11 Jul
2014
11 Jul
'14
2:30 a.m.
I've seen 16, 22, and 24 cited as the official number. The VAX architecture
manual seems to suggest 24. If the number is indeed more than 16, how are
they encoded? Seeing as how the manual says 4 bits are used for specifying
the specific mode, and 4 for the register selection, and seems to imply
that the operand specifier is always a byte long.
Sorry for the potentially dumb question.
11 Jul
11 Jul
2:37 a.m.
On Fri, Jul 11, 2014 at 3:30 AM, Kevin Keith <krfkeith at gmail.com> wrote:
I've seen 16, 22, and 24 cited as the official
number. The VAX architecture
manual seems to suggest 24. If the number is indeed more than 16, how are
they encoded? Seeing as how the manual says 4 bits are used for specifying
the specific mode, and 4 for the register selection, and seems to imply
that the operand specifier is always a byte long.
IIRC, it's like the PDP-11, which only really had eight addressing modes,
but some were considered different modes when used with the program counter
vs. a "normal" register.
7:10 a.m.
On Jul 11, 2014, at 5:37 AM, Eric Smith <spacewar at gmail.com> wrote:
On Fri, Jul 11, 2014 at 3:30 AM, Kevin Keith
<krfkeith at gmail.com> wrote:
Both VAX architecture books I have show that the number is 21. There are 13 ?general
register? modes, and 8 more ?program counter? addressing modes.
It?s reasonable to do that, because from the programmer?s point of view, the PC modes have
distinct properties (like literal or absolute mode) even though at the hardware level they
are constructed out of the same machinery as the general register modes. And yes, that
same reasoning applies to the PDP-11.
Interestingly enough, they don?t count another dozen or so modes for all the various
flavors of indexed mode ? where you get a second mode byte after the initial one that has
mode 4 (indexed). That fits, because indexing doesn?t add another set of conceptually
different modes.
paul
I've seen 16, 22, and 24 cited as the
official number. The VAX architecture
manual seems to suggest 24. If the number is indeed more than 16, how are
they encoded? Seeing as how the manual says 4 bits are used for specifying
the specific mode, and 4 for the register selection, and seems to imply
that the operand specifier is always a byte long.
IIRC, it's like the PDP-11, which only really had eight addressing modes,
but some were considered different modes when used with the program counter
vs. a "normal" register.
4:14 p.m.
Thanks for the answers guys!
I was wondering if any of you VAX gurus might be able to clear up another
question that has been confusing me:
In the architecture manual, in the addressing mode summary table (it's
Table 2-2, on page 2-17, or page 66 on the bitsavers PDF) it says that the
literal addressing mode has specifiers 0-3, yet there appears to be only
one literal mode. The chart shows the format as being:
+-----+--------------+
| 0. | (literal) |
+-----+--------------+
My question is, can the operand specifier for literal mode really be 00,
01, or 10? And if so, why did the designers of the VAX waste the extra bit,
and bit and make literals 7-bits?
Please correct me if I'm missing anything obvious!
On Jul 11, 2014 8:14 AM, "Paul Koning" <paulkoning at comcast.net> wrote:
On Jul 11, 2014, at 5:37 AM, Eric Smith <spacewar at gmail.com> wrote:
counter
On Fri, Jul 11, 2014 at 3:30 AM, Kevin Keith
<krfkeith at gmail.com> wrote:
> I've seen 16, 22, and 24 cited as the official number. The VAX
architecture
> manual seems to suggest 24. If the number is
indeed more than 16, how
are
> they encoded? Seeing as how the manual says 4
bits are used for
specifying
the
specific mode, and 4 for the register selection, and seems to imply
that the operand specifier is always a byte long.
IIRC, it's like the PDP-11, which only really had eight addressing modes,
but some were considered different modes when used with the program vs. a "normal" register.
Both VAX architecture books I have show that the number is 21. There are
13 ?general register? modes, and 8 more ?program counter? addressing modes.
It?s reasonable to do that, because from the programmer?s point of view,
the PC modes have distinct properties (like literal or absolute mode) even
though at the hardware level they are constructed out of the same machinery
as the general register modes. And yes, that same reasoning applies to the
PDP-11.
Interestingly enough, they don?t count another dozen or so modes for all
the various flavors of indexed mode ? where you get a second mode byte
after the initial one that has mode 4 (indexed). That fits, because
indexing doesn?t add another set of conceptually different modes.
paul
6:02 p.m.
On Jul 11, 2014, at 7:14 PM, Kevin Keith <krfkeith at gmail.com> wrote:
Thanks for the answers guys!
I was wondering if any of you VAX gurus might be able to clear up another
question that has been confusing me:
In the architecture manual, in the addressing mode summary table (it's
Table 2-2, on page 2-17, or page 66 on the bitsavers PDF) it says that the
literal addressing mode has specifiers 0-3, yet there appears to be only
one literal mode. The chart shows the format as being:
+-----+--------------+
| 0. | (literal) |
+-----+--------------+
My question is, can the operand specifier for literal mode really be 00,
01, or 10? And if so, why did the designers of the VAX waste the extra bit,
and bit and make literals 7-bits?
The picture is a bit confusing. That field of 0 is 2 bits wide, leaving 6 bits for the
literal. The reason the specifier is listed as ?0-3? is that the addressing mode
specifier is the top 4 bits. So what we have here is that if the top 2 bits of the
specifier are 0, you have literal mode, and the literal is in the bottom 6 bits ? the
bottom 2 bits of the mode specifier combined with the 4 bits that are normally the
register number.
paul
6:13 p.m.
OH! Okay, that makes sense! Yeah, that is kind of vague though
So basically, it means:
00|00. . .
00|01. . .
00|10. . .
00|11. . .
Would all be literal mode (where the first two bits of the literal just
happen to be in the area that is otherwise part of the specifier), while
anything other than 0s in the first two bits would specify something else?
Gotcha!
Thanks so much!
On Jul 11, 2014 7:07 PM, "Paul Koning" <paulkoning at comcast.net> wrote:
On Jul 11, 2014, at 7:14 PM, Kevin Keith <krfkeith at gmail.com> wrote:
Thanks for the answers guys!
I was wondering if any of you VAX gurus might be able to clear up another
question that has been confusing me:
In the architecture manual, in the addressing mode summary table (it's
Table 2-2, on page 2-17, or page 66 on the bitsavers PDF) it says that
the
literal addressing mode has specifiers 0-3, yet
there appears to be only
one literal mode. The chart shows the format as being:
+-----+--------------+
| 0. | (literal) |
+-----+--------------+
My question is, can the operand specifier for literal mode really be 00,
01, or 10? And if so, why did the designers of the VAX waste the extra
bit,
and bit and make literals 7-bits?
The picture is a bit confusing. That field of 0 is 2 bits wide, leaving 6
bits for the literal. The reason the specifier is listed as ?0-3? is that
the addressing mode specifier is the top 4 bits. So what we have here is
that if the top 2 bits of the specifier are 0, you have literal mode, and
the literal is in the bottom 6 bits ? the bottom 2 bits of the mode
specifier combined with the 4 bits that are normally the register number.
paul
3600
days inactive
3601
days old
5 comments
3 participants
participants (3)
-
krfkeith＠gmail.com -
paulkoning＠comcast.net -
spacewar＠gmail.com