Talk:TX-2

Memory using magnetic films

In an article comparing the FX-1 to the TX-2, a description of cache memory appears:

FX-1 -- Lincoln Laboratory (1961) Also a part of TX-2 is a small fast memory using thin magnetic films, the first such memory to be installed in a computer. -- Jamplevia (talk) 19:51, 21 February 2022 (UTC)

The TX-2 used different technologies for different parts of its memory. For example the S-memory was core memory while the T-memory was based on transistors (both were part of a linearly addressable RAM as we'd describe things today). Neither of those was cache and no cache was mentioned in the comprehensive TX-2 Technical Manual (which dates from 1961-1963). The TX-2 was an experimental machine, The TX-2 was changed quite often, so it's possible the memory described in the ONR Digital Computer Newsletter was in fact cache as we'd understand it today, but fitted later in the lifetime of the machine. JamesYoungman (talk) 18:10, 6 July 2023 (UTC)

There's a reprint of that very interesting article on page 1B of the October 1961 Computers and Automation magazine, as "The FX-1 Computer at Lincoln Laboratory" by John A. Kessler, with some photos as well as nicer print. RW Dutton (talk) 20:38, 8 October 2025 (UTC)

Prof. Sutherland states (in an email to User:JamesYoungman on 2025-07-2025):

The “index” memory was a special thin-film random access memory built from

dots of permalloy on glass substrates. It had an incredibly short access time,

but less capacity than main core memory.

Here the "index" memory is the storage for the index registers (X-registers in the machine's assembly language). The description seems to match Plated-wire memory. JamesYoungman (talk) 20:24, 3 August 2025 (UTC)

Compiler

In their edit of 2019-09-04T00:14:33, 888 Manny added the text "The compiler was developed Lawrence Roberts..." but it is not clear from the text (which is current now) what compiler is being described. There is a BCPL compiler from 1969, but Roberts' work was much earlier (1963). If he did develop a compiler, this work would be important even if only because of its early date in the development of TX-2 software.

JamesYoungman (talk) 18:21, 6 July 2023 (UTC)

Roberts refers to the "compiler" in the final page of his 1963 Ph.D. thesis. In an in-person conversation (2025-08-04) with JamesYoungman, Ivan Sutherland confirmed to me that this was a reference to the TX-2's assembler (which was called "M4"). So I have updated the text to add this clarification. I was not sure how to cite an in-person clarification like this, and I also was not sure that the difference is material enough to require some kind of reference. Advice appreciated. -- JamesYoungman (talk) 07:14, 5 August 2025 (UTC)