On Thursday, 02/21/2008 at 04:07 EST, Rob van der Heij <[EMAIL PROTECTED]> wrote: > On Thu, Feb 21, 2008 at 2:27 AM, Alan Altmark <[EMAIL PROTECTED]> wrote: > > > Hard evidence of the performance of emulated FBA can be found at: > > http://www.vm.ibm.com/perf/reports/zvm/html/530scsi.html > > > > Don't just look at the tables; read the text that goes with them. > > Since I have not been able yet to measure it myself, I've hesitated > before responding to this thread. So you're excused if you want to > skip my long post with no real numbers... ;-) > > As with some of the measurements from others, the significance for > Linux on z/VM installations isn't as much as we would hope. The most > significant part in these is a potential reduction in CPU usage. > That's good news. The "native SCSI" was meant as a realistic option to > hold the z/VM data while Linux has its data on FCP directly. That is > also the motivation for the measurements in the performance report. > The context that Sir Santa introduced is to have Linux run on emulated > devices. A simple iozone test does not address the complexity of the > performance issues involved.
All very true. I just wanted to head off the inevitable discussion about "overhead" which is often disguised under the broader term "performance". > Most of the measurements folks seem to do are "top speed" measurements > (like maximum single-user throughput). Of course. The performance measurements we give are meant to allow you to gauge whether a new release is "better" or "worse" than a prior release in a specific area. The deltas are more important, IMO, than the absolute numbers, since the absolute values are based on controlled laboratory conditions (STP!). Holding the "temperature and pressure" constant, the deltas reflect code changes. They should not be used to predict the performance of a workload? And consider what Humpty Dumpty might say about "performance". Does he mean response time? Throughput? CPU seconds? I/O rate? They Performance Report can provide guidance as to whether the workload you have will be affected by the changes we make. You may have a workload that is positively affected by Enhancement #1 and negatively affected by Enhancement #2. How to know? MEASURE IT. Create a repeatable workload that represents the needs of YOUR business. As new releases come in, study the Performance Report, consider whether a configuration change is warranted, then measure the effect on YOUR business. Alan Altmark z/VM Development IBM Endicott
