On Thu, 12 Apr 2012 10:14:34 -0500, McKown, John wrote: >"Has IBM upgraded the PL/AS(current name?) compiler to emit >Relative and Immediate instructions?".
Probably, since the R&I facility was required for OS/390 2.10. >The new instructions may (or may not) be less resource >efficient that the older versions. Well, on the z10, 668 of the 894 instructions are implemented entirely in hardware. I would assume that they are very efficient. >IMO, the only time to use a relative instruction is in "baseless" >code. We can disagree here. I use relative branches in all new code unless I need the branch to be indexed. Relative branches avoid Address-Generation-Interlocks. >Again, the only time to use immediate instructions is when the >value is a constant and will fit into a halfword. For my code, I can not use the Extended-Immediate instructions because our customers may be running on processors that do not include those instructions. >Oh, and one other reason to use at least one Branch instruction >in your program instead of a Brance Relative. If the start of a >program starts with the sequence: > > USING *,15 > B AROUND > DC C'some character string' >AROUND DS 0H > >Then the SYSUDUMP will print "some character string" in the >save area trace portion of the dump. In most cases, this >works because GPR15 points to the B instruction. You get the same thing if the first instruction is J AROUND If your program is ATTACHed in AMODE 64, the B AROUND won't work. In that case, register 15 does not contain the entry point address. You are expected to use relative branching. -- Tom Marchant ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
