Skip to main content

Full text of "sds :: sigma :: cp-v :: memos :: CP-V D00 Presentation May75"

See other formats




CP - V D00 RELEASE 



• MULTIPROCESSING 



FECP 



• CP-V "PLUS" FEATURES 



CP-V MULTIPROCESSING 



o Des ign History 

o Hardware Requirements 

o Operations 

o MP in Acti on 

o Performance 



CP-V MULTI-PROCESSING 



o N O PROGRAM CONVERSION 

o NO CHANGE FOR USERS 

o TIGHTLY COUPLED 

o PRIMARY AND 1, 2, OR 3 SECONDARY CPU's 

o IMPROVED AVAILABILITY 

o IMPROVED PRICE/PERFORMANCE 

o IMPROVED THROUGHPUT 



MULTIPROCESSING DESIGN HISTORY 



o Why Mu I tiprocessi ng 



o Implementation Choices 



o Some data 



o More data and models 



^processor 



100%- 



V/////////////////A 

EXECUTION OFy 
OPERATING '/ 
/^SYSTEM CODE '/, 





EXECUTION OF 
CODE TO 
PROVIDE 
SERVICES TO 
USER PROGRAMS 




MASTER MODE 

PRIVILEDGED 

PROCESSING 




MONITOR SERVICE 
REQUEST VIA CAL 
OR PROGRAM TRAP 



* MAJORITY OF CPU TIME 
USER EXECUTION 



EXECUTION OFzd 
USER 
PROGRAM 
PROCEDURE 



SLAVE MODE 
COMPUTING 



16 



isfribufion 



90% OF TOTAL CPU TIME 




COMPUTE TIME BETWEEN CAL'S 
IN MILLISECONDS 



18 



£\$ 



nfiquratio 



SIGMA 9 

CPU 
(8610E) 



MEMORY 
32 K 




HIGH 

SPEED 

RAD 



1 DIO BSA \ 
I .._ J 



SECONDARY STORAGE 



LINE PRINTER 



CARD READER 



COMMUNICATIONS 
76XXOR3625 



28 



CP-V MP IN ACTION 



o INITIALIZATION AND RECOVERY 



o OPERATOR REQUIREMENTS 



o PROCESSORS COMMUNICATE VIA MEMORY 



o SCHEDULER FLOW 



o PERFORMANCE CONTROL 



MP-7 



CP-V MULTIPROCESSING PERFORMANCE 



o Benefits 

o Relative Performance 

o Primary Secondary Utilization 

o Recent Performance Data 



PRIMARY 
PROCESSOR 




SECONDARY 
PROCESSOR . 




25 



MP PERFORMANCE 



to 

CU 



2000 
1500- 
1000- 
500- 




Two 
Sigma 9's 

or 
370/158 

or 
KL 



JzM 



m 





PERFORMANCE GAIN 



:pus 


Thruput 


Hardware 
Capacity 


Percentage 
of Capacity 


i 


1.00 


1.00 


100 


2 


1.75 


1.82 


96 


3 


2.42 


2.55 


95 


4 


3.00 


3.15 


95 



STATISTICS ON-LINE 

INTERVAL FROM 10:32 TO 10:46 OCT 31, 75 



MINS SINCE STARTUP = 


335 








NUMBER OF USERS 


= 


53 








NUMBER OF ONLINE 




40 








NUMBER OF BATCH 


= 


6 








NUMBER OF GHOSTS 




7 








90% RESPONSE TIME 


= 


500 








CPU % 


ALL 


SNAP 


I/O PER MIN 


ALL 


SNAP 


BATCH EXEC 


63.9 


63.3 


SERVICE REQ 


8674 


12585 


BATCH SERV 


11.7 


5.4 


INTERACTIONS 


22 


113 


ONLINE EXEC 


8.8 


21.7 


CHAR IN 


544 


5766 


ONLINE SERV 


6.0 


23.9 


CHAR OUT 


1730 


5458 


GHOST EXEC 


.4 


.6 


TERM WRITES 


82 


266 


GHOST SERV 


J 


.1 


I/O ACCESSES 


1759 


3343 


MONITOR SERV 


11.4 


29.3 


# TRUNCS 


7 


17 


IDLE 


25.6 


.0 


AIR ATTEMPTS 


85 


82 


SWAP WAIT 


J 


J 


AIR HITS 


34 


25 


I/O WAIT 


6.9 


1.4 


SYMBIONT 


71 


121 


I/O&SWP WAIT 


1.1 


5.2 


IN SWAPS 


58 


256 


TOTAL 


135.8 


151.0 


OUT SWAPS 


42 


181 


SCPU USE % 


ALL 


SNAP 


EVENT RATE/MIN 


ALL 


SNAP 


SCPU # 1 EXEC 


36.0 


52.6 


MASTER CALS 


7066 


9745 


SCPU # 1 IDLE 


35.8 


45.4 


SCPU # 1 CALS 


1607 


2840 


SCPU # 1 TOT 


71.8 


98.0 


MASTER SCHEDS 


3474 


6702 








SCPU # 1 SCHEDS 


1790 


3173 



CP-V D00 - 256K Dual Sigma 9 



MULTI-PROCESSING 

WE SAY MASTER/SLAVE BUT HERE 
ARE CP-V's ANONYMOUS FEATURES 

e EACH ADDED CPU IS A COMPUTE PERIPHERAL 

e ONE JOB IS USUALLY WORKED ON BY 

SEVERAL CPU's IN THE COURSE OF ITS 
EXECUTION 

• A NY CPU CA N BE MASTER 

o MASTERS MAY BE CHANGED DYNAMICALLY 

• SINGLE COPY OF THE SYSTEM 

o CPU's CAN BE DYNAMICALLY PARTITIONED OUT 

o ALL PERIPHERALS & FILES ARE AVAILABLE TO ALL JOBS 



CP-V "PLUS" AND OTHER DOO FEATURES 



• REMOTE PROCESSING 

• TIME SHARING TERMINAL 
t> LOADER 

e FILE MANAGEMENT 

• MISCELLANEOUS ENHANCEMENTS 



REMOTE PROCESSING 



• IBM 2780 ERROR RECOVERY 

• IBM 3780 SUPPORT 

e CP-V TO CP-V FILE TRANSMISSION: 
ISCL/RATLER 



TIME SHARING TERMINAL FEATURES - I 



• AUTOSAVE ON HANGUP 

• TERMINAL COUPLING 

• OUTPUT "HOLD", "IGNORE" MODES 

• EASIER INSTALLATION OF TRANSLATION TABLES 



TIME SHARING TERMINAL FEATURES - II 



\ 

• HALF-DUPLEX LINE SUPPORT - 1200 BAUD 

• READ TIMEOUT SPECIFICATION 

• INPUT AND /OR OUTPUT DELETE SPECIFICATION 

• READ CONDITIONAL ON EXISTING INPUT 

• TERMINAL BLOCK/UNBLOCK IN SECONDS 



LOADER 



• 2X TO 10X FASTER 

• LIBRARY LOAD MODULES WITH DSECTS 

• LYNX ON-LINE INTERFACE 

• LEMUR LIBRARY MAINTENANCE 
ROM LOADING FROM LIBRARIES 



1.50- 



1.00- 



.50- 




- r - 
20 



— r - 
60 



40 60 80 

FORTRAN SOURCE LINES 



100 



120 



FILE MANAGEMENT 



• NAME AND ATTRIBUTE CHANGE AT CLOSE 

• MULTIPLE UNDER NAMES FOR EXECUTE 

• PRIVATE PACK REMOUNT AFTER RECOVERY 

• HGP DUMP BY FIX 

• REDUCED OVERHEAD FOR FILE OPENS 



MISCELLANEOUS USER ENHANCEMENTS 



• LOAD AND LINK 

NORMAL MODE FOR FORTRAN DEBUG 

• REAL TIME M:CAL AND COUNTER ZERO 

• DELTA ASSOCIATION AFTER ABORT 

• DELTA BREAKPOINTS AT OVERLAY CHANGE 



MISCELLANEOUS SYSTEM ENHANCEMENTS 



• KEYIN TO CHANGE EXECUTION PRIORITY 

• OPERATOR BROADCAST-IMMEDIATE MESSAGE 

• ON-LINE VOLINIT 

• PRIVILEGED PROCESSORS 

• CONCURRENT SYMBIONT OUTPUT 

• GENMD ENHANCEMENTS 




CP-V REAL MEMORY MAP 



THE CP-V ZOO 



• FROG 

• LEMUR 

• LYNX 

• MOOSE 

• OX 

• PIGEON 

• RATLER 



MULTI-PROCESSING 

• TIGHTLY COUPLED 

e NO CHANGE FOR USERS 

o PRIMARY/SECONDARY CPU's 

o IMPROVED AVAILABILITY 

o IMPROVED PERFORMANCE 

o ADD ONE, TWO, OR THREE CPU's 



2000- 



MP PERFORMANCE 



Q_ 



1500- 



1000- 



500- 



"370/ 
145 



560 




Two 
Sigma 9's 

or 
370/158 

or 
KL10 



Three 
Sigma 9's 



Four 
Sigma 9's 



1.50- 



1.00- 



$ 



.50- 




— r- 
20 



40 



60 



80 



100 



120 



FORTRAN SOURCE LINES 



THESE SLIDES ARE FROM A PRESENTATION ON CP-V VERSION D00, 
MULTIPROCESSING AND "PLUS" FEATURES. 

PRESENTED DECEMBER 3rd AT THE SAN DIEGO MEETING OF 
EXCHANGE, THE XEROX COMPUTER USERS GROUP. 



G.E. BRYAN 
December 1975 



CP-V 



CP-VDOO RELEASE 

» MULTIPROCESSING 

o FECP 

• CP-V "PLUS" FEATURES 



Multiprocessing 

• GENERAL CHARACTERISTICS 

• DESIGN HISTORY 

• MP IN ACTION 

• PERFORMANCE 



Multiprocessing 

GENERAL CHARACTERISTICS 

TIGHTLY COUPLED 

PRIMARY AND 1, 2, OR 3 SECONDARY CPU'S 

NO PROGRAM CONVERSION 

NO CHANGE FOR USERS 

IMPROVED AVAILABILITY 

IMPROVED PRICE/PERFORMANCE 

IMPROVED THROUGHPUT 



Multiprocessing 



SIGMA 9 MP CONFIGURATION 



MEMORY 
32K 



SIGMA 9 
CPU 



MEMORY 
32K 



MEMORY 
32K 



MEMORY 
32K 



MEMORY 
32K 



PROCESSOR BUS 



7^ 

I 

I SIGMA 9 
-J"1 CPU 
I 
I 
I , J 



MIOP x " 
^-^CH 
*' B 



INTERRUPT 



DIO BSA 



COMMUNICATIONS 
76XX OR 3625 



MEMORY 
32K 



HSRIOP 



OPERATOR'S 
CONSOLE 

• MAG TAPE 
h-DISK 

LINE PRINTER 
- CARD READER, ETC. 




Multiprocessing 



2.5- 



o 

o 

o" 
§2.0- 

LU 
O 

E 

a. 

LU 1 R. 

< 

o 

oc. 

D 
a. 

s i.o- 

LU 

>- 




.5- 



T- 

20 



"T" 
30 



10 



40 



50 



60 



70 



NUMBER OF ON-LINE USERS 



Multiprocessing 

MULTIPROCESSING DESIGN HISTORY 

• WHY MULTIPROCESSING 

• IMPLEMENTATION CHOICES 

• SOME DATA 

• MORE DATA AND MODELS 



Multiprocessing 



DUAL REDUNDANT SYSTEMS 



MEMORY 



CPU 



MEMORY 



IOP 



-TAPE 



DISK 



-CARDS 



k- 



POWER PARTITIONING 



-PRINTERS 



L-RAD 



TAPE- 



DISK- 



CARDS- 



PRINTERS- 



RADJ 



MEMORY 




Multiprocessing 



LOOSELY COUPLED SYSTEMS 



MEMORY 




IOP 



DISK 



TAPE TAPE 
CARDS PRINTERS 
COMM. LINK 



MEMORY 




Multiprocessing 



TIGHTLY COUPLED SYSTEMS 



MEMORY 



MEMORY 





MEMORY 



MEMORY 




DISK 

TAPE 

CARDS 

PRINTERS 

COMMUNICATIONS 



MONOPROCESSOR CPU UTILIZATION 



ioo%- 



TT-r-r-rr-rrr-r-r. 
/EXECUTION OF/ 
/OPERATING A 
/SYSTEM CODE , 

EXECUTION OF 
CODE TO 
PROVIDE 
SERVICES TO 
USER PROGRAMS 
I I I I I I IJ 1 1 1 1 1 I 



ft 



MASTER MODE 

PRIVILEDGED 

PROCESSING 



{ 




MONITOR SERVICE 
REQUEST VIA CAL 
OR PROGRAM TRAP 



-execution of! 
:user 

-PROGRAM 

: PROCEDURE : 



SLAVE MODE 
COMPUTING 



Multiprocessing 



CP-V CAL PROCESSING DISTRIBUTION 



>• 
o 

z 

Ul 

a 

LU 
CC 
LL 





U — 90% OF TOTAL CPU TIME— ►! 


/ ,/0 \ 

/ BOUND \ 


1 /COMPUTE \ 1 
1 / BOUND \ 1 


1 


— ! 1 



1 



10 



50 



COMPUTE TIME BETWEEN CAL'S 
IN MILLISECONDS 



Multiprocessing 

CP-V MP IN ACTION 

INITIALIZATION AND RECOVERY 
OPERATOR REQUIREMENTS 
PROCESSORS COMMUNICATE VIA MEMORY 
SCHEDULER FLOW 
PERFORMANCE CONTROL 



Multiprocessing 

INITIALIZATION AND RECOVERY 

• MASTER IS Fl RST CPU TO START 

o BOOT IDENTICAL TO MONO PROCESSING 

• MOOSE GHOST INITIALIZES SLAVE CPU'S 

• MASTER OR SLAVE REQUESTS RECOVERY 

• MASTER PERFORMS AUTOMATIC RECOVERY 

• OPERATOR SELECTS MASTER 



Multiprocessing 



CP-V MP IN ACTION 



MEMORY 



























J 


USER 






















PAGE 
1 




PROCESSOR 

CONTROL 

TABLES 




SLAVE'S 

PAGE 

1 






i 




i 


\ 


\ ■ 


k 
i 






MASTER 




SLAVE 








C 


:pi 


J 










CPU 







Multiprocessing 



CP-V MP SCHEDULING 



1 



SCHEDULED 
SWAP 



/SCHEDULE 
\ SLAVE 



SCHEDULE 
MASTER 



IDLE 



ENTER 



Multiprocessing 



MULTIPROCESSING AND SHARED PROGRAMS 



MEMORY 



CP-V 



APL 



USER1 



USER 2 





Multiprocessing 

MP PERFORMANCE CONTROL 

• MINIMUM SLAVE QUANTUM - 200 MSEC 

• MAXIMUM SLAVE QUANTUM - 5000 MSEC 

• MS/CAL THRESHOLD - 



Multiprocessing 

CP-V MULTIPROCESSING PERFORMANCE 

• RELATIVE PERFORMANCE 

• PRIMARY SECONDARY UTILIZATION 

• RECENT PERFORMANCE DATA 



Multiprocessing 



MP PERFORMANCE 



2000- 



1500- 



1000- 



500- 



370/ 
145 



560 
6640 



SIGMA 9 
OR 
K110 
6660 . 



TWO 

SIGMA 9'S 

OR 

370/158 

OR 

KL10 

6680 



THREE 
SIGMA 9'S 



FOUR 
SIGMA 9'S 



Multiprocessing 



PRIMARY 
PROCESSOR 



SECONDARY 
PROCESSOR 




Multiprocessing 



PERFORMANCE AGAINST CAPACITY 



# CPUs 


CP-V 
THRUPUT 


HARDWARE 
CAPACITY 


PERCENTAGE 
OF CAPACITY 


1 


1.00 


1.00 


100 


2 


1.75 


1.82 


96 


3 


2.42 


2.55 


95 


4 


. 3.00 


3.15 


95 



Multiprocessing 



STATISTICS ON-LINE 

INTERVAL FROM 10:25 TO 10:40 OCT 20, '75 

MINS SINCE STARTUP= 590 

NUMBER OF USERS = 49 

NUMBER OF ON-LINE = 38 

NUMBER OF BATCH = 5 

NUMBER OF GHOSTS = 6 

90% RESPONSE TIME = 350 



CPU % 


ALL 


SNAP 


I/O PER MIN 


ALL 


SNAP 


BATCH EXEC 


16.7 


41.1 


SERVICE REQ 


7268 


10458 


BATCH SERV. 


15.6 


7.5 


INTERACTIONS 


8 


56 


ONLINE EXEC 


7.0 


65.1 


CHAR IN 


141 


804 


ONLINE SERV 


2.0 


17.0 


CHAR OUT 


1126 


6589 


GHOST EXEC 


.9 


.3 


TERM WRITES 


48 


250 


GHOST SERV 


.0 


.0 


I/O ADDRESSES 


1628 


2299 


MONITOR SERV 


3.9 


17.2 


#TRUNCS 


2 


28 


IDLE 


38.5 


2.2 


AIR ATTEMPTS 


229 


96 


SWAP WAIT 


.5 


.1 


AIR HITS 


42 


31 


I/O WAIT 


22.4 


5.4 


SYMBIONT 


40 


185 


l/O&SWPWAIT 


.7 


.5 


IN SWAPS 


17 


84 


TOTAL 


102.2 


176.4 


OUT SWAPS 


14 


63 


SCPU USE % 


ALL 


SNAP 


EVENT RATE/MIN ALL 


SNAP 


SCPU #1 EXEC 


8.2 


76.4 


MASTER CALS 


7061 


9484 


SCPU#1 IDLE 


19.9 


23.0 


SCPU #1 CALS 


207 


974 


SCPU #1 TOT 


28.1 


99.4 


MASTER SCHEDS 


911 


4267 








SCPU #1 SCHEDS 


227 


1181 



CP-V D00 DUAL SIGMA 9 256K 



Multiprocessing 

EACH ADDED CPU IS A COMPUTE PERIPHERAL 

ONE JOB IS USUALLY WORKED ON BY SEVERAL 
CPU'S IN THE COURSE OF ITS EXECUTION 

ANY CPU CAN BE MASTER 

MASTERS MAY BE CHANGED 

SINGLE COPY OF THE CP-V SYSTEM 

CPU'S CAN BE DYNAMICALLY PARTITIONED OUT 

ALL PERIPHERALS & FILES ARE AVAILABLE TO ALL 
JOBS 



CP-V DOO 



CP-V "PLUS" AND OTHER DOO FEATURES 

• REMOTE PROCESSING 

• TIME SHARING TERMINAL 

• TRANSACTION PROCESSING 

• LOADER 

• FILE MANAGEMENT 

e MISCELLANEOUS ENHANCEMENTS 



CP-V DOO 

REMOTE PROCESSING 

• IBM 2780 ERROR RECOVERY 

• IBM 3780 SUPPORT 

• CP-V TO CP-V FILE TRANSMISSION: 

ISCL/RATLER 



CP-V DOO 



TIME SHARING TERMINAL FEATURES - I 

AUTOSAVE ON HANGUP 

TERMINAL COUPLING 

OUTPUT "HOLD", "IGNORE" MODES 

PAGE ERASE SEQUENCES FOR TEKTRONICS AND 
TTY 40 

TERMINAL COMMANDS ON FILE -OX 



CP-V DOO 



TIME SHARING TERMINAL FEATURES - II 

• HALF-DUPLEX LINE SUPPORT - 1200 BAUD 

• READ TIMEOUT SPECIFICATION 

• INPUT AND/OR OUTPUT DELETE SPECIFICATION 

• READ CONDITIONAL ON EXISTING INPUT 

• TERMINAL BLOCK/UNBLOCK IN SECONDS 



4A 



CP-V DOO 

TRANSACTION PROCESSING 

• TP/TS SWITCHING OF TERMINALS 

• ADDITIONAL CONTROL 

• IMPROVED ERROR ANALYSIS 

• IMPROVED REFERENCE MANUAL 



\W 



CP-VDOO 

LOADER 

2XT0 1 OX FASTER 

LIBRARY LOAD MODULES WITH DSECTS 

LYNX ON-LINE INTERFACE 

LEMUR LIBRARY MAINTENANCE 

ROM LOADING FROM LIBRARIES 



A- 



CP-V DOO 



LOADER SPEED 



1.50- 



1.00-1 
$ 

.50-1 




20 



T 
40 



-T 
60 



— T" 
100 



80 



120 



FORTRAN SOURCE LINES 



c-a 



CP-V DOO 

FILE MANAGEMENT 

NAME AND ATTRIBUTE CHANGE AT CLOSE 
MULTIPLE 'UNDER' NAMES FOR EXECUTE 
PRIVATE PACK REMOUNT AFTER RECOVERY 
HGP DUMP BY FIX 
FASTER FILE OPENS VIA CFU SEARCH 



•?A 



CP-V DOO 



MISCELLANEOUS USER ENHANCEMENTS 

LOAD-AND-LINK 

NORMAL MODE FOR FORTRAN DEBUG 
REAL TIME M: CAL AND COUNTER ZERO 
DELTA ASSOCIATION AFTER ABORT 
DELTA BREAKPOINTS AT OVERLAY CHANGE 



<,-* 



CP-V DOO 



MISCELLANEOUS SYSTEM ENHANCEMENTS 

KEYIN TO CHANGE EXECUTION PRIORITY 

OPERATOR BROADCAST-IMMEDIATE MESSAGE- 
PIGEON 

ON-LINE VOLINIT 

PRIVILEGED PROCESSORS 

CONCURRENT SYMBIONT OUTPUT 

GENMD ENHANCEMENTS 



<?r 



CP-V DOO 



CP-V REAL MEMORY MAP 




; TRAPS: 



I I I i i i i i 'i t i t 'i i 'i n 
SYSTEM MANAGEMENT 

, i A.) nt. ii.nl I,,, u 1 i„, ' „;., ,i.,„;.„!„ / m,',,, 1 ,. 



. : ;: : -:.- :: .::::|OQ" : v 



FILE MANAGEMENT 



m 



TT 
JIT 



il/O INTERRUPT HANDLERS 



4U~A 



:lA 



CP-V DOO 



STATISTICS ON-LINE 

INTERVAL FROM 10:38 TO 10:53 NOV 03, f 75 
MINS SINCE STARTUP = 457 



NUMBER OF USERS = 


43 








NUMBER OF ONLINE = 


32 








NUMBER OF BATCH = 


6 








NUMBER OF GHOSTS = 


5 








90% RESPONSE TIME = ! 


500 








CPU % 


ALL 


SNAP 


l/OPERMIN 


ALL 


SNAP 


BATCH EXEC 


31.3 


117.0 


SERVICE REQ 


10483 


8168 


BATCH SERV 


22.1 


13.8 


INTERACTIONS 


11 


41 


ONLINE EXEC 


18.3 


19.9 


CHAR IN 


157 


511 


ONLINE SERV 


2.5 


9.6 


CHAR OUT 


1331 


4845 


GHOST EXEC 


1.2 


.4 


TERM WRITES 


51 


202 


GHOST SERV 


.0 


.0 


I/O ACCESSES 


2091 


2297 


MONITOR SERV 


7.2 


19.5 


#TRUNCS 


4 


17 


IDLE 


22.2 


.0 


AIR ATTEMPTS 


245 


235 


SWAP WAIT 


.3 


.0 


AIR HITS 


232 


173 


I/O WAIT 


15.5 


.3 


SYMBIONT 


72 


314 


l/O&SWPWAIT 


1.0 


.5 


INSWAPS 


48 


149 


TOTAL 


121.8 


180.9 


OUT SWAPS 


38 


110 


EVENT RATE/MIN ALL 


SNAP 


SCPU USE % 


ALL 


SNAP 


MASTER CALS 


10068 


6050 


SCPU #1 EXEC 


23.4 


80.9 


SCPU #1 CALS 


415 


2118 


SCPU#1 IDLE 


23.7 


17.5 


MASTER SCHEDS 


1524 


5384 


SCPU #1 TOT 


47.1 


98.4 


SCPU #1 SCHEDS 


479 


2228 








RESOURCES IN USE 


■ # 


RAHD-AIR/MIN 


ALL 


SNAP 


MPOOLS 







#AIR ATTEMPTS 


245 


235 


COCBUFS 




22 


#AIR HITS 


232 


173 


IOQ ENTRIES 




3 


MIR TIMEOUTS 


4 


1 


CFUS 




37 


#RAHD ATTEMPTS 18 


101 


GRAN PACK 




36702 


#RAHD STARTS 


18 


100 


GRAN SYMB 




3957 


#RAHD USED 


16 


97 


GRAN RAD 




1272 


#RAHD I/O WAIT 


2 


6 








#RAHD NOT USEC 


I 











#RAHD TIMEOUTS 2 


1 



CP-V DOO DUAL SIGMA 9 256K 



/A 







CP-V DOO 

THE CP-V ZOO 

FROG -FECP CONTROL 

LEMUR - LIBRARY MAINTENANCE 

LYNX - ON-LINE LOADER 

MOOSE - MULTIPROCESSING CONTROL 

OX - ON-LINE COMMAND FILES 

PIGEON - BROADCAST MESSAGES TO TERMINALS 

RATLER - CP-V FILE TRANSMISSION 



!^