Re: Assembling span-dependent instructions

On Friday, July 29, 2022 at 1:38:25 PM UTC-7, anti...@math.uni.wroc.pl wrote:


(snip)


> Z architecture (modern versions of IBM 360) has
> such problems too: there are variants of instruction having different
> lengths but even longest variant have limited range of available
> offsets. At least some versions of Z architecture had severe penalty
> for simultaneusly accessing the same cache line for instruction fetch
> and data access, so putting constant pools in separate cache line was
> very important.


I presume this is true for any system with separate data/instuction
cache. It might be more of a problem for z/, with especially long
cache lines.


From the S/360 days, it was usual for data, even variable data, to be
close to code. That is, for non-reentrant programs. Most assembly
code and Fortran did that.


Otherwise, I believe the original question comes up on any machine
with variable sized branch instructions. Many of the stories I
remember are from the PDP-8.


A similar question comes up generating the Table of Contents
with LaTeX. When you run it, it generates the file used to make
the ToC next time. You run it again to generate the ToC, and
it will tell you if anything moved since the time before.
Hopefully it converges.

Return to the comp.compilers page.
Search the comp.compilers archives again.