On 2/4/2014 6:22 AM, John Gilmore wrote:
Any "complete implementation" that assigns more of it to functionally meaningful categories is to be strongly encouraged. Misattribution, inclusion in the wrong category, would, on the other hand, be a step backward. I strongly suspect that this is really EJ's concern, but he can of course speak for himself.
My concern was a simplistic one. I hoped that the (still experimental) instruction counts were handled in a way that was analogous to what's done today for CPU time, where smart choices were made to ensure things rightly classified as "system overhead" are not attributed to a running job. Obviously, there is still much work to do to make them work as a viable alternative to CPU time.
-- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 http://www.phoenixsoftware.com/ ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
