... or there's some implication of AMODE switching that has not occurred to us. "It takes that long because it must _____."
I would think the cost might depend on how you measured it. (Doesn't it for every instruction nowadays!) I could certainly see SAMxx invalidating any branch prediction, because the effective (for address generation) contents of the registers changes. Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Blaicher, Christopher Y. Sent: Wednesday, May 04, 2016 2:45 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: How expensive is AMODE switching? All I can say is WOW, also. There are only two reasons I can think of to have it cost that much. 1) It is a milli-code instruction, or 2) It purges the pipeline. If that time is for each instruction, then I tend to think it is milli-code. If the time is for a SAM64/SAM31 pair, then I think it is a hardware level instruction that happens to purge the instruction pipeline. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
