G wrote:

> this came up in a thread last year.

Yes, it did. The self-appointed experts were spouting
off as usual, so I had to step in.

> HASP uses a word compare, and in a number of tests

I ran those tests (and others that folks sent to me)
and posted the results.

> that came out faster than various alternatives,
> both with S/360 and S/370 instructions.

Nothing has changed with the z196 hardware. If anything,
more instructions execute in the same amount of time
as all the other instruction. Basically, they all take
the same amount of time now (except for the obvious
exceptions: long moves, etc.). This type of performance
enhancement now basically consists of simply reducing
the total _number_ of instructions executed -- and not
optimizing the specific ones that _are_ executed. The
new instructions (in both this machine and immediately
previously) are tremendously useful in this regard.

--
WB

Reply via email to