22 Mar
2009
22 Mar
'09
3:15 a.m.
-------------Original Message:
Date: Fri, 20 Mar 2009 16:36:28 -0700 (PDT)
From: Michael Hart <imsaicollector at yahoo.com>
Subject: Re: CompuPro CPU-68000
You know the hiring managers I normally talk to have no knowledge of hardware design.
So even if I am showing off my electrical engineering prowest by implementing a IEEE-696
memory board.
I have the Protel product (www.protel.com), a $4500 piece of software back in 1999/2000
which became Altium (http://www.altium.com/) used to design the most sophisticated
hardware. The manufacturing of the boards I understand have become quite cheap. Anyway
thank for URL.
But remember I am primarily as very senior software developer with lots of hardware know
how. So how would that hardware know how be exploited in getting me contracts?
At least that has not been the interest when people look at me. But still it is a skill I
would one day like to exploite in getting a contract
Michael
------------Snarky Comment:
Cleaning up your spelling and grammar couldn't hurt in getting those
contracts; some people do still care...
another michael
22 Mar
22 Mar
9:18 a.m.
New subject: HP 21mx
Hi
I have a question about micro sequences.
If I had the sequence SLA,INA
would the skip happen before the INA or
after? Would the test for the skip happen
before or after the INA?
Dwight
_________________________________________________________________
Internet Explorer 8 ? Now Available. Faster, safer, easier.
http://clk.atdmt.com/MRT/go/141323790/direct/01/
12:16 p.m.
New subject: HP 21mx
dwight elvey wrote:
Hi
I have a question about micro sequences.
If I had the sequence SLA,INA
would the skip happen before the INA or
after? Would the test for the skip happen
before or after the INA?
Dwight
Micro sequence execution order is discussed in the 2100 manual
(http://oscar.taurus.com/~jeff/2100/index.html)
in the section "3.3 REGISTER REFERENCE INSTRUCTIONS" - Alter-Skip Group (page
3-15/16).
(or for quick ref my own summary:
http://www.cs.ubc.ca/~hilpert/e/HP21xx/pgmref.html#micro)
According to the table there, the SLA is executed before the INA, so that
should answer the question of which value is tested.
I guess you could say it still leaves ambiguous or open to interpretation as to
whether the INA is skipped or not, but I would interpret the skip instructions
description "skip next instruction" as skipping the next instruction word,
rather than aborting the remainder of the micro-sequence.
Alternatives interpretations would be
- skip only the next micro-instruction (would require extra hardware to sort
it all out, for marginal benefit)
- skip remaining micro-instructions as well as the next instruction word
(would contradict "skip next instruction", as multiple
'instructions' were skipped)
Skipping only the next instruction word makes sense in hardware terms as it
requires simply an extra increment of the PC.
So, in short, my understanding would be that the pre-increment value of A is
tested in the skip, but the increment is executed in either case. (I could test
it on my 2116 if necessary.)
1:06 p.m.
New subject: HP 21mx
----------------------------------------
From: hilpert at cs.ubc.ca
dwight elvey wrote:
Hi
This is my interpretation as well. It should test
before going to the next micro op.
Hi
I have a question about micro sequences.
If I had the sequence SLA,INA
would the skip happen before the INA or
after? Would the test for the skip happen
before or after the INA?
Dwight
Micro sequence execution order is discussed in the 2100 manual
(http://oscar.taurus.com/~jeff/2100/index.html)
in the section "3.3 REGISTER REFERENCE INSTRUCTIONS" - Alter-Skip Group (page
3-15/16).
(or for quick ref my own summary:
http://www.cs.ubc.ca/~hilpert/e/HP21xx/pgmref.html#micro)
According to the table there, the SLA is executed before the INA, so that
should answer the question of which value is tested.
I guess you could say it still leaves ambiguous or open to interpretation as to
whether the INA is skipped or not, but I would interpret the skip instructions
description "skip next instruction" as skipping the next instruction word,
rather than aborting the remainder of the micro-sequence.
That is the most interesting part. Does it abort the the micro
sequence and do the skip on the next instruction or does it
finish the micro sequence. I agree that it wouldn't be logical
to consider the micro sequence steps and instructions for the
skip.
Alternatives interpretations would be
- skip only the next micro-instruction (would require extra hardware to sort
it all out, for marginal benefit)
- skip remaining micro-instructions as well as the next instruction word
(would contradict "skip next instruction", as multiple 'instructions'
were skipped)
Skipping only the next instruction word makes sense in hardware terms as it
requires simply an extra increment of the PC.
So, in short, my understanding would be that the pre-increment value of A is
tested in the skip, but the increment is executed in either case. (I could test
it on my 2116 if necessary.)
It would be interesting to know. The manual isn't absolutely
clear on this.
I'm writing an assembler in Forth and was wondering about that.
This last weekend I inventoried my I/O boards and wrenched my
back. I'm just beginning to recover from that.
The assembler is coming along well. I've got all the instructions
coded and all I need to do now is the file I/O and adding
the macros for flow structures.
The instructions are well organized except for the CLE instruction.
One needs to know if it is in a SRG or ASG sequence before
one knows what to do. My assembler is single pass so this
required a little extra fiddling. I assume it is a SRG unless
I see the next instruction as part of a ASG.
I expect to bring my 21MX out of storage when I have some
code to run.
Thanks
Dwight
_________________________________________________________________
Windows Live? SkyDrive: Get 25 GB of free online storage.
http://windowslive.com/online/skydrive?ocid=TXT_TAGLM_WL_skydrive_032009
2:03 p.m.
New subject: HP 21mx
On Sun, Mar 22, 2009 at 1:06 PM, dwight elvey <dkelvey at hotmail.com> wrote:
If I've got this correct, SLA, INA is 002014
On my 2117F I stored 002014 at 000100, set A to 000000 and P to
000100, then pressed INSTR STEP, then A contained 000001 and P
000102.
Then reset P to 000100, then pressed INSTR STEP, then A contained
000002 and P 000101.
So it would appear from this experiment that the SLA does execute
before the INA, and the INA executes regardless of whether or not the
SLA skips.
So, in short, my understanding would be that the pre-increment value of A is
tested in the skip, but the increment is executed in either case. (I could test
it on my 2116 if necessary.)
?It would be interesting to know. The manual isn't absolutely
clear on this.
3:07 p.m.
New subject: HP 21mx
----------------------------------------
From: glen.slick at gmail.com
On Sun, Mar 22, 2009 at 1:06 PM, dwight elvey wrote:
If I've got this correct, SLA, INA is 002014
On my 2117F I stored 002014 at 000100, set A to 000000 and P to
000100, then pressed INSTR STEP, then A contained 000001 and P
000102.
Then reset P to 000100, then pressed INSTR STEP, then A contained
000002 and P 000101.
So it would appear from this experiment that the SLA does execute
before the INA, and the INA executes regardless of whether or not the
SLA skips.
Hi Glen
Thasnks, that is kind of what we expected. So, the
micro sequence does not abort to do the skip, the sequence
continues.
It might be intersting to see what happened when
A started with 177777. I'd expect to end at 101 and not
102.
Dwight
_________________________________________________________________
Internet Explorer 8 ? Now Available. Faster, safer, easier.
http://clk.atdmt.com/MRT/go/141323790/direct/01/
So, in short, my understanding would be that the pre-increment value of A is
tested in the skip, but the increment is executed in either case. (I could test
it on my 2116 if necessary.)
It would be interesting to know. The manual isn't absolutely
clear on this.
3:24 p.m.
New subject: HP 21mx
On Sun, Mar 22, 2009 at 3:07 PM, dwight elvey <dkelvey at hotmail.com> wrote:
Hi Glen
?Thasnks, that is kind of what we expected. So, the
micro sequence does not abort to do the skip, the sequence
continues.
?It might be intersting to see what happened when
A started with 177777. I'd expect to end at 101 and not
102.
Dwight
Yes, starting with A=177777 ends with the same result (except for
Extend also being set) as starting with A=000001, as expected after
executing SLA, INA.
-Glen
11 Apr
11 Apr
10:02 a.m.
New subject: HP 21mx
This is an old one, but it just showed up on cctech.
On Sunday, March 22, 2009 at 9:18, dwight elvey wrote:
If I had the sequence SLA,INA would the skip happen
before the INA or
after? Would the test for the skip happen before or after the INA?
The general rules are:
1. The sequences are executed left to right.
2. The next instruction will be skipped if any skip test is true.
If a skip test was true, P is incremented just before completing the
current instruction. This sets the stage for the instruction skip when the
next instruction is fetched.
So SLA,INA tests if bit 0 of A is 0 and then increments A. If the test was
true, then P is incremented.
An instruction such as RAL,CLE,SLA,ERA rotates A left, then clears the E
register, then tests if bit 0 of A is clear, and then rotates A with E
right. If the test was true, then P is incremented. The effect is to
clear the MSB of A and skip the next instruction if it had been zero at the
start of the instruction.
To be really perverse, the instruction CMA,SEZ,CME,SSA,SLA,INA:
- one's complements A
- tests if E is 0
- complements E
- tests if bit 15 of A is 0
- tests if bit 0 of A is 0
- increments A
So the effect is to two's complement A, complement E, and skip if E was
zero or the MSB or LSB of A was one on entry (because the test on E is done
before its complement, but the tests on A are done after its complement).
There is a caveat when RSS is combined with multiple skip instructions; see
one of the 21MX Operating and Reference Manuals for details.
-- Dave
5739
days inactive
5759
days old
7 comments
5 participants
participants (5)
-
dkelvey＠hotmail.com -
dm561＠torfree.net -
glen.slick＠gmail.com -
hilpert＠cs.ubc.ca -
jdbryan＠acm.org