[ale] sorta [OT] IO vs clock cycles

[Joseph A Knapka]

Cade Thacker wrote:
> Need a little bit of knowledge from you guys/gals. I got in a discussion
> about running Java in an enterprise system, and one of the gentlemen I
> repect a great deal said that Java would be much happier running on an NT
> box with a very high clock cycle, then a Unix work horse from IBM. His
> reasoning was that the Unix box is built(and build well) for IO intensive
> tasks, DB type stuff, where NT is build to take advantage of clock speed.

Why does he think I/O performance has to be exclusive of computational
performance? If the OS isn't totally lame, both of those will be more
dependent on the nature of the hardware (how fast is the CPU, do
the peripherals permit DMA or do they require PIO?) than on the OS.

Furthermore, *any* process that has to run in an "enterprise"
environment (by which I take you to mean "has to stay up a lot")
is *not* going to be happier on NT.

Finally, Linux is built to be fast (can't speak for commercial
Unices, but I presume their implementors don't strive for
pokiness). Linux has much faster fork and thread-creation times
than NT, IIRC. Lots of stuff in the kernel that could be made
faster *is* being made faster in the 2.5 series, so 2.6/3.0
should show significant improvement.

> Therefore the JVM would be happier on NT.
> 
> On the same topic, we always hear that the mainframe downstairs has less
> horse power then our PCs, but yet the mainframe handles Millions of
> transactions a second, and my PC dogs out running 5 browsers at once.
> Please explain.

The mainframe probably has less horsepower per cubic centimeter,
as it were, but a lot more horsepower overall in terms of the
quality of equipment (lots of RAM, wide and fast I/O channels,
etc) than the PC. Also the mainframe software was implemented
by people who really care about their product being *fast* and
*stable* (because they're dependent upon long-term corporate
customers who demand this); unlike the vast majority of PC
software, whose authors are mainly concerned with making a buck
from people who are accustomed to rebooting their Redmond crap on
a regular basis.

> Where do you guys think linux would fall in this debate?

I posted an example of Linux-vs-Windows stability running
lots of Java tasks the other day. Linux wins by default:
it finished the job, and Windows XP didn't even get out
of the starting gate before locking up hard.

(FWIW, my experience with XP thus far has been most disappointing.
I'd rather run NT4 if I have to use a Redmond product.)

Cheers,

-- Joe


---
This message has been sent through the ALE general discussion list.
See http://www.ale.org/mailing-lists.shtml for more info. Problems should be 
sent to listmaster at ale dot org.