700220_An_Instruction_Set_For_The_18 Bit_PDP K 700220 An Instruction Set For The 18 Bit PDP

User Manual: 700220_An_Instruction_Set_For_The_18-bit_PDP-K

Open the PDF directly: View PDF .
Page Count: 33

J).l. s

tr

i.bu

ti

or;.

K~ys:

R\~vigion

~

Obsolete:

Date:

An

Instruction

Set

for

the

i8-bit

PDP-K

Instruction

Set

Op

C()d~

Data,

~.ry·p~~s

None

Ncne

February

20,

1970

~

.•

0 i

';

.

~r:

( " ,J

c:

t 1

:'

, n

An

instruction

set

for

an

IB~bit

computer~

is

proposri].

It

cJmbines

the

best

features

of

the

F~P-llts

archltec-

tnre

.:ind

the

PDP-olD

~

s

instruct~ion

set

For

::~t~"e:ral

.r'ed50ns

f

an

lB-'bi

t

(:omput!;;r

was

considered

s

u:>e

~<.('jr;

i t

solves

both

t.he

Of:)

code

:tnd

addr€-!ss

spac.

...

~

prob~ernB

of:

a

16·-bit

computer.

In

addit.~6.on"it:

i;:: a

better

data

base

in

two

important

aredls.

Pulse

Height

A:lalysis

(PHA)

programs

have

proven

the

need

for

18

bits.

1\.1so,

thE:

36-bit

floating-point

rt;prt~sentation

has

much

wider

acceptance,

due

to

its

superio~ity

of

32-bit

for~

mats·

..

1 r

"e

..

, a

computer

wi

th

a

'word

len1Jth

of

18

bi

t~'

"

2.0

.

L~'

~:~ructi()n

_f~rmat

a!?

..

d

T.e~rminology

'!'hc

ins

tr'uction

format

of

most

~inary

(two

address)

instruct.ions

is

shown

below.

It

resembles

that

of

the

PDP-II

and

has

three

fields

Field

oc:

s:

D:

In

s

fruG+

Ion

~.~ription

Operation

Code·

Wi

.J

D

Specifies

the

binary

instruction.

Source

.

Specifies

the

Effective

AdC:ress

(EA)

,of

the

source.

Destination

Specifies-

the

Effective

Addr...

(EA)

of

the

destination

..

The

formats

of

the

Sand

D

fields

are

identical

and

shown

below.

~

.r\,

0

F'/eIJ

Field

R:

"..--._._"""

L3

..

M

Description

Reqister

I

.,

•

,

3

:::J

,,'

...

R

Denotes

lout

of

8

general

reqisters.

Mode

Specifies

the

addressinq

mode

in

a

simila::-

way

to

those

for

tile

PDP-li.l

lSee

PDP-II

Handbook.

0:

1:

2 :

5 :

6:

7:

R

@R

@(R)+

@J>~CR)

(R)+

-

(R)

@-

(R)

A(R)

:

'R

conta~ns

dat~

~

COl1tains

o.ddress

of

datu

l

Autoincrcment,

defer

Index,

defer

Autoincrement

J

;

Autodecrement

Aut0decrement,

d

fer

;Index·

r:['he

address

as

computed

from

the

Rand

M

f,ields

is

called

the

Effective

Address

'tEA".'

When

M-O,

this

is

from

o

to

7.

The

-tocatio·n

of

the

memory

cel1

2

actually

addressed

is

called

the

Effective

Location

"ELM.

For

most

binary

instructions,

EL-EA,

i.e.,

the

effective

location

=

the

effective

address.

In

some

instructions,

the

S

or

D

field

denotes

an

integer

number;

for

example,

to

specify

the

number

of

shifts

in

a

shift

instruction"

The

format

i4

as

follcws:

'---.

__

-."

l'

EP

,:;

e

fJ

-,---

l"@·'

is

used

as

the

"indirect"

symbol.

2/\

register

is

also

considered

"m~moryu.

Fi.'ld

P:

FR:

DescriJ(tion

Register

Denotes

lout

of

8

general

rcgistars.

:1odcs I

Short

Specifies

the

addressing

rnodco

They

are

identical

to

the

first

4 modes

of

the

M

field:

o : R

Epl

is

R

1:

.

@R

;

EP

is

(R)2

2 :

~

(R)+

EP

is

@

(R)

f

autoincrement

3:

@A{R)

EP

is

(R)+A

Free

Bit

Bit

not

used

to

determine

EP.

The

Effective

Position

"EP"

can

be

interpreted

as

the

number

representing

the

EA

when M

would

be

restricted

to

the

first

4

combinations.

The

table

below

shows

the"

values

EP

can

have:

Values

of

EP

MS

Signed

Integer

Unsigned

Integer

0:

R

-4

to

0;

0

to

+3

0

to

+7

1:

@R

-2

17

to

0;

0

to

217_1

0

to

2

18

2 :

.@

(R)

+

...

2

17

to

0;

0

to

217_1

0

to

2

18

. .

@A(R~

-2

17

to

0;

0

to

217_1

0

to

2

18

-.

lttEpn

..::

Effective

Position

3.0

('1,Hupatibili·ty_

Introducing

a

different

word

length

will

cause

some

compatibility

problems.

3.1

Peripheral

COffipatibilit:i

A

separate

memorandum

will

be

devoted

to

this

problem.

The

incompatibility

can

be

reduced

by

r.aving

the

same

bus

st~ucture

for

the

PDP-K

as

the

PDP-Il.

This

is

be-

ing

consider~dc

3.2

P~ogram

Compatibility

Two

aspects

have

to

be

considered.

3.2*1

Word

Length

Compatibilit:t

This

can

be

done

,by

hardware

by

having

a

16-

and

an

IS-bit

mode;

by

software;!

through

a

conversion

pr)grarn

simi

lar

to

that

for

converting

PDP-8

to

~DP-9/l5

?rograms

leaving

certain

portions

to

be

recoded

"by

hand"

'e.g.,

shift

and

rotate

instruct-ions).

3.2.2

Instruction

Set

compatibilitl

This

can

be

accomplished

through

microproJramming.

,

Because

of

the

PDP-Res

lS-bit

word

lenqth,

microprogram-

ming

becomes

very

attractive

,because

the

lOP-IO

can

be

emulated.

-(j-

4.

a

Propos~~d

I?DP-K

Instructio,!l

Sc

t

The

proposed

instruction

~et

is

shown

in

Appendix

A.

Only

tte

major

instructions

ar~

shown.

These

are

the

essential

ones

or

thos~

requiring

lots

of

op

code

'space.

It

is

.;\ssur"".cd

that

the

reader

has

some

knowledge

of

the

PDP-II

instruction

set.

The

instructions

operate

on

5

data

types.

4.1

Bit,

"En~

A

bit

is

a

Boolean

quantity

which

is

true

itT"

or

false

'F".

4 • 2

Byte,·

.::1.:

A

byte

is

a

character

..

4.

3

Word,

.

..:!!.:.

A wore

is:

I

.....

2.

3.

A

Boolean

Array

with

18

elements

A

signed

integer

(2's

comple~ent)

A~

unsigned

integer

4.

4

~Jle

Word,

"0"

A

double

word

is

a

single

precision,

floating-point

num.)er.

4.5

Qu<,druple

Word,

"0"

A

'~uadruple

word

"is

a

double-precision,

floating-point

number.

lDenotes

abbreviat~.6n

for

the

particular

data

type.

By

tC$

are

handled

in

a

\vay

simi

lar

t.o

th,(~

PDP

-1

0 ,

as

described

in

Appendix

B. Few

instructions

operd

tf~

on

byte

because

bytes

~re

considered

a

riata

~ormat

for

characters

only.

Most

instructions

operate

on

words

as

the

word

is

con-

sidered

the

data

format

for

program

control

ani

integer

numbers.

It

is

felt

that

higher

level

languages

(FORTRAN, ALGOL,

etc.)

use

integers

T.rtostly

for

subscript-

ing

and

program

control

an~,

therefo:e,

a

singlt.

IS-bi

t

integer

is

considered

sufficient.

The

condition·

code

"ce"

is

handled

in

a

way

as

described

in

Appendix

c.

-

-8-

5.0

!?Et.scr!I?tion

of

InstructiQns'

Appendix

D

d~scribes

the

instruction

formats

and

the

interpretation

of

the

fields

'of

the

format.

The

data

ty.pe

of

the

instruction

will

be

indicated

by

a

letter

following.

the

mnemonic

of

the

instruction

..

The

letters

are,

as

defined

before:

B~;

bi

t,

Y

:;:

byte

I

W =

word

or

no

letter

(default)

I 0 =

double

word

and

Q -

quadrupl.e

word.

Henc'e,

MOV

can

be

designated

by

MOVY

I

MOW

or

MaV,

MOVD

and

MOVO

•

.

The

,operation.

of

the

individual

instl'uct.ions

is

qiven

below.

MNEM

Operation

• H·Ule

Format

,"

------

..

-------~------------------

-----------------

MOV

.CQM

.

(8)+0

(

(S),~)2+Cl

(tS)

.0)

'+«:2

({S»O).t.c3,

(D)

~'(S)

.'

(r

<0)

3-+<:1

(r=O)

-+<:2

(r>O)+C3

~carry-O)

"C

(Overflow-l)·Y

Move

ISOl

-Compare

.so

(D)

-

(S)

,

(8+1.)4_(1»'

Com.pare

'SD

«D)

«&»

+Cl..

,. . wi.th.

Limits

«(D)~(S»'

«S+n»)(D»)~2

«O)-(S+n»..c3

(D)+(S)+D

(r

<o)-+el

(r==O)-+C2

(r>O)+C3·

,:(

Carr:y-l

)

-+c

~

(OVerfl()W=ll)

~V

'SD

"~FO~

in.t~ction

format

see

Appendix

0 *

:,t:.

'.

~,:~.=:".(

(5)

<Ol-+cl"

means:'

if

(5)

<fo()

~en

l+Cl

j

else

O-+Cl,.

, 3

..

!-,_.

~'esult

of

operation.

4·S+

n

-=

next

data

loca~ion

from

the

source,

....

9-

~NEM

_J?1?=.E.~

......

·.~_o_n

__

.

__

..

__

"

____

e_

..

__

...

IJ_!1_me

____

F_O_r_.

m_a_t.

____

_

SUB

MUL~""'"

DIV

(D)

-

C.sl

.....

D

For

CC

see

COM

10)

*-csj'~D.t:o-~l

(r

<Or~-ct'

(:r=O)

'~C2

(r>O)+C3

( I r

i?

211)

:2

..

V

(D)

,

(D+~j)

/

(S)

-~

D ,

D+

1

(q

--0)

3-+Cl

(q-O)...,C2

(q>

0)

-+C3

(Jql~217)

.. v

IMUlr

(D). *

($.)

-

•.

D

I.

~>

For

CC

flee

MDl",

ID'IV

(B.)

I

(~

...

~),

q

tf:/

"_"

Por

CC·:

....

:,;·DIV

.

Subtract

'SD

Multiply

ISD

Divide

'SD

.

Integer

f.SD

Multiply

Integ-er

Divide

ISD

'".

.,

J

.,:-,.,."<,~.

'.,

''''.

".

•

BXCH

(S)+t"':';]J»~S,

(temp.)·,.i)

Exchange

ISD/

.

.'

,~'.':

.~;.!.

~

~

.

[fa)

'<O)~:l::

(

(I)

-0

).ea.

«(S»O)+el

" .

COML

(D)'

(8)

,«p>'

(S)

).

(S)~cc>

¢ospare

'SD

Loqical

AND

«D)

&

(5)-0)

.

«((D)&(S».(S)+O)

(

(D)"

(S)

.0)

'+CC

«(D)'(S».(S)~O)

(D)

,

{S)

·

...

D

(r<O)+Cl

(rcO)

+C2·

(7:>0)

+C3

2Jrl

at

absolute

value

or:

r.

lq

lIlI\

quotient.

of

divi.aion~

All

,1

11

••

in

(S)

are

0

in

7

(D)

SON

-l

's

in

( S )

are

0

in

-( D )

Some

l's

in

(S)

are

1.

:in

(D)

All

lis

in

(S)

are

1

in

(D)

.'~>9i:eal

AND

'SD

10-

ANDes

(0)&(5)

$.0

l'''or

CC s

((~e

AND

~

·IOR

For

CC

see

AND

lORes

(D)!

(S)

v

-.0

XOP.

{D)8(S)+D

For

CC

see'AND

XORCS

(D),

(5)

'.+D

FADD

(I)} +

(5)

-~D

(x-

<0)

~'::l

(r-O)

-+C2

(r>O)

~~3

(Overflow-i)

,..,..v

ana

'trap

(Underflow-l)

..

U

~~40.

F'SUB

(D)

<-

(S)"D

.For

CC

.ee

FADD

F~'10L

( D ) ., (S»

-t;·,D

POI"

CC

see

FJ\DD

F'D!V

)/{S)~D

CC

..

'

see

«PAnD'

LOC,li

cal

I

SO

~.ND

wit.h

Comple-

m.:-:nted

Source

Logical

.SO

Inclusive

OR

LO'g i c

Z/.l#

S D

Inclusive

OR

with

cal

'SD

Bxcl

ua

l. ·"e

OR

Ex<:,l'Jsive

th

F

n9'~

AOti

Floating'

ISD

Subtract

?l·oatlng

tso

Mult.iply

~"1.oat.in9

t

Dl~lide

-1:-

ftU~Elw1

Operation

Name

:F'ol~mat

-.---..-.

....

~~-

__

.-..-~~""_

.....

_,

__

-.-.t~~

___

.''''''''''JW~;~~

__

Q~1f.9

_______

~

__

''-~~,,

__

~._

AOS

(D)

+l~'D

Add

One

#CSKI'

and

Skip

1.

f

(CC=T),

then

(PC)

+R·~PC

When

skip

condition

is

satisfi.ed#

the

PC

wi

th

th(~

va.lue

in

the

R

field

(0

t;.., 7

if

(CC=T),

then-

(PC)

+R

..

PC

if

(CC=T),

then

(PC)+R+PC

~

i.£

SOJ

if

(CC:.:T),

then

(D)

-+PC

1f

STJ

(R)+J\

if

(CC-T)~

then

(D)·.PC

Slmtract

tCSl(P

One

and

Skip

Test

and

fCSKP

Skip

Add

One

'C,JblP

and

Jump

subtract

fCJMP

One

and

Jump

Test

and

tCJMP

(;Jutnp

.

LSH

Shift

. (D)

+0

Loqical

fEPD f

iLSH

Shift

The

shif't

direction·

and

the

nu.raber

of

shifts

depend

on

the

siqn

and

absolute

value

of

the

number

determined

by

~

the

EP

in

the

S

field

of

the

instruction.

(r

<0)

· ..

·Cl

.

(r-O)

+C2

(r.>O)

-+C3

(last

bit

shifted

out)+C

(OvE.~rflow=l)

l .. v

LSHC

Shif

t

Combined

(O)

i!

(D-+

1)

·,.D

,D+1

For

explanation

and

CC

see

LSH'

Logical

fEPD, 'LSHC

Shift

Combined.

lOverflow

c,'ccurs

(on

le"ft

shifts

and

rotat,es

onl~{)

wheneveI;':

the

value

of

the

two

most

significant

bit.s

of

(D)

become

unequal..

Once

V

is

set,

it

stays

set.~

On a

rlqh.t

shift

()l'

rotate,

V

is

cleared

..

M}iEt"'l.

Operatior:..

Nan1€~

F;'o:emrtt

-~"'~.>eto_"''''''~~''''-~''''Ar~~_~~_'l'''''·',)Y''>l!I...,.,..,_-.._.~_~

...

.w;..'~·~'_'':'l.>"""~'_"_'''''W'''''''''~'_·'''~''''·-''_.'.<II\_''''':-_'

__

~'>#_'''''_':l;'''_,

__

.

__

Rotate

{D)~D

Rotate

fEPO,

tROT

The

rotate

direction

r

bit

ons

'rotated

depend

on

the

sign

and

absolute

the.

number

{l,;:termined

by

t:he

E;P

t.h.t; S

str'Uction

..

(r<O)-+Cl

(r=O)+C2

(r>O)-+C3

(last

bit

rotated

out)~C

(Overflow=l)l-1V

ROTC

Hota.te

Camhi

(n~+"l)

t:e

ASHe

-~

D

,D+

1

Combint:~d

FCir

explanation

and

CC;f

see

Rt)T

..

Shift

Arithmetically

(,0)'''4D

For

explanatit)n

and

CC

I'

see

SH

..

Shift

A.t'itJunetici"lly

C<)m-'

bined

(D),

(r,+

1)

-+t~

i

D+l

J."or

explana.t.ion

a.n.d.

CC

i'

see

LSH.

tnllle;

t.!

c

Shi.ft

IA.:::

i

t.tlmE:

ti

C'

Shift

Com

..

.,

bine.d

BIS

1+EBL2.

Bit

Set

tEPD

The

EBL

is

determined

lUS

follt)w8;'

t.he

EA

of

the

instructic)ft.

is

t.aken,

starti.'!l9

from

the

of

the

wbrd

-denoted

by' EA,

EP

bit.

locations

Note:

EP

allowed

,to

be

b,:tqqer

thAn

18"

SIeL

~,

.

)

..:~,

)

~l>Cl

(BaL)')!

(C)

....

C2

(EBL)

3f;

(C)..c3

"

(EBL)

J

~C

j{or

explarLation

andCC"

see

SIS<\>

Bit

Clear

tEPD

Overflow

occ\..u:~

t

..

he

value

of

left

shifts

and

rotates

only)

whenever

two

most

significant

bi

ts

of

(D) hecolr.e

llnequal

is

set

it

stays

set.

On

it

.tight

oX:'

.

t'otat:e

t

'l

ared

..

'"

bi

t.

loc<ttion,

#I

meant

the

1

--

l ,.,-'

aICH (EBL)

'.EBL

c"or

expla:nation

and

CC

~

see

BIS~

arcp

{C)

....

EBL

(EBL)

1&

(C).Cl

CE.BL)

11

(C)-t-C2

(EeL)

1.

(C)

..,.C3

.

BIT

(C)..;..EBL

(EI:lI.)

....

EBL

For

~xplan.aticn

lUld

CC

I

see

SIS"

BITe'

(EElI)

(EBIt)

~,

)

-+C

1 .

(EBL)

~1(C)+C2

(EB,L)

'.

(C)

~C3

(ESl;)

'

...

c

,Bit

COIn-

tEPD

p1.ement:

B1 t.

Copy

tE:PD

Bi

t T&:!st tF';PD

compleme:nt

BIMS

<EBL)+{-SP

Bit

Move 'EPD

to

Sta(~l{

l'h,e (f,fn.,)

is

pushed

on

t.he

stack

as

if

it

were

an

18--

c-

bi

tWCJ:d,

BIJVt1

(:,P)

+

..

F;SL

Bi

t

Move

iEPD

.

to

Memory

If

(SP)·

t,

then

O+EBL

e~lse

l

....

EBL~

SMOV

(D)+-{SP)

Stack

tEPD

Move

This

isa

ft'nve

from

memt

,ry

to

the

stack

(R6 i.a

impli

staekpo~ntt,l:)

.,

EP

is

'interpreted

.,S

~

~

t.n:!;.~,and

the

FR

fl.eld

:ls

interpr~·~ted

as

a 2£.!t

::!E....'!:~!';;.~!

.)l,t"-

El~~'~-

if

FRa;,a

\hen

EA+EP e

18e

(EA+EP)

MMOV

(Sp)

+

....

n

Memory

tEPH

Move

This

is

a

JrD;"efl,"~m

sta

;ck

to

memory,. .

For

furt.t:er

tion

and

CC,

see

,\~,tOv

..

BR

if

(CC=T)-

t;iu~.,

(Pt::)

+ COFPS)

Branch

IBR

+PC

lana~·

When

the

b.ranch

cond:.l"

~on

is

satisfied',

the

off!:et

;~

a

9-bi

t

signed

qt~,antl.ty)

is

added

to

t.he

PC

..

..

In

her..e

it

is

,meant

the

(EBL)

prior

t'o

change

..

NNEf.1

Operation

"

Name

_,,_-

_~

___

".,.,_.

___

...

_"_

........

,

....

~""""""'*.,..

__

....

_~~_.

____

........

,

.....

h_,

Format

JSR,

Jump

to

tBPD

JSP

Subroutine

Specia.l

suLroutine

call,

passes

parameters

to

the

stack

automatically.

~ee·

Appendix

E.

ANAL

To

be

defined

later.

Analyse

#SD

REPS

Repeat

:#

REP

Sinqle

The

EP

is

interpreted

as

an

unsigned

integer

repre-

senting

the

repeat.

count

·'Re"..

The

repeat

action

is

stopped

when

(Re-O)!

(CC=T)..

'When REPS

stops

and

(CC=T)'

(RC~O),

then

the

remaind.er

of

the

repeat

c:ount:

is

pushed

on

the

st,ack,

i

..

e

..

I

RCre-"m

-1>

.....

(8P)

..

'

:

BPD

Repeat

tREP

Dcnlble

,

Repeat

next

two

instructiOlu;.

f~()r

explanation,

see

UPS.

'

JMP

if

(Cc-T)

the

(D)

"PC

Jump

'JMP

'J~

takes

place

when

jump

condition

,is

satisfied

..

~

XCT

if

(CC-Tf,<,then

Execute

Execut.e

fJMP

:When

condition

satisfied,

the

instruction

denoted

by (

~~!:ta·.

executed.

"'.'.'

-''\!''\'.,

xc:ru.

if

(CC~')

the

Execute

Execute

JJMP

Undisturbed

'

Undis-

,

turbe4,

.

·1Iben:condi

tfon

aati.fied,

the;iinst~uctt.on

denoted

by

1~~

is

executed

undisturbed,

l~.~i·"',~tre.ult

of

the

"Qpe:ration

i8

not

st.ore3

only

~.,

,CC'

is

aet.

-

-15-

6"

0

Registe£._~eV~!!

General'register

"R7"

is

used

in

the

PDP-li

as

the

PC

(program

Gounter)*

Because

of

this,

certain

addressing

modes

are

no~

advisable

or

lead

to

"self-

destruction"

of

the

program.

'The

table

below

shows,

this.

ADDRESSING

MODES

FOR'

R7

WI,.,

r

............

.... .

...........

------------~-.---,.

Source

Destination

----~~~-~-~.-.---------------------

R7 OK

@R7

OK

,eR7)+

OK

fA

(It?)'

OK

(R7)+

OK.

--

(It"

Error

@-

(lti)

.Error

At.')

" -Oft

-J a

.....

..

r"<.''-·

Of'.

~

R1

@R7

,,:a7) +

'A(R7)

eR7)+

-(R1)

@-'(R7)

A (1t7)

J lilt ;

OK

Error

OK

OK

Error

Error

OJ(

~tt.

suq

...

tedllOt"cmlY" to.p:.;."eJrt

~,fJle

proqrattmter

!J

r

9l'

,

-&king

these

,~~ror.,

ba~,.l.oto,turn

these

'~~lty

com.biJl-atione

into

.ome~in9

useful.

'.'';'. :...ff

;.Use

the

destination

mode

.("')+

tn.'

nOhlal

way

exc~pt

i40

not

store

the

result

of

the

operation.

This

way

~l-ornary

instructions

.become

-te.t

iaaediate"

in-

'structions

~

6

..

2

Use

the

dest.inat.ion

modes

-(R7)

and

@-(R7)

as

flags

indicating

the

follO\riin9~

Consider

the

instruction

a

st.ack

operation

with

the

stack

(i

..

e,. f

there

where

R6

o(;)in'ts

to)

as

the

d~;stination

and

as

source

the

contents

of

"(R5)

+EN"

•

The

E:.:

fecti

ve

,

Number

"EN"

is

the

contents

of

the

S

field

of

tI\€

Instruc

....

Elan

-rnterpreted

as

an

unsigned

integex

(i.

e,,;

fr'om

o

to

63.)

...

The

binary

instructions

look

like:

...

(SP)

Operation

«R5)+EN).-+

(SP)

6.2.2

@-(R?)

C~.~~

Operation

'similar

to

the

-(R7)

cas(~

excent

ILS

source

the

contents

of

«(RS)

+EN)

is

taken..

Bin~iry

instructions

look

like:

(SP)

Operation@«R5}+EN)-+(SP)

lOW

c

double

20W

..

quadruple

word

':ot

}

-S+n

=

next

data

word

~'c~~-·_.ct

_._.~c;.

___

...

____

•

___

•.

___

._____

.

.--

--

..

---

....

--.-------

...

---

c

--

-

..

--

.-.

-

.-----.-

---.--.-.>

...

--...-.

-----~--.~-

-~

....

---.-

c~·-

,SG\UD~~.a.atJ;>uction

J;tescription

Bit

Byte

\'Jord

,irl~

\

....

.

~

__

~'L

1

AOJ

(R)

+l-;.R,

jump?

Add One

and

Jump

.;

1

SOJ

(R)-l+R,

jump?

Subtract

One

and

Jump

.:

1 TSTJ

(R)-..R,

jump?

Te.t

and

Jump

"

1/2

LSD

Logical

Shitt

,:

1/2

LSHC

Logical

Shift

Combined "

1/2

ROT

Rotate

.

.;

1/2

RCT£

Rotate

Combined

I

1/2

ASH

Arithmetic

Shift

I

...

1/2

ASHe

Arithmetic

Shift

Combined

I

1/2

BIMS

(BBL)

....

-(SP)

4

Bit

Move

to

S'tack

I

1/2

BIMM

(SP)++EBL

Bit

Move

to

Memory I

'I

1/2

BIS 1

...

£81,1

B.lt

Set

"

1/2

BIeL

O+IBL

Bit

Cleat;

I

1/2

BIeN (EBL)

'+EBL

Bit

Complement ,

1/2

step

(C)2-+EBL

Bit

Copy "

1/2

BIT

(BBL)

+ce

l

Bit

Test

.;

1/2

tlI'l'C

(BBL)

'-+CC

Bit

Test

Complement

I

3

SHOY

.

(0)

+-

(SP)

4

Stack.Move,

Multiple

Indexed

.;

.;

./

3

MMOV

(SP)++D

Memory

Move,

.

Multiple

Indexed

.; .;

,I ,

~

2

Bit

Branch

OJ

I

Z

JSR,

JSP

Subroutine

Call

1 ANAL '

Analyze

1/8

REPS

Repeat

Single

Cond,

N

l/8

REPD

Repeat

Double Cond, N

1/4

JMP

Jump Cond, D

1/4

XCT

Execute

Cond, D

1/4

XCTU

Execute

Undiaturbed

Cond,

D

4/64

TST

(D)+CC

Teat

.;

.;

.;

.,""'!>:;1I!'C'<f~~

.""

lEBL

~

effective

bit

location

2(C)

:::::;'

content.s

of

carry,

status

bit·

,

'l

,~CC

-~.

condition

code

4sp

=

stack

pointer

/;';;'~:~:"~.'::':~";C;:;:L>;:.;._...;;.~._

."

........

__

._.

_____

-----_.

'--_'

._--_

.

.-

-:

Count

;truction

.

Description

C

Bit

Byte

Word

~C

Q\.'

-

-"

-'

.,

....

,.,

.t!!k.!.t

..•....

,,,.

.~.

$",

'---'

---.--

..

...

I:~~

..

_,

~

4/64

SETZ

O+D

Set

to

all

Zeros

I I l I

t'

1/64

SETPO l

....

D

Set

to

plus

One

I

1/64

SETMO

-l+D

set

to

Minus one

1/64

ADDe

(D)+(C)

....

D Add

Carry

.I

1/64

suac

(D)-(C)+D

Subtract

Carry

i

1/64

roc

(D)

t .... D Take

One's

Complement

.;

3/64

'rrc

(D)

'+l+D "Take

Two'.

Complement I v .

1/64

CIFS

(0)+-(8P)

convert

InteCjer

to

Float.

Single

.;

1/64

CIFD

(D)+-(SP)

Convert

Intec}er

to

Ploat.

Double

/

1/64

.

Df:SI

(D)+~(SP)

Convert

Float.

Single

to

Integer

.;

1/64

CPSD

(D)+-(SP)

Convert

Float.

Sin91e

to

Float.

D. I

t

1/64

..

CFOI

(D)+-(SP)

Convert

Float.

D.

to

Integer

I

'1/64

C!o'DS

(D)"-(SP)

Convert

Float.

D •

to

Float.

s.

I

. .

1/64

!NCBP

.Inore.nt

Byte

Pointer

I

1/64

DEtSP

Dec:r_nt

Byte

Pointer

I

1/256

MCCS

(CC)+-(SP)

Move

CC

tb

Stack

1/256

MCCC

(ee)

+C

Move

CC

to

C

Bit

I

~

1/4096

MSCC

.

(S,)+

...

C

Move

Stack

to

C

Bit·

,.!)

I

(D)~(D+n)

"'.

. I I

3/64

NEeH

Next

••

change

1/64

LOCK

(

(D)

-0

)::a.( 81')

++D

Lock

J~

APPENDIX

B

PDP-K

Byte

Handlin~

The PDP-K

will

handle

bytes

in

the

same

manner

as

the

PDP-lO.

The

format

of

the

byte

instructions

will

be

similar

to

all

other

instructions.

I

nst

roc

t.on

..--~

[ b b 6 ]

."---r-J .I..-....,-J J

•

OC.

S "

'rhe

possible

oct.

areiMOVY,

COMY,'

and

COMLY.

The

Sand

locations

a.nd DYP".

as

the

EP

D

fieids

are'identical

in

format

and

define

the.;,

of

the

source

and.

destination

byte

pointers

"sn

The

Sand

Dfields

art

interpreted

the

same way

field,

described

in.¥ction

2.0

and

a8

shown

belOfl.·

if,'

~.

,.

1>1

.....,..,

FA

,

s

",

2

•

IfS

~

0

..a

........

~

F,e;Id

; .

.f

[ \ 3

t.

",

II

~~

________

~*~

______ J

.

JP

I

The

locations

of

"the

SYP

(source

byte

pointer)

and

the

DD

(d

..

t.laation

byte

pointer)

.~e

4eter.mine4)~

tbecontents

of

the

EP'

s

of

th.

•

and

D

fiel~

of

the

i

••

tructions.

The

fzwe

bi

t.s

"Fa-

are

..

ad·

to

alloW

for

iacrementinq

the

byte

poinur.

'the

formats

of

til.

sn

ana

ow'

.....

id

••

ti~l

fMl4

shown

bel~'.

jYP.,

Oyp

.,.

~.....

'

..

r

Field

yp

YS

YL

-2~.-

Description

;t'he

position

of

the

first

bit

of

.the

byte

in

the

double

word

addressed

by

YL.

The

length

of

the

byte

in

bits.

YL

is

interpreted

as

a

reqular

destination

ani

de-

notes

the

location

of

the

double

word

contain:ng

the

byte.

-22-

APPENDIX

C

Condition

Codes

The

PDP-K

condition

code

differs

from

the

PDP-ll

because

of

the

specitl

require~nts

imposed

by

the

single

bit

diddling

ins,tructions

of

POP-It.

The.

instructions

making

use

of

th~

condition

code

have

4

bits

to

specify

the

,condition.

The

function

of

4

of

the

condition

code

flip-flops

will

be

disct.ssed

below.

Condi·h'on

c.~

FI;

p-

Flops

ell

C2:

.

,.-

~

I

(1

1·

C2

I

C3

J C ]

£

,

indicate.

indicate.

indicates

"indicates

"til!"

-6-

-.-

"I"

J ?

•

in

arithmetic:

in

.~ft,le

bit

operations

operatioas

~

arl~tlc

operations

in

sinlle

bit

operations

e31'

indicates

..

,.

....

in

.r1~tic

operations

indicates

•••

in

8ingle

bit

operation.

Ci

.

carq

bit

also

"sed

.~

.t.eat

bit

iA

single

bit

operation.

In

hlthMt:ic

operations

the

flip-flops

el,

e2,

C3

and

C

are

used. a8

listed

in

the

table

below

and

interpretttd

as

follows.

Cl-1

'When

the

renlt

i.

4();

.C2-1 when

result

-0#

C3-1

when

r

••

ult.,i1!':O,

and

C.l

when

there

i.

a

carry

or

when

there

is

no

borrow.

Xatbeca.e

of

bit

diddling,

.the

flip-flops

are

used

&8

follOWS:

!:;.'

(BBL)

2,

(C)

J+Cl

(EBL) I (C)

+c2

(EBL).(C)..-C3

(EBL)"-C .

. ,

'''Ii

---------

.

ls.e

Appendix

'A

iDatructio68

alS,

BIeL,

BleD,

BIT,

BltfC

and

BICP.

't

2EaL

-

cont.ents·

of

EffectiVe

Bit

Locatioa

,complemente.c! wheD

the

8TC

(bit

test

ca.plement)

instruction

ia

used.

3(C)

,=

contents

of

the

carry

flip-flop.

-23-

The

operation

above

allows

all

16

boolean

operators

between

2

variables

directly

and

allows

domplex

boolean

equations

to

be

evaluated

easily.

The

interpretation

of

the

contents

of"

the

flip-flops

el,

c2

and

C3

for

signed

arithmetic

and

bit

didlinq

is

shown

below

and

required

8

"condition

code

combinations"

out

of

the

16

total.

TABLE

Cl

, 1 •

Siqned

Bit

< = >

Arithmetic

Diddling

CI

C2

C3

Interpretation

Interpretation

.

O·

0 0 0

PaIse

BHOT

1 0 0 1 > BG'l' •

BXOR

.

2 0 1 0 •

BEO

.,

BIOR

3 0 1 1

~

BGE'

"

SHAND

4 1 0 0 <

BLT

, BAND

5 1 0 1 -

BNB

I •

BHIOR

.

6 . 1 I 0

.s

BLE

.'

BNXOR

'.

7 1 1 1

True

BRA

True

BRA

The

remaining

8

combinations

are

used

as

shown

in

the

table

below.

Toqether

with

the

\BEO

and

BHE

conditions

fr~m

above

they

contain

all

conditions

for

unsigned

arithmetic.

;

-24-

TABLE

C2

..

--.---

--

---~-

-.~~

..

-.

._-----_

...

_

.......

-

Speciz..l

Unsigned

Condition

Arithmetic

Interpretation

Interpretation

r--'-

1

_J

0

Repeat

count

:::I:

0

BZR

1 >

BHI

2

OVerflow

BOV

3

No

Carry

BSCA

~

BHIE

4 I

Carry

BCA

<

BLO

5

No

Overflow

BNOV

6 .

~

BLOE

•

1

Repeat

count

~

0 I BNZR J

-25-

APPENDIX

D

Instruction

Formats

..

0.1

~ormat

lSD,

Source

Destination

Instruction

has

3

fields

of

6

bits

instruction

1\

"

I 6 6 I 6 I

OC

a

operatior~

code

"

""

A.

I 5 -

source

'v

v v D -

destination

OC

S 0

The

S

and

D

fields

have

the

SUle

forsu.t

as

ahown

below.

S

or

D

field

,

I\,

"

I 3 I 3

I.

R

a'"

reqi8ter

j:ield

V M • mode

fi~la

.V

M R

0.2

For.a.~

'CSKP,

Conditional

Skip

Instruction

has

4

fields.

'!'he

SC

field

(skip

,

.

condition)

i8

interpreted

as

in

Table

Cl

of

Appendix

c.

The R

field

contains

the

number

of

words

to

be

skipped

(from

0

to

7).

] UJ • r

6 6

v

\.

v

OC

sc

• D

-

26~-

D.3

Porm~t

#CJMP,

Conditional

Jumc

_.-'_~I'f

__

...

_____

~",.'4

________

'

____

"_'

••

"-,,,.......L.-,

D.5

'1'

i

i.

i

sin

s t r

11

C'

t.

ion

has

4

fie

1.

d 5 •

'1"

h e J C

fie

1 d

con

t a

ins

t'1f::

jump

condition,

interprete,d

as

shown

in

Table

Cl

of

Appendix

c~

The

R

field

denotes

the

register

to

be

tested

after

an

increment

(decrement

or

test).

oc

JC

R D

[

6"JLGJ

-_.---v".- - .

...1

I

~~

"----v--

I

OC

'L

E~

D

JC

= jump

condition

EP

=

effective

positior

field

FR •

free

bit

used

to

extend

the

OC

field

D

is

a

regular

destination

field,

EP

is

a

regular

ef-

fective

position

field.

Format

'LSH,

L02ical

Sh.i.t~

(D)

,---

A

------""

Q~[hi9h

II

:

lOW:

""1--

-Q

18

D06

Format

tLSHC

A

Logical

Shi~~

~ombine~

(D)

,..-

A.

f-:l--

r

! 0

.~igh

L_-1

18

low

D.

7

Format

*

¥<?T,

I

Rotate

l---'--

(D)

I

LF---------;~-

r~~h

________

~

___

19-

_____ ._.o-l_o_w_

I

I

L-..

_________

..

__

.

____

.

______

,---.J

0.8

Format

'ROTC,

Rotate

Combin~d

,-------.-----

--------------

----

j

(0)

(D+l)

"

--

1\

18

J~I

high

l

,--;

low

14--1

0

'---

___

~

___

~

_____

,J

l-.~

___

~

(

D+1)

D.9

Format

'ASH,

Ax!~met:ic

Shift

.

(D)

r-------------~-------------~----,

[ 5 I 4 -I

'-

" "--v--"'-

OC

BC

9 I I '

v

.)

!'\F"~At-

-"_.

-------·1

r:~-'

I

!

highJ

J

~

bit

r--

Ut:tJJ

ac -

branch

cOlldi

tion

Offset

=

9-bit

signed

integer

i

f";

:c

I

0.12

Format

'REP,

Repeat

r ----

,

8 4 1 I 5

I •

.

'=

A-~

,

'W"

oc

COHD

EP

-

single/double

bit

D.13

Pormat

fJMP

tJump

oc

COHO

D

APPENDIX

E

Subroutine

Calls

Besides

the

standard

PDP-II

JSR,

the

PDP-K

will

have

a

ncre

powerful

subroutine

call.

'rhis

new

call

"JSP"

(Jump

to

Subroutine

with

Parameters)

automatically

passes

paramet:ers

to

t.he

stack

a.nd

does

"stack

house-keeping"

in

such

a way

that

subroutine

returns

can

be

done

ift

a

trivial

way

while

the

stack

is

ucleaned

up"

automatically

..

The

format

of

the

call

is

iEPD

where

the

EP

field

is

in-

terpreted

as

the

number

of

parameters

to

be

pushed

on

th,!

stack.

Register

R5

is

used

to

point

to

the

first

passed

parameter

after

it

has

been

pushed

on

the

stack..

The

ex,lmple

below

~hows

how

th.e

JSP

could

be

implemented..

Note

that

in

addi

ti0n

to

t.he

par-arneters

themse

1

ves,

three

other

quant.j.

ties

have;:.o

bE:;

pushed

on

the

stack

to

allow

for

automatic

up·'

dating

upon

return

from

the

subroutine.

1.

The

number

of

parameters

wNp·

2 •

/J.

..

link

to

'the

previous

call

"LNKIlt

3.

The

return

address

"RAft

Below

is

shown how

the

JSP

actually

operatese

The

left

stack

shows

the

situation

jus~t

p.rior

to

the

call

of

subro'ttine

2,

the

right

stac.k,

shows

tht~

si

tuation

just

after

the

cal:.

I

...

J

~

~.

Free

stack

area

.

Scratch

1

RAl

I

LNKl

NPl

-

..

Pl

..

O

Pl.l

..

.

•

Stack

just

prior

to

the

call

"JSP

n,

SUB2"

.-

3 1 -

I

I

...

~

I

Free

stack

area

-

RA2

LNK2

NP2

P2.0

p2.1

. . . .

P2.n-2

P2.n-l

Scratch

RAl

LNKl

NPI

Pl.O

Pl.l

1

I

I

i

.,

I

l

stack

just

after

the

call

MJSP

n,

SUB2"

-32-

The

passing

on

of

para.meter.

which

are

pasted

.s

pa,rameters

is

taken

care

of

by

givin9

the

to-be-passed,-on

parameter

an

address

relative

to

the

parameter

pointer,

i.e.,

(RS). A

parameter

following

a

subroutine

call

is

considered

a "new"

parametlerwhen

its

value

is

~'4

and.

a

passed

parameter

otherwise.

See

ex~le

below:

JSP

Pl.0

'

..

Pl.1

Pl.n-l

"

..

n,SUSl

Js,p

.,9UI2

.2.0

P2.1

'2.2

leall

SU8l

Wi1~

n

parameters

/eall

SUB2

with

m

/par

...

tu.

,/par8lD8ter.,

~

63

so

it

.

1.

interpreted

as

a

.....

4

parameter,

not

-1-

but

«R5)+1)

will

be

pushed

on

the

stack.

~hi.

is

just

parameter

Pl.l

of.tbe

previou8