Re: is Vax addressing sane today

On Sun, 22 Sep 2024 2:10:40 +0000, Lawrence D'Oliveiro wrote:

On Sun, 22 Sep 2024 01:23:35 +0000, MitchAlsup1 wrote:
>

On Sat, 21 Sep 2024 23:29:01 +0000, Lawrence D'Oliveiro wrote:
>

On Sat, 21 Sep 2024 20:39:38 +0000, MitchAlsup1 wrote:
>

In the days before <good> branch prediction having a conditional
branch after each instruction that could have an execution problem was
an extremely poor choice. Thus, exceptions were invented (circa 1958).

>
So all that does is push the conditional branch into the microcode. And
make the instruction more complicated. Why should that be faster?

>
It pushes the branch into the mispredict-recovery path and does not
occupy any code space.
>
There is no microcode outside of Z-system these days.

>
It occupies some space, either microcode or circuit logic, or both.

It has sequencers, but none of them are in ROM or PLA form.

And why should that be faster?

It is faster if for no other reason that it did not fetch the branch
that is always predicted non-taken. ICache and Fetch argument. It is
of lower power because it did not fetch, decode, or execute the branch.
If every calculation instruction had to be followed by a conditional
branch, then the code would be 150% its original size (or worse).