On Thu, Jul 24, 2014 at 10:22 PM, Eric Smith
<spacewar at gmail.com> wrote:
>The VAX-11/780 introduced the concept of patchable control store,
>where most of the control store was PROM but there was a small amount
>of patch RAM. Costs of high speed RAM declined quickly enough that
>most later non-microprocessor VAXen used entirely RAM control store.
>With the advent of VAX microprocessors, the larger die area required
>for RAM vs ROM shifted the economics shifted back to favor ROM with a
>small RAM patch area.
The 370/145 had entirely writable control store, loaded from
every time the machine was powered on. I'm pretty sure the
and 370/16x also had at least partially writable control
could also be loaded from floppy, for diagnostic purposes,
have firmware emulation of older machines.
Some models of the 360 family had alterable control store
form of patterned cards. The 360/30 had mylar cards that could
actually be punched on a keypunch to create custom microcode.
These cards had one plate of a capacitor plus the dielectric
and by punching out the plate, you made a zero at that location.
The 360/50 and 360/65 had a scheme where there was a word
line board and a bit line board with traces at right angles, and
a solid mylar sheet between them. The bit line board traces had
wide flags on either a 1 line or a zero line, depending on what
the bit at that position was to be. So, the board etch pattern
was changed to alter the firmware.
I think most modern, high performance CPUs do not use microcode,
as the two-stage instruction execution would be slower than
just doing the decode in pure logic. Also, modern CPUs are so
massively pipelined, and most microcode implementations
are pretty much uniprocessor-like, I don't know how you'd handle
all the pipelining with microcode. They may use some microcoded
functions to handle special cases, though.
My main point to respond to was that alterable microcode
the VAX 11/780 by quite a bit. I'm pretty sure Burroughs also
had machines with writable control store earlier.