On 10/26/06, John Campbell <[EMAIL PROTECTED]> wrote:

I was told some time ago that you don't want to migrate a heavy
CPU user to a z/VM instance... though, I suspect, an I/O bound
environment would be just peachy.

I don't know who shared that wisdom with you, but I suggest you don't
buy a car from him either.
I assure you that a CPU bound application will run in a virtual
machine not slower than native in LPAR. The only thing that causes
overhead is when the virtual machine does things where z/VM has to
ensure that virtual machines cooperate. But that's the entire point of
sharing.

Sure, if you have some task that usefully runs the CPU flat out 7x24,
then you would not enjoy sharing because the task does not leave any
spare cpu cycles that could be shared. In that case it might be better
to put the task on a machine for itself with a CPU that can be faster
because it was not designed to allow for sharing (and maybe did not
have all the stuff for reliability since it would be only that single
task to suffer when it breaks).

--
Rob van der Heij
Velocity Software, Inc
http://velocitysoftware.com/

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to