This is an old revision of the document!
This page is not limited to TOPS-20, but includes also information on other operating systems that run on the PDP-10 architecture, as there is a great deal of overlap and much PDP-10 software can run on either system with little or no modification. It may also serve as an instructive comparison between the operating system versions and hardware models to show what hardware support is required for each version.
(Some pictorial illustration would be helpful here, to lighten up an otherwise dense piece of tables and text)
(A very brief historical background of the PDP-10 architecture could be useful here, just a single passage or two)
Much of the information in this section has been obtained from the following documentation:
The PDP-10 is originally designed for a memory with a maximum capacity of 218 or 262,144 words (this maximum has since been substantially increased; see the section on Extended addressing for details), each word holding 36 bits of data.
The processor has 16 general-purpose registers or accumulators, each also 36 bits wide, numbered from 0 to 178 (AC0–AC17). These accumulators may be addressed also as part of the main memory, meaning that an operation referring to memory cell 0000178 will actually access AC17, wherefore the contents of any accumulator can be copied to or otherwise used together with any other accumulator in a single instruction.
15 of these accumulators, AC1–AC17, may also serve as index registers; see the section on Effective address calculation for details.
The processor also has an 18-bit program counter, a set of status flags, plus a few other special registers for internal use.
The PDP-10 uses a 36-bit word format, with bits conventionally numbered 0–35 from most to least significant bit (big-endian). Portions of this word may be referenced as arbitrarily sized bytes (1–36 bits long) and used to represent smaller data types such as characters of text.
|6-bit bytes||Byte 0||Byte 1||Byte 2||Byte 3||Byte 4||Byte 5|
|7-bit bytes||Byte 0||Byte 1||Byte 2||Byte 3||Byte 4||0|
|8-bit bytes||Byte 0||Byte 1||Byte 2||Byte 3||0|
|S||1||0||0: number is positive; 1: number is negative|
|N||35||1–35||Two's-complement integer (excluding sign)|
|Exponent||8||1–8||Binary exponent + 128|
|Fraction||27||9–35||Two's complement binary fraction|
As 36 is divisible by three, in PDP-10 assembly language programming integer numbers are often written in octal (base 8) rather than hexadecimal (base 16) or decimal (base 10) notation. Exceptions here are bit positions, byte sizes, and software major/minor version numbers, which are written in decimal notation. When confusion is likely, decimal integers are suffixed with a period (“.”).
|Position||6||0–5||First (most significant) bit of byte field|
|Size||6||6–11||Number of consecutive bits in byte field|
|Basic instruction||Basic opcode||AC||I||X||Y|
|I/O instruction||7||Device||I/O opcode||I||X||Y|
|Flags||13||0–12||See Processor status flags|
|Basic opcode||9||0–8||See Basic instruction set|
|AC||4||9–12||Accumulator to be used|
|Device||7||3–9||Input/output device unit|
|I/O opcode||3||10–12||See Input/output instruction set|
Note: I/O instructions are not available when the processor is in user mode, but can be executed only by the operating system kernel in system (privileged) mode. They are thus of limited interest to regular application programmers under most operating systems.
Note: As with I/O instructions, hardware interrupts are not directly available to programs running in user mode. The operating system however provides software interrupt signals to user-mode applications.
|Processor||Includes||Memory size||Sections||I/O||Front end||Operating systems|
|DEC KA10||8–256 KW||1||—|
|DEC KI10||?–4096 KW||1||—|
|DEC KL10 A||FlPt||?–4096 KW||1||Console|
|DEC KL10 B||FlPt, ExtA||?–4096 KW||32||4 DTE20, 8 RH20||PDP-11/??||?–7|
|DEC KS10||FlPt||?–512 KW||1||Console?||?–5|
|XKL Toad-1||FlPt, ExtA||?–32 MW||?||?–7|
|XKL Toad-2||FlPt, ExtA||?–256 MW||4096||?–7|
|FlPt||Floating point operations|