Re: Making Guile slower?

[Greg Harvey <Greg dot Harvey at thezone dot net>]

Mark Probst <schani@mips.complang.tuwien.ac.at> writes:

> On Thu, Dec 16, 1999 at 05:57:25AM -0330, Greg Harvey wrote:
> > I also tried it malloc as the thing being done to the cell, instead of
> > just setting it to an inum, but I got some odd results. In fact, I got
> > overall odd results:
> > 
> > With nimp check: 1000000 iterations
> > Time: 0.010000
> > Without nimp check: 1000000 iterations
> > Time: 0.100000
> > [greg ~/devel/tests]$ ./imp 1000000 100
> > With nimp check: 1000000 iterations
> > Time: 0.010000
> > Without nimp check: 1000000 iterations
> > Time: 0.110000
> 
> rather interesting numbers. what processor did it run on? which
> compiler did you use with which flags? i used egcs-2.91.66 with
> -O9 on an alpha 21164pc and it gave:
> 
> schani@quinta:/tmp > ./guiletest 
> With nimp check: 10000000 iterations
> Time: 3.050000
> Without nimp check: 10000000 iterations
> Time: 2.973872
> 
> the same compiler, same flags on a pentium2 gave:
> 
> schani@vader:/mnt/homes/nethome/hansolo/schani > ./guiletest 
> With nimp check: 10000000 iterations
> Time: 1.490000
> Without nimp check: 10000000 iterations
> Time: 1.390000
> 
> if everything else fails, look at the generated assembler code.

An amd k6, pgcc 2.91.66, -O6 -mk6. I did look at the asm to make sure
it was actually doing everything instead of agressively optimizing out
a bit in the imp_test, because I was quite surprised (obviously ;). I
could only guess at why this should be the case (amd's processors tend
to be funky like that :).

-- 
Greg