The CDC 6000 Series Computer

---

---

The CDC 6000 family started with the 6600, at the time of its introduction the fastest computer in the world. After the 6600 was introduced, Control Data Corporation introduced the 6400, a slower and cheaper model of the 6600, and that is the machine U. Ammann mentions implementing a Pascal compiler for. [1]

The 6600 was designed by Seymour Cray. In many respects there are parallels between Seymour cray and Niklaus Wirth. Both believed that a clean, sparse and well constructed archetecture could outperform a complex and arbitrarily constructed one.

Perhaps most telling of Seymour Cray was the exchange Thomas Watson, the CEO of IBM, had with Seymour Cray: [3]

"Last week Control data... announced the 6600 system. I understand that in the laboratory developing the system there were only 34 people including the janitor. Of these, 14 are engineers and 4 are programmers.. Contrasting this modest effort with our vast development activities, I fail to understand why we have lost our industry leadership position by letting someone else offer the world's most powerful computer." - Watson

"It seems like Mr. Watson has answered his own question." - Seymour Cray

The 6600 had 60 bit words with 18 bit addresses, for a potential of 256k words (or about 2 megabytes) of memory, but was commonly only shipped with 128k words, while the 6400 was commonly equipped with 64k words. It processed 10 million instructions per second, awesome speed for 1964.

Although the Design of Pascal was not supposed to be tied to a particular machine, there is actually a lot to suggest that features of Pascal have roots in the CDC 6000 series:
 

• The CDC 6400 had a 64 element character set (6 bits per character), and Wirth describes sets as consisting of 59 elements [2]. So character sets on the machine fit in a single machine word, and were easily manipulated.
• The CDC 6400 could contain 10 6 bit characters per word, which made 10 character strings a special, efficient case. This matches Pascal's original 8 character identifier "suggested limit" on length. The Pascal compiler was able to store identifiers in a single machine word.
• The CDC 6400 could not address individual bytes, but was a (60 bit) word only machine. Hence, Pascal's emphasis on packed types, packed strings, and the pack/unpack procedures.
• The CDC 6400 did not have EOLN as a character, but instead signaled End-of-Line condition via record structure in text files. Hence, there were good reasons for Wirth to implement eoln and eof as functions and not characters.
• Mark Riordan (see web links below) describes the difficulties with getting Scope/NOS (the OS Wirth used) to access a permanent file by name from the program itself. Basically, external job control instructions were used to designate a "local" file as an alias for a named permanent file. Local files used by each program were deleted (considered temporary files) by the OS at the end of a job. Sound familar ? These limitations go a long way to explain why Pascal only accepted existing files as program header arguments whose names were provided elsewhere, and why all internal files were nameless temporary files.

None of this is "bad", it just indicates that the machine you use can influence your work in ways you don't think about. People who know me, know I am a big Wirth fan. However, I believe in an objective view. Wirth tried very hard to make Pascal machine and system independent, but could not have come up with a construct that would have been difficult or impossible to implement on the host CDC 6400 machine.

The CDC 6600 has a simulator so that you can run software for that computer:

http://members.iinet.net.au/~tom-hunter/

A good historical page on the CDC 6600 is:

http://www.cs.nmsu.edu/~pfeiffer/classes/473/notes/cdc.html

Finally, there is a newsgroup, Control Freaks, that deals with matters CDC.

Courtesy of the first class bitsavers project, I have put together a list of manuals for the CDC 6000 computers. Here they are:

6600systemBul3_Oct66.pdf                     The systems bulletin for 6400/6600. Appears to be a newsletter.

020167_6600_CP_Vol1_Mar67.pdf       Working theory of the 6600 CPU.

60239700_6600_CP_Vol2_Mar67.pdf    6600 CPU theory of operation, volume 2.

60045100_6000_IO_Spec_Sep67.pdf      Hardware manual for 6400/6500/6600 input/output units.

60100000D_6600refMan_Feb67.pdf        Reference manual for 6400/6500/6600 computers, including instruction set.

60141900_Codes_Rev-A_May65.pdf      A compilation of hardware information, instruction and I/O unit codes.

60156100G_periphRef_Jan69.pdf             Programming manual for tape, card reader, etc.

60156100J_periphRef_Mar70.pdf             Later version of above.

60173800_ScopeRefMan_Sep66.pdf       Scope operating system programming.

60305200D_SCOPE3.3ug_Sep71.pdf     Scope reference, later version. [Transfer problem]

60307200D_SCOPE3.4ref_Oct72.pdf     Again.

60179000_Scope3.0oper_Apr67.pdf        Scope operaing guide.

60189100_Export-Import_8231.pdf          Remote terminal management.

60189400L_SCOPE3.1.4_Apr71.pdf       Later version of Scope reference manual.

60225100_extdCoreStor_Feb68.pdf         Extended core hardware and programming.

60250400_6000IntroAndPProc.pdf           6000 Series overview

60257200_6400_CP_TrM_Feb67.pdf      6400 CPU hardware repair training.

60299500_SMM_Instant_Nov69.pdf        Systems diagnostics manual.

60306500A_1971_SCOPE34_SPRM.pdf Scope programmers reference manual.

60307400C_SCOPE3.4ins_Nov72.pdf      Scope installation handbook.

60334000_6603DiskControl.pdf                Disk controller reference.

60342500H_UPDATE_RefMan.pdf          Update manual, a source deck manager.

60344000A_SCOPE2ins_Dec72.pdf        Scope 2 installation manual.

60363900_7054_OperAndPgmg.pdf        Disk storage operation and programming.

60372600A_SCOPE2ug_Nov72.pdf       Scope 2 Users guide.

60372600E_SCOPE2.1ug_Mar78.pdf     Scope 2.1 Users guide.

82100010_dd60a_Mar65.pdf                  Control console hardware manual.

References:

[2] Pascal: Users Manual and Report [Jensen and Wirth].

[3] To be fair, this is a famous "exchange" documented in several places, including the web and in books, but there is little hard proof that it actually took place. If one of you has a hard reference to this, I would appreciate it.