Edward Jaffe wrote:
The first several pages would have to explain the pipeline design and the delays that can occur. After that, specific instruction timings would have to assume the pipeline is flowing freely.
Management doesn't care about individual instruction timings, but the overall throughput and response time. The service bureaus I worked at had a standard job stream, with matching initiator setup, that was used to determine new processor performance and rates for CPU, I/O, etc. These days I'd add some transaction processing to that. A similar set of benchmarks could and should be used by IBM for processors, to give management an idea of what they're getting. I realize there are different ways of configuring and capping multiple processors, but several reasonable combinations should be adequate to determine "management throughput" <g>
Gerhard Postpischil Bradford, VT ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
