Soup -- thanks, now we are talking. Its good to have some
numbers like this as a starting point. But I'm not sure if his
test with mathematica:
`The real proof that faster is better comes when you apply a high-performance Power Mac to a computation-heavy job. Rather than pick from the usual graphics applications, I selected Wolfram
Research Inc.'s Mathematica 3.0, which was recently benchmarked by Karl Unterkofler at fampm201.tu-graz.ac.at;
is what I'm talking about. You see, for a well-compiled binary
running at the assembly code or machine code level yes we'll
see that RISC is better than pentium etc, but that is the processor
which we are seeing, and some things about the OS handling of
tight resources, but not the kind of everyday slugishness I
was talking of. Once you start the compu-intensive mathematica
job you wont see much differenece between win95 and winNT
for example beause you arent opening and closing windows,
using the OS, youre just running the processor which is the same.
The big improvement with mathematica on intel under next is
nice to see but may just mean that a big program like mathematica
was probably written in a high level language and the compiler
for the unix-origin OS were more efficient than windows ones, i.e
Wolfram began life in unix and ported perhaps not efficiently to
windows. Still its a good point (which Philip Lee made) that
big programs will tend to port well into Rhapsody and your numbers
confirm this to me. If it shows up on a range of programs
then I'd be even happier.
But what the everyday user sees, who doesnt run mathematica is things
like I mentioned in my last post. Maybe the only way is to
wait for the product and see. I know only on my computer yes
big compu-intesive jobs run great but I still dread turning
on Next now because of all the general waiting around. Anyhow,
your link is a good frame of reference. Thanks,
Shahn |
