guile-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: What happened to the ex-Guile VM?


From: Keisuke Nishida
Subject: Re: What happened to the ex-Guile VM?
Date: Sat, 24 Mar 2001 22:03:57 -0500
User-agent: Wanderlust/2.4.0 (Rio) SEMI/1.13.7 (Awazu) FLIM/1.13.2 (Kasanui) Emacs/21.0.99 (i686-pc-linux-gnu) MULE/5.0 (SAKAKI)

At 25 Mar 2001 03:58:59 +0200,
Marius Vollmer wrote:
> 
> > > Nice!  With Lightning and hand coding the assembler (probably
> > > pessimizing it on the way), I get a speedup of 28 compared to the
> > > debugging evaluator and 11 compared to the normal evaluator.  Since
> > > you get a speedup of about 7.5, this speaks very much for your VM.
> > 
> > I realized that I had compiled the VM with no optimization (-O0).
> > With -O2 and the fast engine, I get a speedup of about 13.7:
> > 
> >   address@hidden> ,time (fib 30)
> >   $2 = 1346269
> >   clock utime stime cutime cstime gctime
> >    1.73  1.72  0.00   0.00   0.00   0.00
> > 
> > Just for your information.
> 
> Woa!  This makes machine code look quite bad.  Why isn't it faster?  I
> think a native machine code compiler will only win big when it can
> generate inline code for common operations like car and fixnum
> arithmetic and can do type analysis like Stalin (supported by
> declarations).

Actually, machine code is faster.  The speedup is 13.7 compared to the
debugging evaluator (where the speedup by Lightning is 28), so Lightning
still seems to be twice as fast as the VM.

But with larger programs, there may not be significant differences,
because there are probably less CPU cache misses with bytecode.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]