > > > > > >