I'm still waiting for the DWIMNWIS instruction.
> Date: Fri, 24 Jun 2016 15:15:08 -0400
> From: [email protected]
> Subject: Re: Simple assembler question
> To: [email protected]
> 
> On 23 June 2016 at 17:51, Phil Smith III <[email protected]> wrote:
> >
> > With all of the 273 new formats of LOAD, I assume this is hiding in there 
> > somewhere:
> 
> > I have a value in grande register 3. I need the high-order bits in 32-bit R0
> > and the low-order bits in 32-bit R15.
> 
> The RxSB... instructions are marvels in several ways, not least for
> having in some cases five operands, and in some cases 7-character
> names. And because there are some extended assembler mnemonics that
> bear no resemblance to the name or description of the base
> instruction.
> 
> RISBLG[Z] is Rotate then Insert Selected Bits Low [and Zero remaining bits].
> 
> There is an extended mnemonic LLHFR (presumably Load Low from High
> Fullword Register) that generates
> 
> RISBLGZ R1,R2,0,31,32
> 
> but even the RISBLGZ mnemonic is an extension of the base RISBLG that
> turns on the "Zero remaining bits" flag.
> 
> These are actually great instructions for all sorts of things, and are
> recommended for performance reasons in Kevin Shum's IBM z Systems
> Processor Optimization Primer. If there's a catch, it's that they
> require the High Word Facility. I must admit I spent an embarassing
> number of minutes wondering why there are no versions of these
> instruction that rotate right...
> 
> Oh yes, so what I'm trying to say is that two variations on RISBxG[Z]
> would probably do just what you need.
> 
> Tony H.
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN
                                          
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to