Keven Hall wrote that he wanted to promote a discussion of

<begin snippet>
. . . why it might be that System/360 was designed with a set of
registers that contain heterogeneous control information, sometimes
within a single register and yet provides
instructions for loading and storing their values in a whole register
at a time fashion.
</end snippet>

Such a discussion in these terms stacks the deck against this design decision.

The crucial issue seems to me to be not heterogeneity, which is
certainly present, but the circumstances in which the contents of
these registers are changed.  Often, for example, an entire current
environment is saved and an antecedent one is restored; and operations
of this kind are best done in 'whole register at a time fashion'.

Moreover, the task of altering  a single bit setting in a register
image that is to be loaded subsequently is a trivial one.

Mr Hall should give us some examples of operations that he judges
awkward or infelicitous done in 'whole register at a time fashion'.

John Gilmore, Ashland, MA 01721 - USA

Reply via email to