> Ron Rogers wrote:
> some simple tests I ran with timing on.
Not Linux but a significant difference in your hardware. Thus this
simple test does not show that Linux is any faster or better.
Case in point. I run a query on a Sunfire V440 (4 CPUs) and run the
same query on the same table on a Sunfire V20z. The V440 is a
higher-end server. The V20z is an entry-level server (and
significantly cheaper). The select count is much faster on the V20z.
Reason? The V20z's uses dual AMD Opeteron CPus which is acknowledge as
one of the best and fastest 64bit CPUs around.
The select count is a single process. So what is being meassured are
how fast the CPU is (assuming the cache is hot and I/O not a factor -
which it is in this test). I.e. processing data blocks in memory which
means performance is CPU bound. The V440 has a lot more CPU capacity
than the V20z, but the V20z's CPUs are a *lot* faster. The o/s itself
is not a factor. In fact, it is seldom a factor when it comes to
performance issues.
o/s choice decisions that are made based on the notion of actual o/s
performance is flawed.
--
Billy
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This e-mail and its contents are subject to the Telkom SA Limited
e-mail legal notice available at
http://www.telkom.co.za/TelkomEMailLegalNotice.PDF
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
--
http://www.freelists.org/webpage/oracle-l
