Would not L/ST be faster than 4-MVIs? Most every control-block I’ve seen bearing an eye-catcher is Fullword aligned in storage which makes for optimal execution.
Keven > On Oct 24, 2020, at 16:57, Tony Thigpen <t...@vse2pdf.com> wrote: > > I was thinking that the MVI would be some sort of special character, > limiting the tag to just 3 unique characters. But now I understand it better. > My mistake. > > Tony Thigpen > > Charles Mills wrote on 10/24/20 5:23 PM: >> Why not? What's wrong with MVI/MVC for 3? >> Given that it would involve no storage references beyond the target and >> i-cache, I wonder if the fastest way to set four bytes, without the string >> appearing in storage, might not be four sequential MVI's. >> Charles >> -----Original Message----- >> From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] >> On Behalf Of Tony Thigpen >> Sent: Saturday, October 24, 2020 2:07 PM >> To: ASSEMBLER-LIST@LISTSERV.UGA.EDU >> Subject: Re: Reversed string macro >> I kinda like the concept, but a lot of literals we use are just 4 >> characters, so it does not scale down easily.