Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Microcap & Penny Stocks : Patriot Scientific - PTSC -- Ignore unavailable to you. Want to Upgrade?

To: bob who wrote (5530)	9/13/1998 8:55:00 PM
From: cksla	Read Replies (2) \| Respond to of 8581

bob- the latest epistle by Keith Wooten: message ÿAuthor:ÿKeith Woottenÿauthorÿprofile emailÿreply viewÿthread postÿnewÿúÿpostÿreply subscribe Email:Keith@wootten.demon.co.ukDate:1998/09/13Forums:comp.lang.forth view for bookmarking ÿúÿ text only ÿ ------------------------------------------------------------------------ In article <6tg2g3$5sf$1@news.IAEhv.nl>, Marcel Hendrix <mhx@iaehv.iaehv.nl> writes [snipped] >This seems reasonable, but it isn't so simple. The PSC1000, P21, F21 >get their peak speed by executing operations in parallel (a 20 bit word >can contains 4 instructions if no immediate data or address is >necessary). The 68060 does not have parallel execution units and gets >speed by pipelining. There seems to be some confusion here regarding PSC1000 (ShBoom) and P21, F21. As I understand it, the original ShBoom was a Chuck Moore design which was not produced. P21 and F21 are current Chuck Moore and Jeff Fox MISC designs which are in various prototytpe stages. The Patriot Scientific ShBoom (PSC1000) is based on Chuck's original ShBoom but is a silicon reality currently in 0.5um at 100MHz with 0.35uM at - I think - 150MHz expected very soon. The PSC1000 has some things in common with the MISC chips from the architectural point of view, but is fabricated using normal, everyday techniques. National Semiconductor made the first ones, and I believe the current parts are made by Taiwan Semiconductor - the point is, they can be made at any modern foundry. The PSC1000 does not execute instructions in parallel (there is a separate IO processor which can run in parallel with the CPU in some circumstances), but fetches four 8bit opcodes in parallel from 32bit memory; these are executed sequentially at full clock rate. (Incidentally, the part may be configured to run from 8bit memory where performance is not a priority - it's still makes an inexpensive embedded processor.) [snipped] > >Let me explain why I think 1/100th the price can't be true. First >question: 1/100th of what? You mentioned a 68060 system that couldn't >do what was needed. We can double the price and get two 68060s (one for >the IRQ which is then a simple poll, one for the rest). I'm sure the >68060 plus minimum peripherals is less than $200, so the cost would be >$400. To make your claim of 1/100 true, I must assume a PSC1000 costs >$4. Now wouldn't it be very stupid to sell this chip for only $4? I >would sell it for $200, or maybe $395. > I agree that the numbers game gets a bit out of hand, but I think the point is that the MISC chips - if they become working silicon - could be many times faster, smaller and cheaper. This could allow, for example, multiple processors on a chip with memory for intensive tasks such as image recognition at low cost. The PSC1000 supposedly costs $10 in quantity, I've paid nearer $40 (in the UK) for small quantities. If you want an embedded processor running Forth at high speed, it's great. Maybe you tell your boss that your using a macro assembler on this stack machine (I don't) so as not to use the F-word. Cheap, Fast, Good, Available - all four? (Sorry, Wolfgang) Bye from England -- Keith Wootten