A9(ID:2774/a::007)

Interpreter for SILLIAC 

A9 interpretive programming system for the SILLIAC, Australia's ILLIAC
"A floating point interpretive scheme (called A9 because it used SILLIAC subroutines A1 - A8) was constructed in less than a week to demonstrate programming for a group of teachers. (This system attracted a large following and a compiler group was later constructed for the KDF9 to run A9 programs - 300 times faster than on SILLIAC."


Hardware:
References:
  • Bennett, John M "Computer Science at the University of Sydney 1953-1986" BasserNet Volume 3, Number 6, December 1994. Basser Department of Computer Science U of Sydney view details External link: BasserNet back issue Extract: A9
    We developed on SILLIAC a complete program library for the Snowy Mountains Computer, SNOCOM, before the hardware was available using an interpretive technique.

    A four machine network with 14 keyboard interactive inputs to one of them was developed in the sixties (when networks were a novelty) and was in regular use for years.

    In the late sixties a colour movie of the forming of a shock wave under increasing pressure conditions was made with the PDP8 (using the network).

    A floating point interpretive scheme (called A9 because it used SILLIAC subroutines A1 - A8) was constructed in less than a week to demonstrate programming for a group of teachers. (This system attracted a large following and a compiler group was later constructed for the KDF9 to run A9 programs - 300 times faster than on SILLIAC.

    With a grant of $(US)250,000 from GTE, several networking projects were carried out - including an infrared intra-office network and a network which laid the groundwork for AARNet.

    Various mathematical programming packages (e.g. integer programming and exploiting of tableau structure) were constructed. And so on.

  • Dakin, Robert "a matter of interpretation" Sixteen Bits, July 1998 view details Extract: A9 compiler
    There were two strands to SILLIAC software - machine code and an interpreted language called A9 [...] A9 was more like a set of powerful machien code instructions than a computer language as we know it. It was astill a lot easier to use than anything else that was available, providing convenient access to floating point arithmetic, I/O and array handling facilities.

    A9 provided a much more disciplined programming environment than machine code. You simply couldn't do such fun things as bit twiddling and code modification on the fly, and this meant that there was no compelling reason to use an interpreter. So I took on the task of developing an A9 compiler which generated, and then immediately executed, KDF9 machine code.  [...]

    Compared to the SILLIAC A9 interpreter it was quite spectacular - with a faster processor, hardware floating point and no interpretation overheads. It became rather more than a transition tool - quite a lot of new software A9 was  written for the KDF9.