-----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Charles Mills Sent: Thursday, March 29, 2007 6:08 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: What's a CPU second?
Gee, thanks everybody for telling me about MP overhead and SRB time and time-slicing and the rationale for MSUs. <g> Would someone care to go out on a limb and give me a direct yes or no on my basic question? If a jobstep is reported to have used .02 CPU seconds, for example, that means it spent .02 seconds executing on one CPU, or a total of .02 seconds spread across multiple CPUs -- the number reported is irrespective of the total MSU horsepower of the box, correct? Presumably on a faster CPU, it would use fewer seconds -- but not fewer on a box rated at more MSUs due to more CPUs of the same speed. Is that right (neglecting MP overhead, marketing MSUs, SRB time, etc., etc.)? <SNIP> "... the number reported is irrespective of the total MSU horsepower of the box, correct?" Yes. "Presumably on a faster CPU... ...?" Yes. A second of CPU time is a second regardless of how many instructions get processes in that second -- which means that a faster, slower, cached (or not), pipelined (or not) CPU is still measured in standard seconds. It is the same unit that you and I deal with on our respective watches (assuming you wear one and don't used a cell phone to get the time). Regards, Steve Thompson PS. This is true for today. Get one of those quantum computers and all bets are off. And has anyone seen Schrodinger's cat lately? ---------------------------------------------------------------------- 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
