[HN Gopher] The Transputer Handbook (1990) [pdf]
___________________________________________________________________
The Transputer Handbook (1990) [pdf]
Author : twoodfin
Score : 30 points
Date : 2022-08-20 17:34 UTC (5 hours ago)
(HTM) web link (www.transputer.net)
(TXT) w3m dump (www.transputer.net)
| mark_undoio wrote:
| The Transputer always sounded like a computer from an alternate
| future when I read about them before.
|
| I don't know how practical the tech would have been to scale but
| I wonder how different computing could have been.
|
| A conceptually somewhat related system (in the sense of a
| computer as a network of smart-ish components) was the (research
| prototype) Desk Area Network
| https://www.cl.cam.ac.uk/research/srg/netos/projects/archive...
|
| Mainframes had significantly powerful IO processors, meanwhile
| modern PCs contain many processors linked by effectively an
| internal network. With the right network hardware you can DMA to
| another system too. So I guess the lines have always been a
| little blurry between systems!
| Rochus wrote:
| Occam-2 was definitely not my most favorite language, but an
| interesting experience. Thanks for the link; I only had the
| original manuals and Pio2 at the time. Using
| http://any2djvu.djvu.org/ I was able to generate a fast, compact
| version of the referenced PDF.
|
| Btw. Pio2 can be accessed from here:
| http://www.cs.ox.ac.uk/people/geraint.jones/publications/boo...
| atrn wrote:
| I did a lot of work with transputers and occam in the 1980s and
| wrote an OS for a machine that used them. The exposure to the
| CSP way of thinking was a blessing. occam-2 on the other hand
| was a little too primitive.
|
| I remember when Inmos marketing came up with "We wanted to make
| occam the FORTRAN of parallel processing". My response was "You
| have." (a reference to occam-2 having only static arrays and no
| concept of a record/struct)
| 01100011 wrote:
| Wtf... I was just talking about transputers with a coworker
| yesterday. The ST20 was a popular core in early settop boxes(i.e
| DirecTV) and was based on a lot of transputer concepts.
| codetrotter wrote:
| I wanted Dall-E 2 to show me what a Transputer might look like,
| so I entered the following as prompt for Dall-E 2, based on the
| text from the preface and the intrpoduction of the linked PDF:
|
| > The software and hardware implementation of transputer parallel
| processing systems. Why do we need all this computing power, and
| how are we going to be provided with it? Modern developments have
| shown that good software should be intuitive. If the user has
| learned one program package then the next package, albeit from a
| different software house, should behave in a similar fashion.
|
| Results:
|
| * https://i.imgur.com/koOGgFI.png
|
| * https://i.imgur.com/BuU8YrX.png
|
| * https://i.imgur.com/7fZplFV.png
|
| * https://i.imgur.com/mrMUHEp.png
|
| The generated images are very stock photo looking. And
| unfortunately also not very informative..
| kitd wrote:
| Entertainingly, the second one has been flagged as potentially
| "erotic".
|
| That's parallel processing for you...
| codetrotter wrote:
| Another attempt.
|
| Prompt:
|
| > A photo in the style of a still from the 1995 science fiction
| movie Johnny Mnemonic, showing a man sitting in front of a
| Transputer.
|
| Results:
|
| * https://i.imgur.com/5ESVag2.png
|
| * https://i.imgur.com/8APGO8b.png
|
| * https://i.imgur.com/n38em4Q.png
|
| * https://i.imgur.com/rM4bDit.png
|
| I like these more. Still not sure the output is in any way
| influenced by the word "transputer" though, as these look like
| pretty regular computers.
|
| Guess I am left with no choice but to read the PDF itself. And
| hope that it's worth it :p
| jschveibinz wrote:
| In the early 80's, optical signal processing was the next big
| thing in array processing. Multiply-accumulates we're done in
| optics and on the camera detector, and it was super fast, but the
| dynamic range was limited to less than 12 bits. By 1990, optical
| signal processing was dying and parallel array processing was the
| "next big thing" in array processing. Systolic arrays and
| transputers were going to be array processing engines of the
| future. The problem was to optimally fit the algorithm into the
| computing architecture---compilers were complicated. By
| 1995-2000, general purpose processors, operating systems and
| MATLAB were fast enough to put most of that to bed except for
| graphics cards and special effects. GPU's continue to be the
| latest in the history of array processing hardware solutions. But
| I believe their days are also numbered with the advent of quantum
| computing systems and other technologies.
| zackmorris wrote:
| This is great!
|
| I independently figured a lot of this stuff out around 1995 while
| writing shareware games in college (limited by absurdly slow 30
| MHz memory busses and single-threaded execution before video
| cards) and 2005 at a job where I used MATLAB. Basically that
| computers ran many thousands of times slower than they should
| have, which today translates to many millions of times slower for
| the transistor counts. We perceive computers as fast today, but
| that's just the video card doing most of the work, at enormous
| programmer expense due to having to use separate languages and
| paradigms separate from the host OS.
|
| The article jumps the shark at chapter 5 Transputer Languages on
| page 63, so I'd recommend skipping that. It's fun to learn about
| stuff like Occam, but it's kind of the assembly language of
| Erlang or Go (if Go had full process isolation) so is probably
| too low-level for large programs to be understandable by humans.
| They didn't really understand process isolation like we do today,
| so got lost in the weeds with libraries/intrinsics/threads,
| rather than having a language self-parallelize all code like
| something like Julia (from what I understand, I haven't fully
| learned it yet).
|
| I'd largely skip Chapter 6 Software Environments on page 88 as
| well, which covers various environments for transputers, which
| are fun for their historical significance. But again, I wouldn't
| use any of them today, since we have much simpler approaches with
| stuff like containerization now.
|
| Chapter 7 Transputer Family Hardware on page 101 is quaint. The
| principles are good, but break down with the memory and network
| speeds available today and the fact that we work at higher levels
| of abstraction.
|
| I didn't know there was an Atari Transputer Workstation (ATW) in
| 1988, wow!
|
| I've written extensively about multicore processing in my
| comments here for over a decade, with no meaningful progress in
| industry from what I can tell. Personal hardware got abandoned
| for stuff like elastic compute on AWS.
|
| That said, it's pretty obvious what a transputer today would look
| like. I'd suggest a 1 GHz processor with at least 1 GB of onboard
| ram and at least 256 cores for under $100. It should have gigabit
| ethernet and use something like content-addressable memory to
| allow the chips to self-organize with LRU caching for data
| locality. It would use web metaphors so that the location of a
| core or memory doesn't matter, even running across a warehouse or
| over the internet. So 4 of these would look like a single desktop
| computer with a kilocore (kC, like IBM's from 2006) and easily
| scale to probably 64kC or even 1 MC without too much extra work.
| It would be the software equivalent of something like BitTorrent
| vs FTP. And largely buildable with FPGAs, had they continued to
| evolve instead of getting hamstrung by their GPU competition.
|
| It would be programmed with existing C-style languages and a
| self-parallelizing compiler that uses techniques from compiler
| optimization to detect for-loops on immutable data and spread the
| workload over any number of processes. Programs would be inverted
| to identify mutability and treat it as unsafe code that can't be
| statically analyzed or evaluated lazily. So synchronous blocking
| higher-order methods on immutable data would be highly
| encouraged. Stuff like shader memory layout optimizations would
| largely go away due to the use of content-addressable memory and
| the avoidance of hardware caching other than what's needed to
| make the distributed memory appear as a single address space.
|
| It's all so easy and obvious to me that I've all but given up on
| it ever happening. Writing this out for the umpteenth time makes
| me feel tired. To the point that I struggle to program in my
| daily life, because it feels like I'm having to do everything the
| hard way with async approaches (today's GOTO) and highly-
| specialized shaders that break constantly because I just can't
| see what I'm doing or debug them properly so it feels like all I
| ever get is a blank screen.
|
| I wish so badly that a startup would jump on this and bring us
| real computers with real power so that we could play around with
| stuff like genetic algorithms and running machine learning in
| simulation so that we could evolve our own AI digital assistants.
| Instead of having to rely on closed/proprietary/cookie-cutter
| solutions that are pay-to-play or have other artificial
| limitations to prevent terrorism or whatever.
|
| It's the same old story, I'm thrilled to see stuff like DALL-E 2
| bring the predictions of sci-fi to life, but saddened that it
| happened decades later than it should have due to poor hardware
| architecture choices. It's like having to watch electric cars go
| mainstream 20-30 years after the lithium ion battery was
| invented, because the fossil fuel industry prevented that from
| happening every step of the way.
| nickdothutton wrote:
| Brings back memories from university programming the T414.
| __d wrote:
| The T414 was the first useful but affordable Transputer.
|
| It was 32 bit but lacked the floating point hardware of the
| more expensive T8xx series. Later on there were even cheaper
| T4xx chips, which were relatively easy to get on a PC expansion
| card with slightly crippled tool chains for occam.
|
| I still have a couple of T414 boards. I mostly run Helios on
| them: a parallel universe version of Plan9.
| bear8642 wrote:
| Just to check - by Helios you mean
| <https://en.wikipedia.org/wiki/HeliOS>? I've not heard of it
| before.
|
| > a parallel universe version of Plan9
|
| I've enjoyed learning about Plan9, how would you say they
| compare as systems?
| adrianmsmith wrote:
| I have this book, still sitting in my bookshelf, from back in the
| day. I never got to use a Transputer for real which is a shame as
| I would have loved to try them out, but the book was still
| excellent and worth the read.
___________________________________________________________________
(page generated 2022-08-20 23:01 UTC)