🚀 go-pugleaf

Troy Beukema wrote:
> 
> In article <31699471.5C0@noran.com>, Dave Greunke  <vid@noran.com> wrote:
> 
> >So, lets see what you're saying here.... A 100 MHz+ Pentium is faster
> >than a 66 Mhz Sparc.... Wow what an incredible revelation.
> [rest of hate-flame snipped]
> 
> I am sorry if I offended you with my observations.  I completely
> disagreed with a comment that someone posted, so I offered my
> perspective, based on real life experience with many different
> workstations.

I am not offended by your article... It would take a lot more than that
to offend me.... I am, however, concerned with the ease with which you
spew out your mis-information and the amount of general mis-information
out on the net. After reading the first article you posted I came away
thinking that you were saying that a 66 Mhz Sparc was the same class
machine as a 100 Mhz Pentium and that "Based on Your REAL world
Experiences" this was evidence that the Pentium was indeed a better
"workstation class" machine. Which of course is total crap. If you want
to make a valid comparison try comparing a 100 Mhz Sparc 20 to a 100 Mhz
Pentium. When you do that you may have some "Real World Experiences"
which make you think about the situation quite differently. You also
failed to mention what kinds of programs you used to make your
comparisons with. Oddly enough, people who compare computer systems for
a living (ie: professionals) find that kind of information very useful. 
 
> Here is some extra information a nice person sent me if you need more info
> to be convinced that the P5-P6 PCs are workstation class machines :
> 
> [sender's name snipped intentionally]
> >to provide some real world program experiences (under linux):
> >
> >IBM RS/6000-590         44 secs         95 secs
> >DEC 21064A-275          33 secs         --
> >P6-200                  20 secs         50 secs
> >
> >heavy floating point (crystallographic least squares of protein structures,
> >the kind of stuff pharmaceutical companies do in drug development - I work
> >for a chemical company in that sector).
> >
> >The P6-200 has a modest 64 MB memory and 5.5 GB disk (we have a total of
> >around 20 GB on all our P5/P6 systems, and none except for an old P5-60,
> >has less than 32 MB memory - all of course run exclusively linux)


A classic example of the problem I am trying to point out. The above may
have been useful information at one point but in the format that you
present here it makes it useless for comparisons. What program is being
run???? All we know is that it appears to be floating point intensive
which doesn't give us any feel for it's integer performance. We don't
know anything about the DEC or IBM machines (and 64 Meg is hardly a
modest amount of memory), or what the posted times represent. This makes
it impossible for anyone with any understanding of computer benchmarking
to draw any reasonable conclusions.

> 
> Assuming the P5 is around 1/2 as fast as the P6 machine, we see the PC
> is a workstation class machine, and linux is getting very close to
> a professional workstation-quality OS; not to mention the fact that
> many commercial unices are also available.  I believe the P6 is commonly
> regarded as a workstation killer, and MMX will only strengthen its position.

Again.... All meaningless jibberish.... You make assumptions you cannot
support and draw conclusions without providing information to back them
up. Based now on 2 of your postings I would again state that your
articles are at best opionions and probably closer to mis-information. I
see no reason to bother with reading any of your future articles as
there is clearly nothing to be learned from them(aside from posssibly
how not to compare computer systems). The net does not need any more of
these types of postings.

> 
> Your snide observation about clock frequency is incorrect.  Tell me if a
> 100Mhz 486 will outrun the Sparc 20-66 or a 100MHz P5.
> 
> Have a nice day.
> 
> -Troy
> 

Again you display your inexperience/ignorance.... The CPU clock is one
of the major factors in determining the performance of a system. In
response to your question: Running what kind of benchmark???  If you're
running a small program that all fits in cache and spins in a CPU loop I
would bet on the 100 Mhz machines... If you're talking about real world
types of programs that involve significant amounts of I/O as well as
computational resources and require operation in a multi-process
environment I would expect the 66 Mhz Sparc to perform quite favorably.

BTW: This will be my last posting on this subject. I have real world
concerns to attend to. If you want to learn more about benchmarking and
comparing system I suggest you tune into something like comp.benchmarks.


-- 
-------------------------------------------------------------------------------
Dave Greunke                | Internet:   vid@noran.com
Systems Administrator       | aka: webmaster@noran.com
Noran Instruments, Inc.     |
Middleton, WI.              |
(608)836-4125               |