On Friday, 07/21/2006 at 06:55 ZE2, Rob van der Heij <[EMAIL PROTECTED]> wrote: > Those improvements in z/VM 5.2 would only apply to virtual n-way with > n>2, isn't it? > For n=2 that improvement was already with earlier releases. I have > read that discussion several times but the "nice improvement" was not > obvious to me.
Correct. I would be remiss if I didn't point out that the benefit of the change (both in z/VM and in Linux) is proportional to frequencey of diag 44s, not the number of CPUs. If you aren't burning CPU in CP for diag 44s, then there is minimal benefit since your workload isn't "excessively" spinning on locks. Alan Altmark z/VM Development IBM Endicott ---------------------------------------------------------------------- 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
