John,

To me, the only valid count of instructions is that which one arrives at be
recognizing distinct binary operations codes.

Using that rational, SIO/SIOF were distinct instructions (9C00, 9C01),
whereas PC/PCF must be counted as one (1) instruction (B218).

You have correctly pointed out that multiple instructions may be implemented
by a single microcode/millicode routine.  That is really not relevant to the
"number of instructions", since that is hidden from the programmer.

As far as we are concerned, every instruction could well be hardwired.

John P. Baker

-----Original Message-----
From: IBM Mainframe Assembler List [mailto:[email protected]]
On Behalf Of john gilmore
Sent: Tuesday, August 31, 2010 12:33 PM
To: [email protected]
Subject: Re: number of new instructions

The game of trying to count distinct instructions is, I suppose, innocuous;
but doing so serves no good purpose.

The PROP treats MVI, MVIY, and MVC as variants of a generic MOVE
instruction; but it treats MVCLIN and MVCL as distinct instructions.  There
is, I believe, a hardware, instruction-implementation rationale for this
distinction, but there is no functional one.

Or again, as Tony Harminc has just pointed out, the mnemonic pair SIO, SIOF
may be counted as one instruction or two.

Examples of this sort can be multiplied ad nauseam by paging through the
PROP.  A count of distinct instruction mnemonics at time t--They are in an
HLASM table having defined content at time t--can be precise.  A count of
instructions cannot.

Qualitatively, it can be said that there are many more instructions than
there once were; and that is enough.

John Gilmore Ashland, MA 01721-1817 USA

Reply via email to