Rocksolid Light

Welcome to novaBBS (click a section below)

mail  files  register  newsreader  groups  login

Message-ID:  

"I DO want your money, because god wants your money!" -- The Reverend Jimmy, from _Repo_Man_


devel / comp.lang.forth / Re: 6 GHz stack machine

SubjectAuthor
* 6 GHz stack machineStephen Pelc
+* Re: 6 GHz stack machineBrian Fox
|+* Re: 6 GHz stack machineStephen Pelc
||`* Re: 6 GHz stack machineClive Arthur
|| `* Re: 6 GHz stack machinenone
||  `- Re: 6 GHz stack machinePaul Rubin
|`- Re: 6 GHz stack machineBrad Eckert
+- Re: 6 GHz stack machineMarcel Hendrix
+- Re: 6 GHz stack machineFourthy Forth
+* Re: 6 GHz stack machineIlya Tarasov
|+* Re: 6 GHz stack machineJurgen Pitaske
||`* Re: 6 GHz stack machineIlya Tarasov
|| `* Re: 6 GHz stack machineJurgen Pitaske
||  +- Re: 6 GHz stack machineIlya Tarasov
||  `* Re: 6 GHz stack machineHugh Aguilar
||   +- Re: 6 GHz stack machineJurgen Pitaske
||   `* Re: 6 GHz stack machineIlya Tarasov
||    +- Re: 6 GHz stack machineFourthy Forth
||    `- Re: 6 GHz stack machineHugh Aguilar
|+* Re: 6 GHz stack machinePaul Rubin
||`* Re: 6 GHz stack machineIlya Tarasov
|| +* Re: 6 GHz stack machinePaul Rubin
|| |`* Re: 6 GHz stack machineIlya Tarasov
|| | `* Re: 6 GHz stack machinePaul Rubin
|| |  +- Re: 6 GHz stack machineRick C
|| |  `* Re: 6 GHz stack machineIlya Tarasov
|| |   `* Re: 6 GHz stack machinePaul Rubin
|| |    +- Re: 6 GHz stack machineJurgen Pitaske
|| |    `- Re: 6 GHz stack machineIlya Tarasov
|| `* Re: 6 GHz stack machineJurgen Pitaske
||  `* Re: 6 GHz stack machineIlya Tarasov
||   `- Re: 6 GHz stack machineJurgen Pitaske
|`* Re: 6 GHz stack machineStephen Pelc
| +* Re: 6 GHz stack machineFourthy Forth
| |`* Re: 6 GHz stack machineStephen Pelc
| | `- Re: 6 GHz stack machineFourthy Forth
| `* Re: 6 GHz stack machineBrad Eckert
|  +- Re: 6 GHz stack machineJurgen Pitaske
|  +- Re: 6 GHz stack machinedxforth
|  +* Re: 6 GHz stack machineStephen Pelc
|  |`* Re: 6 GHz stack machineJurgen Pitaske
|  | `* Re: 6 GHz stack machineStephen Pelc
|  |  `* Re: 6 GHz stack machineJurgen Pitaske
|  |   `* Re: 6 GHz stack machineStephen Pelc
|  |    `- Re: 6 GHz stack machineRick C
|  `* Re: 6 GHz stack machineRick C
|   `* Re: 6 GHz stack machineBrad Eckert
|    +* Re: 6 GHz stack machinePaul Rubin
|    |`* Re: 6 GHz stack machineBrad Eckert
|    | `* Re: 6 GHz stack machinePaul Rubin
|    |  +* Re: 6 GHz stack machineBrad Eckert
|    |  |+- Re: 6 GHz stack machinePaul Rubin
|    |  |`- Re: 6 GHz stack machineS Jack
|    |  `* Re: 6 GHz stack machineAnton Ertl
|    |   `* Re: 6 GHz stack machineJurgen Pitaske
|    |    +- Re: 6 GHz stack machineJurgen Pitaske
|    |    `* Re: 6 GHz stack machineAnton Ertl
|    |     +* Re: 6 GHz stack machineJurgen Pitaske
|    |     |`* Re: 6 GHz stack machineJurgen Pitaske
|    |     | `- Re: 6 GHz stack machineAnton Ertl
|    |     `- Re: 6 GHz stack machinePaul Rubin
|    `* Re: 6 GHz stack machineRick C
|     +- Re: 6 GHz stack machineFourthy Forth
|     `* Re: 6 GHz stack machineJurgen Pitaske
|      `* Re: 6 GHz stack machineRick C
|       `* Re: 6 GHz stack machineFourthy Forth
|        `* Re: 6 GHz stack machineRick C
|         +- Re: 6 GHz stack machineJurgen Pitaske
|         `* Re: 6 GHz stack machineRick C
|          `- Re: 6 GHz stack machineHugh Aguilar
+* Re: 6 GHz stack machineRick C
|`* Re: 6 GHz stack machineWayne morellini
| +- Re: 6 GHz stack machineRick C
| +* Re: 6 GHz stack machinePaul Rubin
| |`* Re: 6 GHz stack machineWayne morellini
| | `* Re: 6 GHz stack machinePaul Rubin
| |  `* Re: 6 GHz stack machineRick C
| |   +* Re: 6 GHz stack machineWayne morellini
| |   |`* Re: 6 GHz stack machineRick C
| |   | `* Re: 6 GHz stack machineWayne morellini
| |   |  `* Re: 6 GHz stack machineRick C
| |   |   `* Re: 6 GHz stack machineWayne morellini
| |   |    `* Re: 6 GHz stack machineRick C
| |   |     +* Re: 6 GHz stack machinePaul Rubin
| |   |     |`- Re: 6 GHz stack machineRick C
| |   |     `* Re: 6 GHz stack machineWayne morellini
| |   |      `* Re: 6 GHz stack machineRick C
| |   |       `* Re: 6 GHz stack machineWayne morellini
| |   |        `* Re: 6 GHz stack machineRick C
| |   |         `* Re: 6 GHz stack machineWayne morellini
| |   |          `* Re: 6 GHz stack machineRick C
| |   |           +* Re: 6 GHz stack machineWayne morellini
| |   |           |`* Re: 6 GHz stack machinedxforth
| |   |           | +* Re: 6 GHz stack machineJurgen Pitaske
| |   |           | |`- Re: 6 GHz stack machineRick C
| |   |           | +- Re: 6 GHz stack machineWayne morellini
| |   |           | +- Re: 6 GHz stack machineWayne morellini
| |   |           | `- Re: 6 GHz stack machineWayne morellini
| |   |           +* Re: 6 GHz stack machineRick C
| |   |           |`* Re: 6 GHz stack machinePaul Rubin
| |   |           | +* Re: 6 GHz stack machineNickolay Kolchin
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +* Re: 6 GHz stack machineRick C
| |   |           | +* Re: 6 GHz stack machineNickolay Kolchin
| |   |           | +- Re: 6 GHz stack machineHugh Aguilar
| |   |           | +- Re: 6 GHz stack machineWayne morellini
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +- Re: 6 GHz stack machineNickolay Kolchin
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +- Re: 6 GHz stack machineNickolay Kolchin
| |   |           | +* Re: 6 GHz stack machineAnton Ertl
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +- Re: 6 GHz stack machineNickolay Kolchin
| |   |           | +- Re: 6 GHz stack machineHugh Aguilar
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +* Re: 6 GHz stack machineHowerd Oakford
| |   |           | +- Re: 6 GHz stack machineWayne morellini
| |   |           | +- Re: 6 GHz stack machineJurgen Pitaske
| |   |           | +- Re: 6 GHz stack machineHugh Aguilar
| |   |           | +- Re: 6 GHz stack machineJurgen Pitaske
| |   |           | +- Re: 6 GHz stack machineminf...@arcor.de
| |   |           | +- Re: 6 GHz stack machineWayne morellini
| |   |           | +- Re: 6 GHz stack machineWayne morellini
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +- Re: 6 GHz stack machineWayne morellini
| |   |           | +- Re: 6 GHz stack machineJames Brakefield
| |   |           | +- Re: 6 GHz stack machineRick C
| |   |           | +* Re: 6 GHz stack machineWayne morellini
| |   |           | `- Re: 6 GHz stack machineWayne morellini
| |   |           `* Re: 6 GHz stack machineWayne morellini
| |   +* Re: 6 GHz stack machinePaul Rubin
| |   `* Re: 6 GHz stack machineBrian Fox
| `* Re: 6 GHz stack machineAndy Valencia
+* Re: 6 GHz stack machineWayne morellini
`* Re: 6 GHz stack machine

Pages:123456789
Re: 6 GHz stack machine

<35e98296-578a-421b-80ce-0e34e95f0a47n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=13863&group=comp.lang.forth#13863

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a05:622a:1c4:: with SMTP id t4mr3493503qtw.140.1625509434095;
Mon, 05 Jul 2021 11:23:54 -0700 (PDT)
X-Received: by 2002:ac8:7d4c:: with SMTP id h12mr14040021qtb.322.1625509433842;
Mon, 05 Jul 2021 11:23:53 -0700 (PDT)
Path: i2pn2.org!i2pn.org!paganini.bofh.team!usenet.pasdenom.info!usenet-fr.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Mon, 5 Jul 2021 11:23:53 -0700 (PDT)
In-Reply-To: <87k0m59weo.fsf@nightsong.com>
Injection-Info: google-groups.googlegroups.com; posting-host=79.172.109.247; posting-account=1qVL5woAAABqIkvU7Nz8Hn6p94_dJQrA
NNTP-Posting-Host: 79.172.109.247
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<875yxqaqlz.fsf@nightsong.com> <bce49b56-45ec-4565-95ce-b29216387487n@googlegroups.com>
<87wnq5amkl.fsf@nightsong.com> <e36b0cf3-0851-42ac-a971-dbff3d3fce82n@googlegroups.com>
<87k0m59weo.fsf@nightsong.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <35e98296-578a-421b-80ce-0e34e95f0a47n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: ilya74.t...@gmail.com (Ilya Tarasov)
Injection-Date: Mon, 05 Jul 2021 18:23:54 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Ilya Tarasov - Mon, 5 Jul 2021 18:23 UTC

понедельник, 5 июля 2021 г. в 08:12:18 UTC+3, Paul Rubin:
> Ilya Tarasov <ilya74....@gmail.com> writes:
> > OMG, you are completely missing this. EDA tools for silicon are not
> > masterpiece generators, it looks like MS Paint with basic set of features,
> > and you need to create something really smart.
> Ok, that's interesting to hear, I was under the impression that you
> write some HDL and the EDA and the fab house take care of (most of the rest).

I didnt write EDA tools. Even using tools is a kind of art. Industry leading tools
like Cadence or Synopsis are a large set of utilities with complex design flow.
There is no 'pushbutton flow'.

> Here is a good tutorial on SAT and SMT solvers:

Looks like a single screw for building a car.
> https://yurichev.com/writings/SAT_SMT_by_example.pdf
> > If an idea is 'Intel can do it, so we can do it too, especially
> > because we are touched by the Forth spirit', you will definitely fail.
> I think the OKAD approach mirrors other stuff written in the 1980's
> after Mead & Conway's book "Introduction to VLSI Design" shook things up
> a lot. You'd lay out rectangles on a screen and iirc, if a red wire
> crossed a green wire, the intersection was a transitor. People did
> design chips using those methods, including the GA144, but it was a lot
> of work and probably becomes unmanageable for much more complex chips.

OKAD is a good example of how wrong and naiv people may be. VLSI design is
not a laying rectangles.
There is a story about a worker who wanted to assemble a television set.
He annoyed engineers by asking them for a schematic for a specific CRT.
They gave him such a scheme in the end. When after a while they asked
him ironically how he was doing with the TV, he brought them home and
they were shocked. The entire wall was covered with a sheet of plywood,
on which parts were nailed, assembled exactly as shown in the diagram.
This TV worked! OKAD looks the same. Printed circuit boards differ
significantly from circuits, and silicon dies differ significantly from circuits
and HDLs. There are so many design rules that are collected in huge
technology files. These rules are validated in CAD, and it is a difficult
skill to understand which rules and how tightly should be set in a project.

Re: 6 GHz stack machine

<87r1gct0f2.fsf@nightsong.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=13866&group=comp.lang.forth#13866

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: no.em...@nospam.invalid (Paul Rubin)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Mon, 05 Jul 2021 23:35:13 -0700
Organization: A noiseless patient Spider
Lines: 15
Message-ID: <87r1gct0f2.fsf@nightsong.com>
References: <sbmuh0$l0m$1@dont-email.me>
<544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<875yxqaqlz.fsf@nightsong.com>
<bce49b56-45ec-4565-95ce-b29216387487n@googlegroups.com>
<87wnq5amkl.fsf@nightsong.com>
<e36b0cf3-0851-42ac-a971-dbff3d3fce82n@googlegroups.com>
<87k0m59weo.fsf@nightsong.com>
<35e98296-578a-421b-80ce-0e34e95f0a47n@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: reader02.eternal-september.org; posting-host="c179265554a4dfe6605bd318c736ae5c";
logging-data="9350"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1/JMrKL4HHgs4tlLIlYCpMi"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
Cancel-Lock: sha1:47VEBYAkyIBQ5X0nFfbhaL6nVcQ=
sha1:K2qLWCNuEy8rXPgDcImCeoivK9A=
 by: Paul Rubin - Tue, 6 Jul 2021 06:35 UTC

Ilya Tarasov <ilya74.tarasov@gmail.com> writes:
> There are so many design rules that are collected in huge
> technology files. These rules are validated in CAD, and it is a difficult
> skill to understand which rules and how tightly should be set in a project.

Design rules in the "old days" were pretty simple: X unit line width, Y
units between lines, Z extra space around any via. There may have been
one or two other parameters but DRC was pretty simple. There was a
timing simulator that worked by estimating capacitance of rectangles.
People did make working digital chips this way. SPICE was more for
analog chips and if you were after higher performance than you could get
with the simple stuff.

It's interesting to hear that the workflow from HDL to finished chips is
not as simple as I'd imagined. Oh well.

Re: 6 GHz stack machine

<454a39b0-c7dc-4031-9e24-dfa14bab02f4n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=13867&group=comp.lang.forth#13867

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:ac8:6ec4:: with SMTP id f4mr15993637qtv.133.1625555631349; Tue, 06 Jul 2021 00:13:51 -0700 (PDT)
X-Received: by 2002:ac8:4e42:: with SMTP id e2mr16087246qtw.311.1625555631124; Tue, 06 Jul 2021 00:13:51 -0700 (PDT)
Path: i2pn2.org!i2pn.org!aioe.org!feeder1.feed.usenet.farm!feed.usenet.farm!tr1.eu1.usenetexpress.com!feeder.usenetexpress.com!tr2.iad1.usenetexpress.com!border1.nntp.dca1.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Tue, 6 Jul 2021 00:13:50 -0700 (PDT)
In-Reply-To: <87r1gct0f2.fsf@nightsong.com>
Injection-Info: google-groups.googlegroups.com; posting-host=85.210.153.81; posting-account=eAOrwQkAAABheFES5y-02sBOFdTlBRio
NNTP-Posting-Host: 85.210.153.81
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com> <875yxqaqlz.fsf@nightsong.com> <bce49b56-45ec-4565-95ce-b29216387487n@googlegroups.com> <87wnq5amkl.fsf@nightsong.com> <e36b0cf3-0851-42ac-a971-dbff3d3fce82n@googlegroups.com> <87k0m59weo.fsf@nightsong.com> <35e98296-578a-421b-80ce-0e34e95f0a47n@googlegroups.com> <87r1gct0f2.fsf@nightsong.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <454a39b0-c7dc-4031-9e24-dfa14bab02f4n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: jpita...@gmail.com (Jurgen Pitaske)
Injection-Date: Tue, 06 Jul 2021 07:13:51 +0000
Content-Type: text/plain; charset="UTF-8"
Lines: 25
 by: Jurgen Pitaske - Tue, 6 Jul 2021 07:13 UTC

On Tuesday, 6 July 2021 at 07:35:16 UTC+1, Paul Rubin wrote:
> Ilya Tarasov <ilya74....@gmail.com> writes:
> > There are so many design rules that are collected in huge
> > technology files. These rules are validated in CAD, and it is a difficult
> > skill to understand which rules and how tightly should be set in a project.
> Design rules in the "old days" were pretty simple: X unit line width, Y
> units between lines, Z extra space around any via. There may have been
> one or two other parameters but DRC was pretty simple. There was a
> timing simulator that worked by estimating capacitance of rectangles.
> People did make working digital chips this way. SPICE was more for
> analog chips and if you were after higher performance than you could get
> with the simple stuff.
>
> It's interesting to hear that the workflow from HDL to finished chips is
> not as simple as I'd imagined. Oh well.

Just for people who might have the time and look for some background:
One extreme - all of the aspects involved in ASICs and the complexity:
https://anysilicon.com/
And the other extreme, probably the closest there is to Chuck's way at the time
https://www.asic-gmbh.de/array_engl.html
Here you can even do Mixed-Signal ASIC prototyping using a breadboard
https://www.asic-gmbh.de/breadboard_engl.html
I remember the ASIC they designed with EEPROM at the time
- no microprocessor yet on-board - used by a model train company.
Manufactured by Hughes microelectronics

Re: 6 GHz stack machine

<9d7666e5-4263-4710-ad79-4a1536f2f0d6n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=13869&group=comp.lang.forth#13869

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:ad4:48c4:: with SMTP id v4mr17416051qvx.30.1625559118602;
Tue, 06 Jul 2021 01:11:58 -0700 (PDT)
X-Received: by 2002:a05:620a:643:: with SMTP id a3mr771001qka.128.1625559118383;
Tue, 06 Jul 2021 01:11:58 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Tue, 6 Jul 2021 01:11:58 -0700 (PDT)
In-Reply-To: <87r1gct0f2.fsf@nightsong.com>
Injection-Info: google-groups.googlegroups.com; posting-host=79.172.111.97; posting-account=1qVL5woAAABqIkvU7Nz8Hn6p94_dJQrA
NNTP-Posting-Host: 79.172.111.97
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<875yxqaqlz.fsf@nightsong.com> <bce49b56-45ec-4565-95ce-b29216387487n@googlegroups.com>
<87wnq5amkl.fsf@nightsong.com> <e36b0cf3-0851-42ac-a971-dbff3d3fce82n@googlegroups.com>
<87k0m59weo.fsf@nightsong.com> <35e98296-578a-421b-80ce-0e34e95f0a47n@googlegroups.com>
<87r1gct0f2.fsf@nightsong.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <9d7666e5-4263-4710-ad79-4a1536f2f0d6n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: ilya74.t...@gmail.com (Ilya Tarasov)
Injection-Date: Tue, 06 Jul 2021 08:11:58 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Ilya Tarasov - Tue, 6 Jul 2021 08:11 UTC

вторник, 6 июля 2021 г. в 09:35:16 UTC+3, Paul Rubin:
> Ilya Tarasov <ilya74....@gmail.com> writes:
> > There are so many design rules that are collected in huge
> > technology files. These rules are validated in CAD, and it is a difficult
> > skill to understand which rules and how tightly should be set in a project.
> Design rules in the "old days" were pretty simple: X unit line width, Y
> units between lines, Z extra space around any via. There may have been
> one or two other parameters but DRC was pretty simple. There was a
> timing simulator that worked by estimating capacitance of rectangles.
> People did make working digital chips this way. SPICE was more for
> analog chips and if you were after higher performance than you could get
> with the simple stuff.

It works up to 250 or 180 nm, maybe. Just like breadboard and wires can be
used for 8 MHz MCU in a DIP package. High-frequency effects will add
problems for 100 MHz clock, and will prevent normal working of DDR3
memory - multilayer PCB wit controlled impedance, differential routing,
shielding, decoupling capacitors etc is strictly required.

For ASIC, at least two major technology shifts took place. First was at 130-90
nm, when a wire delay became equal/greater than a gate delay. Tricks with
short spikes, asynchronous schemes, and many other things are gone.
Design must be fully synchronous, clocked by carefully built clock tree.
Second is near 28 nm, with several non-obvious problems. Clock tree is no
longer able to cover an entire chip area - welcome to Globally Asyncronous,
Locally Synchronous architectures (GALS). Variations (process, voltage,
temperature) are huge, so timing analysis is very complex. For ASICs, local
overheating is a little bonus - is it not enough to place and route, active gates
may be so dense, so temperature may raise too high in a certain small area.
Oh, above 1 GHz a pack of routing tricks are also needed. For example, a net
may require differential and co-planar routing with shielding around curves..
This is a black magic known by people who are deeply inside physical design
process - I cannot provide a complete list of what is needed.
> It's interesting to hear that the workflow from HDL to finished chips is
> not as simple as I'd imagined. Oh well.

Even for FPGA, synthesis and implementation are separated in the CAD
workflow. For ASICs, implementation is a complete different from a
synthesis, heavily dependent on a factory and technology process.

Re: 6 GHz stack machine

<153765a2-a04f-42a5-9a60-5ccd36073090n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=13870&group=comp.lang.forth#13870

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a0c:9ac7:: with SMTP id k7mr17238335qvf.49.1625560937059;
Tue, 06 Jul 2021 01:42:17 -0700 (PDT)
X-Received: by 2002:ac8:67d5:: with SMTP id r21mr16388284qtp.92.1625560936875;
Tue, 06 Jul 2021 01:42:16 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Tue, 6 Jul 2021 01:42:16 -0700 (PDT)
In-Reply-To: <865481aa-00de-48a5-8315-d5c8591fdfebn@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=79.172.111.97; posting-account=1qVL5woAAABqIkvU7Nz8Hn6p94_dJQrA
NNTP-Posting-Host: 79.172.111.97
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<303ae4aa-865c-4176-911f-1aafa102cf34n@googlegroups.com> <d60799fd-ee64-407d-a62f-7bb85088ee91n@googlegroups.com>
<28c78743-6939-493b-9dff-b28d5d2c9fb1n@googlegroups.com> <865481aa-00de-48a5-8315-d5c8591fdfebn@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <153765a2-a04f-42a5-9a60-5ccd36073090n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: ilya74.t...@gmail.com (Ilya Tarasov)
Injection-Date: Tue, 06 Jul 2021 08:42:17 +0000
Content-Type: text/plain; charset="UTF-8"
 by: Ilya Tarasov - Tue, 6 Jul 2021 08:42 UTC

> Good analogy, Ilya!
> I have found, when trying to discuss MFX and the out-of-ordering of the instructions,
> that people will say this is easy, and they read a magazine article about it, etc..
> Maybe easily done by Chuck Norris! lol In practice, not done by anybody other than me.

There are a lot of things which cannot be done from a first attempt. Yes, out-of-order
is not an easy thing, and certainly require some exercices to get at least experience.
It's funny many hobbyists expect that any problem will be solved by itself, simply
because the Forth will be used.

> Juergen Pintaske disgusts me.
> He wants the Forth community to cross their fingers for Stephen Pelc. lol
> Shall we also bring Stephen coffee, shine his shoes, and sharpen his pencils?
> How about a free blow-job? The Forth community is queuing up for the privilege!

'The larger the cabinet, the louder it falls' :)
I saw SP-Forth community fall and TechnoForth fall. Ok, SP-Forth is still exist and
available to download, but very far from success. Both teams wanted to be leaders
of community and were catched by a kind of a deadlock. They started to waiting for
followers who must write applications, but their fans are started to waiting bright
news about their success. No serious activity as a result. I told to Andrey Cherezov
he should think about strategic plans, but he was blinded by the temporary success
of SP-Forth. His arguments was very like to what I can see here. Ok when... ;)

Re: 6 GHz stack machine

<7ce4c32e-3181-419a-a524-e515cc91c088n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=13880&group=comp.lang.forth#13880

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a0c:be85:: with SMTP id n5mr31191451qvi.59.1625765505037;
Thu, 08 Jul 2021 10:31:45 -0700 (PDT)
X-Received: by 2002:a37:b42:: with SMTP id 63mr33458040qkl.325.1625765504852;
Thu, 08 Jul 2021 10:31:44 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Thu, 8 Jul 2021 10:31:44 -0700 (PDT)
In-Reply-To: <153765a2-a04f-42a5-9a60-5ccd36073090n@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=49.197.253.49; posting-account=-BkW7AoAAAAPYBmWFpknE3U_nEqEd6J7
NNTP-Posting-Host: 49.197.253.49
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<303ae4aa-865c-4176-911f-1aafa102cf34n@googlegroups.com> <d60799fd-ee64-407d-a62f-7bb85088ee91n@googlegroups.com>
<28c78743-6939-493b-9dff-b28d5d2c9fb1n@googlegroups.com> <865481aa-00de-48a5-8315-d5c8591fdfebn@googlegroups.com>
<153765a2-a04f-42a5-9a60-5ccd36073090n@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <7ce4c32e-3181-419a-a524-e515cc91c088n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: fourthyf...@gmail.com (Fourthy Forth)
Injection-Date: Thu, 08 Jul 2021 17:31:45 +0000
Content-Type: text/plain; charset="UTF-8"
 by: Fourthy Forth - Thu, 8 Jul 2021 17:31 UTC

On Tuesday, 6 July 2021 at 6:42:17 pm UTC+10, Ilya Tarasov wrote:
> > Good analogy, Ilya!
> > I have found, when trying to discuss MFX and the out-of-ordering of the instructions,
> > that people will say this is easy, and they read a magazine article about it, etc..
> > Maybe easily done by Chuck Norris! lol In practice, not done by anybody other than me.
> There are a lot of things which cannot be done from a first attempt. Yes, out-of-order
> is not an easy thing, and certainly require some exercices to get at least experience.
> It's funny many hobbyists expect that any problem will be solved by itself, simply
> because the Forth will be used.
> > Juergen Pintaske disgusts me.
> > He wants the Forth community to cross their fingers for Stephen Pelc. lol
> > Shall we also bring Stephen coffee, shine his shoes, and sharpen his pencils?
> > How about a free blow-job? The Forth community is queuing up for the privilege!
> 'The larger the cabinet, the louder it falls' :)
> I saw SP-Forth community fall and TechnoForth fall. Ok, SP-Forth is still exist and
> available to download, but very far from success. Both teams wanted to be leaders
> of community and were catched by a kind of a deadlock. They started to waiting for
> followers who must write applications, but their fans are started to waiting bright
> news about their success. No serious activity as a result. I told to Andrey Cherezov
> he should think about strategic plans, but he was blinded by the temporary success
> of SP-Forth. His arguments was very like to what I can see here. Ok when... ;)

Maybe somebody not know how to design a forth procese.

Regular material, without deviation, always have statistically same action of action and interference. Not E=mc2, but tell a lot..

Re: 6 GHz stack machine

<7412b62a-f862-4456-ac79-300d18e01ab4n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=13881&group=comp.lang.forth#13881

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a0c:ff01:: with SMTP id w1mr2035579qvt.28.1625790991196; Thu, 08 Jul 2021 17:36:31 -0700 (PDT)
X-Received: by 2002:a05:620a:13ec:: with SMTP id h12mr18471265qkl.217.1625790991014; Thu, 08 Jul 2021 17:36:31 -0700 (PDT)
Path: i2pn2.org!i2pn.org!aioe.org!feeder1.feed.usenet.farm!feed.usenet.farm!tr2.eu1.usenetexpress.com!feeder.usenetexpress.com!tr2.iad1.usenetexpress.com!border1.nntp.dca1.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Thu, 8 Jul 2021 17:36:30 -0700 (PDT)
In-Reply-To: <153765a2-a04f-42a5-9a60-5ccd36073090n@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=148.167.132.245; posting-account=OxDKOgoAAADW0cxAqHqpN1zqeCoSsDap
NNTP-Posting-Host: 148.167.132.245
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com> <303ae4aa-865c-4176-911f-1aafa102cf34n@googlegroups.com> <d60799fd-ee64-407d-a62f-7bb85088ee91n@googlegroups.com> <28c78743-6939-493b-9dff-b28d5d2c9fb1n@googlegroups.com> <865481aa-00de-48a5-8315-d5c8591fdfebn@googlegroups.com> <153765a2-a04f-42a5-9a60-5ccd36073090n@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <7412b62a-f862-4456-ac79-300d18e01ab4n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: hughagui...@gmail.com (Hugh Aguilar)
Injection-Date: Fri, 09 Jul 2021 00:36:31 +0000
Content-Type: text/plain; charset="UTF-8"
Lines: 48
 by: Hugh Aguilar - Fri, 9 Jul 2021 00:36 UTC

On Tuesday, July 6, 2021 at 1:42:17 AM UTC-7, Ilya Tarasov wrote:
> > I have found, when trying to discuss MFX and the out-of-ordering of the instructions,
> > that people will say this is easy, and they read a magazine article about it, etc..
> > Maybe easily done by Chuck Norris! lol In practice, not done by anybody other than me.
> There are a lot of things which cannot be done from a first attempt. Yes, out-of-order
> is not an easy thing, and certainly require some exercices to get at least experience.

Well, I solved the out-of-ordering of the instructions for the MiniForth on the first attempt.
I had a solution that worked, anyway.
Steve Brault complained that in some cases my assembler's solution was not optimal.
It was possible to hand-code machine-language that did the same thing but was more
efficient in the sense of having fewer NOP instruction inserted.
I said that it was necessary for the assembly-language programmer to help the assembler
by writing his code in a "riscified" manner. Steve Brault complained that this was not
documented anywhere and he did not know what I meant.
I said that this was very similar to what Michael Abrash described in:
"Zen of Code Optimization" that covered the Pentium with its U and V pipes.
The idea is simple. You hold values in registers for as long as possible. You don't load a
register and then immediately use the register. You load the register, you do something
unrelated, then you use the register --- the idea is that the unrelated code will parallelize
with loading the register and/or with using the register. I never heard any further complaints
from Steve Brault, so I assume he understood what I was telling him --- I never saw any of
his assembly-language code though (except the function that did 16-bit integer addition),
so I don't know what quality level he was achieving. I never saw any of the motion-control
code written in MFX because the motion-control program was proprietary to Testra.
John and Tom Hart were very afraid that I would steal it and go start my own company
selling motion-control boards in competition with Testra. That was paranoia. I wasn't
going to do that, and this was way beyond my ability anyway.

> It's funny many hobbyists expect that any problem will be solved by itself, simply
> because the Forth will be used.

Programmers tend to be overly focused on the programming language, and not
focused enough on algorithms --- but algorithms can be ported between languages.

Telling people about writing MFX doesn't impress them. They always tell me that
they don't use Forth so all of this is irrelevant --- they are C programmers.
In actuality, a VLIW processor could be built to run C code, and my assembler
ideas would transfer over to it smoothly. Nobody ever builds VLIW processors though.
You are the only person I am acquainted with who knows what a VLIW processor is.
Quite a lot of people use the term VLIW as a synonym for "super-duper."

We had a thread with this hilarious title: "Zero Instruction Computing?"
https://groups.google.com/g/comp.lang.forth/c/dPvjIMFtRVA/m/PEPLhCtvBgAJ

On Sunday, December 27, 2020 at 11:59:07 PM UTC-7, gnuarm.del...@gmail.com wrote:
> It is a VLIW format with individual control points.

Whatever!

Re: 6 GHz stack machine

<2c325066-827c-4b47-9995-b54fd31ee0b9n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14112&group=comp.lang.forth#14112

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a37:b6c5:: with SMTP id g188mr2955947qkf.92.1628140630592;
Wed, 04 Aug 2021 22:17:10 -0700 (PDT)
X-Received: by 2002:ac8:1417:: with SMTP id k23mr2874993qtj.311.1628140630439;
Wed, 04 Aug 2021 22:17:10 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!news.mixmin.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Wed, 4 Aug 2021 22:17:10 -0700 (PDT)
In-Reply-To: <sbt5dt$goo$1@dont-email.me>
Injection-Info: google-groups.googlegroups.com; posting-host=49.182.101.199; posting-account=-BkW7AoAAAAPYBmWFpknE3U_nEqEd6J7
NNTP-Posting-Host: 49.182.101.199
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <2c325066-827c-4b47-9995-b54fd31ee0b9n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: fourthyf...@gmail.com (Fourthy Forth)
Injection-Date: Thu, 05 Aug 2021 05:17:10 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Fourthy Forth - Thu, 5 Aug 2021 05:17 UTC

On Monday, 5 July 2021 at 6:24:31 am UTC+10, Stephen Pelc wrote:
> On 4 Jul 2021 at 17:00:50 CEST, "Ilya Tarasov" <ilya74....@gmail.com>
> wrote:
> > пятница, 2 июля 2021 г. в 14:49:54 UTC+3, Stephen Pelc:
> doing some tool-making for them. They have completed a fair number of chips.
...
> it's a 32 bit integer CPU with floating point and custom instructions. The
> custom instructions help with the goal.
>
> Stephen

Is this GA company. Done number chips and advanced 32bit list. Will run around in circles streaming, or from memory plus stream?

Re: 6 GHz stack machine

<a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14113&group=comp.lang.forth#14113

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a37:9445:: with SMTP id w66mr3187480qkd.410.1628142401552; Wed, 04 Aug 2021 22:46:41 -0700 (PDT)
X-Received: by 2002:a05:620a:7e4:: with SMTP id k4mr3137625qkk.328.1628142401387; Wed, 04 Aug 2021 22:46:41 -0700 (PDT)
Path: i2pn2.org!i2pn.org!aioe.org!news.uzoreto.com!tr1.eu1.usenetexpress.com!feeder.usenetexpress.com!tr2.iad1.usenetexpress.com!border1.nntp.dca1.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Wed, 4 Aug 2021 22:46:41 -0700 (PDT)
In-Reply-To: <sbt5dt$goo$1@dont-email.me>
Injection-Info: google-groups.googlegroups.com; posting-host=2600:8800:1600:ae2:e1f8:84c0:3a7e:4e9e; posting-account=6GCGIQoAAAAyO8IjR_VTwqiqLwx0Q_G8
NNTP-Posting-Host: 2600:8800:1600:ae2:e1f8:84c0:3a7e:4e9e
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com> <sbt5dt$goo$1@dont-email.me>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: hwfw...@gmail.com (Brad Eckert)
Injection-Date: Thu, 05 Aug 2021 05:46:41 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Lines: 75
 by: Brad Eckert - Thu, 5 Aug 2021 05:46 UTC

On Sunday, July 4, 2021 at 1:24:31 PM UTC-7, Stephen Pelc wrote:
> On 4 Jul 2021 at 17:00:50 CEST, "Ilya Tarasov" <ilya74....@gmail.com>
> wrote:
> > пятница, 2 июля 2021 г. в 14:49:54 UTC+3, Stephen Pelc:
> >> An MPE client is currently designing a new dual-stack machine. The
> >> predicted performance is 6 GHz (instructions per second). 40 CPUs
> >> occupy less than 1 sqare mm.
> >
> > So clickbaiting. 6 GHz meaning 40*150MHz? This is IPS parameter (Instruction
> > per second), not clock speed. If instruction needs more than 1 cycle in
> > average,
> > CPI (cycle per instruction) is needed as well.
> What I can say is limited by what I know and what I am allowed to say. MPE is
> doing some tool-making for them. They have completed a fair number of chips.
>
> No click baiting. Let's assume that 6Hz is a target figure for some unknown
> process
> - i.e. I don't know what it is. From what I do know, I would expect a figure
> in excess
> of 2GHz. That's per CPU. And I know very little about chip and FPGA design.
>
> 150MHz per CPU would happen on an FPGA.
> >> It's for real, and they have a paying client for it.
> >
> > I wonder to see a silicon chip designed without a real goal. Of course, toys or
> > results of author's ambitions may be done just to show it may exist.
> Yes, there's a goal but it's not published yet. For Marcel's benefit,
> it's a 32 bit integer CPU with floating point and custom instructions. The
> custom instructions help with the goal.
>
> Stephen
That is impressive. It must be at a reasonably small process node like 28nm, which is affordable for MPW prototype chips.
At 28nm, you get 8M bits of RAM per square mm. Let's suppose this chip has 5M bits of RAM.
That would be 128K bits, or 4K 32-bit words, per core. Sounds about right.

I watched Chuck Moore's interview where he talked about designing his own computer chips with his his own tools. He looked good.
My eyeroll moment was when he said that he couldn't build large, fast RAMs. Presumably they aren't OKAD-friendly.
Okay, but isn't that the trick? Modern processors are big RAMs with processing logic bolted on here and there.

At today's prices, small companies can build their own even less ambitious Forth chips that make sense at the 130nm to 350nm nodes.
130nm is very popular because the masks can be made by laser instead of much slower e-beam and there's no need for
multi-layer phase shift masks. The wafer costs aren't too bad either.
With the current supply crunch, I suspect more companies are asking "Why are we still buying off-the-shelf MCUs?".

What's more is that Forth is the most elegant way of computing ever invented. It taps into mathematical principles that are only now
being discussed in terms of the mathematics of functional programming. Concatenative lambda calculus supported directly in
hardware is very good. Hardware stacks, also very good. GreenArrays proved that stacks are in fact green.

This will help create more Forth programmers. The good thing about a language ahead of its time is that its time hasn't passed.
Perhaps Forthers treat programming the way the French treat food. Would that make C the equivalent of English cuisine?

Re: 6 GHz stack machine

<b9e2e01b-df19-4679-97ce-f3d19eca6f95n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14115&group=comp.lang.forth#14115

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:ac8:665a:: with SMTP id j26mr3712729qtp.254.1628156418609;
Thu, 05 Aug 2021 02:40:18 -0700 (PDT)
X-Received: by 2002:a37:a00c:: with SMTP id j12mr3838388qke.249.1628156418426;
Thu, 05 Aug 2021 02:40:18 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Thu, 5 Aug 2021 02:40:18 -0700 (PDT)
In-Reply-To: <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=85.210.153.81; posting-account=eAOrwQkAAABheFES5y-02sBOFdTlBRio
NNTP-Posting-Host: 85.210.153.81
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <b9e2e01b-df19-4679-97ce-f3d19eca6f95n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: jpita...@gmail.com (Jurgen Pitaske)
Injection-Date: Thu, 05 Aug 2021 09:40:18 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Jurgen Pitaske - Thu, 5 Aug 2021 09:40 UTC

On Thursday, 5 August 2021 at 06:46:42 UTC+1, Brad Eckert wrote:
> On Sunday, July 4, 2021 at 1:24:31 PM UTC-7, Stephen Pelc wrote:
> > On 4 Jul 2021 at 17:00:50 CEST, "Ilya Tarasov" <ilya74....@gmail.com>
> > wrote:
> > > пятница, 2 июля 2021 г. в 14:49:54 UTC+3, Stephen Pelc:
> > >> An MPE client is currently designing a new dual-stack machine. The
> > >> predicted performance is 6 GHz (instructions per second). 40 CPUs
> > >> occupy less than 1 sqare mm.
> > >
> > > So clickbaiting. 6 GHz meaning 40*150MHz? This is IPS parameter (Instruction
> > > per second), not clock speed. If instruction needs more than 1 cycle in
> > > average,
> > > CPI (cycle per instruction) is needed as well.
> > What I can say is limited by what I know and what I am allowed to say. MPE is
> > doing some tool-making for them. They have completed a fair number of chips.
> >
> > No click baiting. Let's assume that 6Hz is a target figure for some unknown
> > process
> > - i.e. I don't know what it is. From what I do know, I would expect a figure
> > in excess
> > of 2GHz. That's per CPU. And I know very little about chip and FPGA design.
> >
> > 150MHz per CPU would happen on an FPGA.
> > >> It's for real, and they have a paying client for it.
> > >
> > > I wonder to see a silicon chip designed without a real goal. Of course, toys or
> > > results of author's ambitions may be done just to show it may exist.
> > Yes, there's a goal but it's not published yet. For Marcel's benefit,
> > it's a 32 bit integer CPU with floating point and custom instructions. The
> > custom instructions help with the goal.
> >
> > Stephen
> That is impressive. It must be at a reasonably small process node like 28nm, which is affordable for MPW prototype chips.
> At 28nm, you get 8M bits of RAM per square mm. Let's suppose this chip has 5M bits of RAM.
> That would be 128K bits, or 4K 32-bit words, per core. Sounds about right..
>
> I watched Chuck Moore's interview where he talked about designing his own computer chips with his his own tools. He looked good.
> My eyeroll moment was when he said that he couldn't build large, fast RAMs. Presumably they aren't OKAD-friendly.
> Okay, but isn't that the trick? Modern processors are big RAMs with processing logic bolted on here and there.
>
> At today's prices, small companies can build their own even less ambitious Forth chips that make sense at the 130nm to 350nm nodes.
> 130nm is very popular because the masks can be made by laser instead of much slower e-beam and there's no need for
> multi-layer phase shift masks. The wafer costs aren't too bad either.
> With the current supply crunch, I suspect more companies are asking "Why are we still buying off-the-shelf MCUs?".
>
> What's more is that Forth is the most elegant way of computing ever invented. It taps into mathematical principles that are only now
> being discussed in terms of the mathematics of functional programming. Concatenative lambda calculus supported directly in
> hardware is very good. Hardware stacks, also very good. GreenArrays proved that stacks are in fact green.
>
> This will help create more Forth programmers. The good thing about a language ahead of its time is that its time hasn't passed.
> Perhaps Forthers treat programming the way the French treat food. Would that make C the equivalent of English cuisine?

Just had a look at the latest news on theGreenarrays website, copied from there:

Latest developments:

As of Spring 2021, shipments of the EVB002 evaluation kit and of G144A12 chips continue to be made.
The arrayForth 3 integrated development system is in use with no reported problems.
Design of a new chip, G144A2x, continues; this will be upward compatible with the G144A12,
with significant improvements.
Development of Application Notes, including that of a solftware defined GPS receiver, continues.

Has there been a hint somewhere when this new chip will be out? It seems to be shifting.

Re: 6 GHz stack machine

<segddb$eaf$1@gioia.aioe.org>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14116&group=comp.lang.forth#14116

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!aioe.org!7AktqsUqy5CCvnKa3S0Dkw.user.46.165.242.75.POSTED!not-for-mail
From: dxfo...@gmail.com (dxforth)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Thu, 5 Aug 2021 20:09:48 +1000
Organization: Aioe.org NNTP Server
Message-ID: <segddb$eaf$1@gioia.aioe.org>
References: <sbmuh0$l0m$1@dont-email.me>
<544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me>
<a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Info: gioia.aioe.org; logging-data="14671"; posting-host="7AktqsUqy5CCvnKa3S0Dkw.user.gioia.aioe.org"; mail-complaints-to="abuse@aioe.org";
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101
Thunderbird/78.12.0
Content-Language: en-GB
X-Notice: Filtered by postfilter v. 0.9.2
 by: dxforth - Thu, 5 Aug 2021 10:09 UTC

On 5/08/2021 15:46, Brad Eckert wrote:
> ...
> This will help create more Forth programmers. The good thing about a language ahead of its time is that its time hasn't passed.
> Perhaps Forthers treat programming the way the French treat food. Would that make C the equivalent of English cuisine?
>

If Forth is anything like the cheap frozen French import pastries that
turn up on local supermarket shelves, then it suits consumers with more
imagination than taste.

Re: 6 GHz stack machine

<seglca$vq8$1@dont-email.me>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14117&group=comp.lang.forth#14117

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: step...@vfxforth.com (Stephen Pelc)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Thu, 5 Aug 2021 12:25:46 -0000 (UTC)
Organization: A noiseless patient Spider
Lines: 14
Message-ID: <seglca$vq8$1@dont-email.me>
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com> <sbt5dt$goo$1@dont-email.me> <2c325066-827c-4b47-9995-b54fd31ee0b9n@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=fixed
Content-Transfer-Encoding: 8bit
Injection-Date: Thu, 5 Aug 2021 12:25:46 -0000 (UTC)
Injection-Info: reader02.eternal-september.org; posting-host="ae2dac938617445ed3f8bacbb89dd674";
logging-data="32584"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX19ib2j+ndNmvE6S0y70uBTD"
User-Agent: Usenapp/1.10.1/l for MacOS - Full License
Cancel-Lock: sha1:NrzrICUMbZ4/UjP1ZTSxDg1gp7Y=
 by: Stephen Pelc - Thu, 5 Aug 2021 12:25 UTC

On 5 Aug 2021 at 06:17:10 BST, "Fourthy Forth" <fourthyforth@gmail.com> wrote:
>
> Is this GA company.

No.

> Done number chips and advanced 32bit list. Will run around in circles
> streaming,
> or from memory plus stream?

I don't understand what you mean. Programs run from memory with separate
areas for code and data.

Stephen

Re: 6 GHz stack machine

<segljn$30c$1@dont-email.me>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14118&group=comp.lang.forth#14118

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: step...@vfxforth.com (Stephen Pelc)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Thu, 5 Aug 2021 12:29:43 -0000 (UTC)
Organization: A noiseless patient Spider
Lines: 8
Message-ID: <segljn$30c$1@dont-email.me>
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com> <sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=fixed
Content-Transfer-Encoding: 8bit
Injection-Date: Thu, 5 Aug 2021 12:29:43 -0000 (UTC)
Injection-Info: reader02.eternal-september.org; posting-host="ae2dac938617445ed3f8bacbb89dd674";
logging-data="3084"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+M1kl6cs3EMKGBSIs9T6w+"
User-Agent: Usenapp/1.10.1/l for MacOS - Full License
Cancel-Lock: sha1:GRPnJku7U/Muqkrx21mwReNoqoo=
 by: Stephen Pelc - Thu, 5 Aug 2021 12:29 UTC

On 5 Aug 2021 at 06:46:41 BST, "Brad Eckert" <hwfwguy@gmail.com> wrote:
>
> Perhaps Forthers treat programming the way the French treat food. Would that
> make C the equivalent of English cuisine?

Please, USAnian cuisine - lots of fat and sugar.

Stephen

Re: 6 GHz stack machine

<371041eb-f2d7-47ba-958a-3c87760f9348n@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14119&group=comp.lang.forth#14119

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a05:620a:b4f:: with SMTP id x15mr5013072qkg.436.1628172121761; Thu, 05 Aug 2021 07:02:01 -0700 (PDT)
X-Received: by 2002:a05:622a:144:: with SMTP id v4mr4547794qtw.220.1628172121590; Thu, 05 Aug 2021 07:02:01 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!news.uzoreto.com!tr3.eu1.usenetexpress.com!feeder.usenetexpress.com!tr1.iad1.usenetexpress.com!border1.nntp.dca1.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Thu, 5 Aug 2021 07:02:01 -0700 (PDT)
In-Reply-To: <seglca$vq8$1@dont-email.me>
Injection-Info: google-groups.googlegroups.com; posting-host=49.197.87.174; posting-account=-BkW7AoAAAAPYBmWFpknE3U_nEqEd6J7
NNTP-Posting-Host: 49.197.87.174
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com> <sbt5dt$goo$1@dont-email.me> <2c325066-827c-4b47-9995-b54fd31ee0b9n@googlegroups.com> <seglca$vq8$1@dont-email.me>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <371041eb-f2d7-47ba-958a-3c87760f9348n@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: fourthyf...@gmail.com (Fourthy Forth)
Injection-Date: Thu, 05 Aug 2021 14:02:01 +0000
Content-Type: text/plain; charset="UTF-8"
Lines: 17
 by: Fourthy Forth - Thu, 5 Aug 2021 14:02 UTC

On Thursday, 5 August 2021 at 10:25:48 pm UTC+10, Stephen Pelc wrote:
> On 5 Aug 2021 at 06:17:10 BST, "Fourthy Forth" <fourth...@gmail.com> wrote:
> >
> > Is this GA company.
>
> No.

Thank you

> > Done number chips and advanced 32bit list. Will run around in circles
> > streaming,
> > or from memory plus stream?
> I don't understand what you mean. Programs run from memory with separate
> areas for code and data.
>
> Stephen

Parallel serial process, small memory, program hard, vers normal

Re: 6 GHz stack machine

<6879585d-a82f-4577-8914-576bd69b9dbbn@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14127&group=comp.lang.forth#14127

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a05:620a:21d8:: with SMTP id h24mr11797085qka.499.1628279937461;
Fri, 06 Aug 2021 12:58:57 -0700 (PDT)
X-Received: by 2002:ae9:e109:: with SMTP id g9mr12104583qkm.95.1628279937290;
Fri, 06 Aug 2021 12:58:57 -0700 (PDT)
Path: i2pn2.org!i2pn.org!paganini.bofh.team!usenet.pasdenom.info!usenet-fr.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Fri, 6 Aug 2021 12:58:57 -0700 (PDT)
In-Reply-To: <60DF15E4.1070701@rogers.com>
Injection-Info: google-groups.googlegroups.com; posting-host=2600:8800:1600:ae2:541d:ff0e:5a0:20e5;
posting-account=6GCGIQoAAAAyO8IjR_VTwqiqLwx0Q_G8
NNTP-Posting-Host: 2600:8800:1600:ae2:541d:ff0e:5a0:20e5
References: <sbmuh0$l0m$1@dont-email.me> <60DF15E4.1070701@rogers.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <6879585d-a82f-4577-8914-576bd69b9dbbn@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: hwfw...@gmail.com (Brad Eckert)
Injection-Date: Fri, 06 Aug 2021 19:58:57 +0000
Content-Type: text/plain; charset="UTF-8"
 by: Brad Eckert - Fri, 6 Aug 2021 19:58 UTC

On Friday, July 2, 2021 at 6:34:32 AM UTC-7, Brian Fox wrote:
> On 2021-07-02 7:49 AM, Stephen Pelc wrote:
> > An MPE client is currently designing a new dual-stack machine. The
> > predicted performance is 6 GHz (instructions per second). 40 CPUs
> > occupy less than 1 sqare mm.
> >
> > It's for real, and they have a paying client for it.
> >
> > Depending on life, there may be more information at EuroForth 21 in
> > Rome in September. I have my EU Covid passport already.
> >
> > Stephen
> >
> That's really exciting Stephen.
>
> I have always found it tragic that Chuck's CPU ideas didn't find a home
> in the bigger world.
>
I would suppose it uses lessons from shBoom but puts stacks in hardware.
In a modern chip, wire delays trump logic delays.
A hardware stack is a bidirectional shift register where all of the bits are adjacent so there are no long wires.
That allows stacks to run at high speed, much faster than a decently-sized memory.
So, a shBoom type of architecture with 8-bit instructions in a 32-bit group makes sense.
6 GHz instructions means 1.5 GHz memory.

Re: 6 GHz stack machine

<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14241&group=comp.lang.forth#14241

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a37:6447:: with SMTP id y68mr3477739qkb.296.1629203540847;
Tue, 17 Aug 2021 05:32:20 -0700 (PDT)
X-Received: by 2002:a37:506:: with SMTP id 6mr3547803qkf.15.1629203540647;
Tue, 17 Aug 2021 05:32:20 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Tue, 17 Aug 2021 05:32:20 -0700 (PDT)
In-Reply-To: <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=24.50.226.128; posting-account=I-_H_woAAAA9zzro6crtEpUAyIvzd19b
NNTP-Posting-Host: 24.50.226.128
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: gnuarm.d...@gmail.com (Rick C)
Injection-Date: Tue, 17 Aug 2021 12:32:20 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Rick C - Tue, 17 Aug 2021 12:32 UTC

On Thursday, August 5, 2021 at 1:46:42 AM UTC-4, Brad Eckert wrote:
> On Sunday, July 4, 2021 at 1:24:31 PM UTC-7, Stephen Pelc wrote:
> > On 4 Jul 2021 at 17:00:50 CEST, "Ilya Tarasov" <ilya74....@gmail.com>
> > wrote:
> > > пятница, 2 июля 2021 г. в 14:49:54 UTC+3, Stephen Pelc:
> > >> An MPE client is currently designing a new dual-stack machine. The
> > >> predicted performance is 6 GHz (instructions per second). 40 CPUs
> > >> occupy less than 1 sqare mm.
> > >
> > > So clickbaiting. 6 GHz meaning 40*150MHz? This is IPS parameter (Instruction
> > > per second), not clock speed. If instruction needs more than 1 cycle in
> > > average,
> > > CPI (cycle per instruction) is needed as well.
> > What I can say is limited by what I know and what I am allowed to say. MPE is
> > doing some tool-making for them. They have completed a fair number of chips.
> >
> > No click baiting. Let's assume that 6Hz is a target figure for some unknown
> > process
> > - i.e. I don't know what it is. From what I do know, I would expect a figure
> > in excess
> > of 2GHz. That's per CPU. And I know very little about chip and FPGA design.
> >
> > 150MHz per CPU would happen on an FPGA.
> > >> It's for real, and they have a paying client for it.
> > >
> > > I wonder to see a silicon chip designed without a real goal. Of course, toys or
> > > results of author's ambitions may be done just to show it may exist.
> > Yes, there's a goal but it's not published yet. For Marcel's benefit,
> > it's a 32 bit integer CPU with floating point and custom instructions. The
> > custom instructions help with the goal.
> >
> > Stephen
> That is impressive. It must be at a reasonably small process node like 28nm, which is affordable for MPW prototype chips.
> At 28nm, you get 8M bits of RAM per square mm. Let's suppose this chip has 5M bits of RAM.
> That would be 128K bits, or 4K 32-bit words, per core. Sounds about right..

I believe that much static RAM would consume a fair amount of current. In devices like this power consumption is typically an important detail to control. So it would be useful to learn what the device actually is using.

> I watched Chuck Moore's interview where he talked about designing his own computer chips with his his own tools. He looked good.
> My eyeroll moment was when he said that he couldn't build large, fast RAMs. Presumably they aren't OKAD-friendly.
> Okay, but isn't that the trick? Modern processors are big RAMs with processing logic bolted on here and there.

"Modern" processors aren't designed to run Forth apps and vice versa. Isn't that the issue being addressed by such chips?

> At today's prices, small companies can build their own even less ambitious Forth chips that make sense at the 130nm to 350nm nodes.
> 130nm is very popular because the masks can be made by laser instead of much slower e-beam and there's no need for
> multi-layer phase shift masks. The wafer costs aren't too bad either.
> With the current supply crunch, I suspect more companies are asking "Why are we still buying off-the-shelf MCUs?".

ADC, DAC and other analog I/Os as well as the many digital peripherals... that's a big part of the reason. There's a lot of IP amortized in such off the shelf MCUs, not to mention the large code base and tool sets. Even when building a custom chip it is not very common to roll your own CPU to put in it. There has to be a compelling case to support such an investment.

> What's more is that Forth is the most elegant way of computing ever invented. It taps into mathematical principles that are only now
> being discussed in terms of the mathematics of functional programming. Concatenative lambda calculus supported directly in
> hardware is very good. Hardware stacks, also very good. GreenArrays proved that stacks are in fact green.

Not to be snide, but I wasn't aware that GreenArrays proved much of anything with the GA144 except that a CPU chip could be designed that sounded so good and was nearly completely ignored by anyone building products. It is one of the most unrealistic CPUs ever conceived, worse than the RCA 1802 COSMAC. As odd as it was, it has found a home in the space community... somehow.

> This will help create more Forth programmers. The good thing about a language ahead of its time is that its time hasn't passed.
> Perhaps Forthers treat programming the way the French treat food. Would that make C the equivalent of English cuisine?

You can't say the time for Forth has passed. I'm not sure it ever existed.... did it? Certainly it is not yet to come.

--

Rick C.

+ Get 1,000 miles of free Supercharging
+ Tesla referral code - https://ts.la/richard11209

Re: 6 GHz stack machine

<47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14277&group=comp.lang.forth#14277

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a37:a80c:: with SMTP id r12mr10104607qke.299.1629485168975;
Fri, 20 Aug 2021 11:46:08 -0700 (PDT)
X-Received: by 2002:a05:6214:1cb:: with SMTP id c11mr21689983qvt.47.1629485168800;
Fri, 20 Aug 2021 11:46:08 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Fri, 20 Aug 2021 11:46:08 -0700 (PDT)
In-Reply-To: <782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=2600:8800:1600:ae2:e00a:803a:5007:cbb0;
posting-account=6GCGIQoAAAAyO8IjR_VTwqiqLwx0Q_G8
NNTP-Posting-Host: 2600:8800:1600:ae2:e00a:803a:5007:cbb0
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: hwfw...@gmail.com (Brad Eckert)
Injection-Date: Fri, 20 Aug 2021 18:46:08 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Brad Eckert - Fri, 20 Aug 2021 18:46 UTC

On Tuesday, August 17, 2021 at 5:32:21 AM UTC-7, gnuarm.del...@gmail.com wrote:
> I believe that much static RAM would consume a fair amount of current. In devices like this power consumption is typically an important detail to control. So it would be useful to learn what the device actually is using.
That's an interesting thought, but it seems like being performance-oriented they aren't running off batteries. But isn't this a loaded question? Otherwise, why would you start off your response with a non-issue?
> "Modern" processors aren't designed to run Forth apps and vice versa. Isn't that the issue being addressed by such chips?
They are designed to run apps. Apps are CPU coupled to memory. I would imagine "the issue" being addressed has to do with the C programming model.
Maybe they are philosophically opposed to RISC's overhead of nested calls. Of course, saving every return address to the stack isn't cheap. Having stacks in memory, definitely not cheap.

But isn't this a language problem? Now we are penalizing factoring?

> ADC, DAC and other analog I/Os as well as the many digital peripherals... that's a big part of the reason. There's a lot of IP amortized in such off the shelf MCUs, not to mention the large code base and tool sets. Even when building a custom chip it is not very common to roll your own CPU to put in it. There has to be a compelling case to support such an investment.

I would characterize that as an investment in our youth. Of course, there is the sunk cost problem. But, Google (why does it have to be them?) found a way around some of that. eFabless gives your kids (or your inner kid) a playground to just do interesting things in.

> Not to be snide, but I wasn't aware that GreenArrays proved much of anything with the GA144 except that a CPU chip could be designed that sounded so good and was nearly completely ignored by anyone building products. It is one of the most unrealistic CPUs ever conceived, worse than the RCA 1802 COSMAC. As odd as it was, it has found a home in the space community... somehow.

Isn't this about where you are personally? Chuck Moore did in fact greatly enrich GlobalFoundries by fixing their broken fab models when it really mattered. The benefits to humanity of Chuck Moore are profound. That is a life well-lived.

Re: 6 GHz stack machine

<87eeaogd2y.fsf@nightsong.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14278&group=comp.lang.forth#14278

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: no.em...@nospam.invalid (Paul Rubin)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Fri, 20 Aug 2021 11:56:05 -0700
Organization: A noiseless patient Spider
Lines: 8
Message-ID: <87eeaogd2y.fsf@nightsong.com>
References: <sbmuh0$l0m$1@dont-email.me>
<544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me>
<a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com>
<47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: reader02.eternal-september.org; posting-host="da5fc49382a50e5e735fc5692de8af4c";
logging-data="29992"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX18V8GJpdvgXR23XTezwRUPo"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
Cancel-Lock: sha1:xQt/wv03/tb7qtFWidK3DHn+z7Q=
sha1:mdjhruoY0/yhwADF43+zyxmcJW8=
 by: Paul Rubin - Fri, 20 Aug 2021 18:56 UTC

Brad Eckert <hwfwguy@gmail.com> writes:
> Chuck Moore did in fact greatly enrich GlobalFoundries by fixing their
> broken fab models when it really mattered.

Wait, what? Do you mean OUR Chuck Moore? You may be confusing him with
another Chuck Moore:

https://en.wikipedia.org/wiki/Charles_R._Moore_(computer_engineer)

Re: 6 GHz stack machine

<d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14279&group=comp.lang.forth#14279

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a05:622a:138c:: with SMTP id o12mr19698527qtk.346.1629489223622;
Fri, 20 Aug 2021 12:53:43 -0700 (PDT)
X-Received: by 2002:a05:622a:305:: with SMTP id q5mr19467027qtw.154.1629489223419;
Fri, 20 Aug 2021 12:53:43 -0700 (PDT)
Path: i2pn2.org!i2pn.org!paganini.bofh.team!usenet.pasdenom.info!usenet-fr.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Fri, 20 Aug 2021 12:53:43 -0700 (PDT)
In-Reply-To: <87eeaogd2y.fsf@nightsong.com>
Injection-Info: google-groups.googlegroups.com; posting-host=2600:8800:1600:ae2:e00a:803a:5007:cbb0;
posting-account=6GCGIQoAAAAyO8IjR_VTwqiqLwx0Q_G8
NNTP-Posting-Host: 2600:8800:1600:ae2:e00a:803a:5007:cbb0
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com> <47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
<87eeaogd2y.fsf@nightsong.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: hwfw...@gmail.com (Brad Eckert)
Injection-Date: Fri, 20 Aug 2021 19:53:43 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Brad Eckert - Fri, 20 Aug 2021 19:53 UTC

On Friday, August 20, 2021 at 11:56:07 AM UTC-7, Paul Rubin wrote:
> Brad Eckert <hwf...@gmail.com> writes:
> > Chuck Moore did in fact greatly enrich GlobalFoundries by fixing their
> > broken fab models when it really mattered.
> Wait, what? Do you mean OUR Chuck Moore? You may be confusing him with
> another Chuck Moore:
>
> https://en.wikipedia.org/wiki/Charles_R._Moore_(computer_engineer)

Wait, what? There's moore than one?

Yeah, that whole temperature model being important when nobody thought it mattered.
Well of course it matters if scaling matters. So, GlobalFoundries became the guys who could deliver.

Shouldn't we be building Forth systems just because we can? Just to pay tribute to such a remarkable human being as Charles H. Moore?

Forth was always doomed because of the sunk-cost problem. Sorry, no hiding the source in libraries, the source is the library. So yes. Forth was always ahead of its time. It doesn't play in the 3D money game. It's a whole different thing. Oh the times, they are a-changing.

C is a legacy of what? What did the 20th century get you? Bigger KaBooms? How will the industry paradigm overcome the same old problems that can't be ignored anymore? Isn't library code the real problem? Notice how Chuck always challenged his thinking. What a wizard.

The soul of a machine, isn't that the heart of the problem? This separation of the creator from his/her creation for reasons of financial empire. This empire of illusion. Too bad Chuck fell for the illusion, but it gave us Forth. Libraries are not your friend because they are built on this legacy.
No, if you are out to build something to materially benefit all of humanity, isn't Forth really your only sustainable option?

Look at Woody Harrelson as a role model for this. Lives in Maui not far from our own Elizabeth Conklin. Had his wedding bands custom made from gold dust panned from streams in Northern California. Not far from Chuck's old stomping grounds. He could have gotten that gold off the market. But he didn't..

Re: 6 GHz stack machine

<87a6lbhmyr.fsf@nightsong.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14280&group=comp.lang.forth#14280

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: no.em...@nospam.invalid (Paul Rubin)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Fri, 20 Aug 2021 13:37:16 -0700
Organization: A noiseless patient Spider
Lines: 32
Message-ID: <87a6lbhmyr.fsf@nightsong.com>
References: <sbmuh0$l0m$1@dont-email.me>
<544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me>
<a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com>
<47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
<87eeaogd2y.fsf@nightsong.com>
<d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: reader02.eternal-september.org; posting-host="da5fc49382a50e5e735fc5692de8af4c";
logging-data="4872"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX18aJ6opzchqYyAllMvdtsNI"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
Cancel-Lock: sha1:kj15g83UVta8cpsQjGynGkjhWWc=
sha1:zPndp95cSOsgTg29OVLYdK1Lb3s=
 by: Paul Rubin - Fri, 20 Aug 2021 20:37 UTC

Brad Eckert <hwfwguy@gmail.com> writes:
> Well of course it matters if scaling matters. So, GlobalFoundries
> became the guys who could deliver... Shouldn't we be building Forth
> systems just because we can? Just to pay tribute to such a remarkable
> human being as Charles H. Moore?

Your entire post is completely confusing, but particularly: are you
saying that Charles H. Moore (the inventor of Forth) had something to do
with GlobalFoundries? Or are you thinking of Charles R. Moore, a
different person who was an architect at AMD, which later spun off its
fab division to become GlobalFoundries? It sounds more likely to me
that you are thinking of Charles R. Moore, but I can't tell.

> Too bad Chuck fell for the illusion, but it gave us Forth. Libraries
> are not your friend because they are built on this legacy.

I completely don't understand what point you are making about libraries.

> No, if you are out to build something to materially benefit all of
> humanity, isn't Forth really your only sustainable option?

This I don't understand either. Forth is interesting historically and
maybe in the present day, but of course there are other ways to write
software, and even ways to materially benefit humanity without involving
software.

> Look at Woody Harrelson as a role model for this. Lives in Maui not
> far from our own Elizabeth Conklin. Had his wedding bands custom made
> from gold dust panned from streams in Northern California.

That sounds pretty cool. We haven't heard from Elizabeth here lately.
I hope she comes back.

Re: 6 GHz stack machine

<c9a84734-9ac0-4b68-962b-0d213338e9ebn@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14281&group=comp.lang.forth#14281

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a37:a905:: with SMTP id s5mr10947450qke.63.1629498681511;
Fri, 20 Aug 2021 15:31:21 -0700 (PDT)
X-Received: by 2002:a0c:ffae:: with SMTP id d14mr22401395qvv.41.1629498681326;
Fri, 20 Aug 2021 15:31:21 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Fri, 20 Aug 2021 15:31:21 -0700 (PDT)
In-Reply-To: <87a6lbhmyr.fsf@nightsong.com>
Injection-Info: google-groups.googlegroups.com; posting-host=2600:8800:1600:ae2:e00a:803a:5007:cbb0;
posting-account=6GCGIQoAAAAyO8IjR_VTwqiqLwx0Q_G8
NNTP-Posting-Host: 2600:8800:1600:ae2:e00a:803a:5007:cbb0
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com> <47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
<87eeaogd2y.fsf@nightsong.com> <d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com>
<87a6lbhmyr.fsf@nightsong.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <c9a84734-9ac0-4b68-962b-0d213338e9ebn@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: hwfw...@gmail.com (Brad Eckert)
Injection-Date: Fri, 20 Aug 2021 22:31:21 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 by: Brad Eckert - Fri, 20 Aug 2021 22:31 UTC

On Friday, August 20, 2021 at 1:37:18 PM UTC-7, Paul Rubin wrote:
> Brad Eckert <hwf...@gmail.com> writes:
> > Well of course it matters if scaling matters. So, GlobalFoundries
> > became the guys who could deliver... Shouldn't we be building Forth
> > systems just because we can? Just to pay tribute to such a remarkable
> > human being as Charles H. Moore?
> Your entire post is completely confusing, but particularly: are you
> saying that Charles H. Moore (the inventor of Forth) had something to do
> with GlobalFoundries? Or are you thinking of Charles R. Moore, a
> different person who was an architect at AMD, which later spun off its
> fab division to become GlobalFoundries? It sounds more likely to me
> that you are thinking of Charles R. Moore, but I can't tell.
> > Too bad Chuck fell for the illusion, but it gave us Forth. Libraries
> > are not your friend because they are built on this legacy.
> I completely don't understand what point you are making about libraries.
> > No, if you are out to build something to materially benefit all of
> > humanity, isn't Forth really your only sustainable option?
> This I don't understand either. Forth is interesting historically and
> maybe in the present day, but of course there are other ways to write
> software, and even ways to materially benefit humanity without involving
> software.
> > Look at Woody Harrelson as a role model for this. Lives in Maui not
> > far from our own Elizabeth Conklin. Had his wedding bands custom made
> > from gold dust panned from streams in Northern California.
> That sounds pretty cool. We haven't heard from Elizabeth here lately.
> I hope she comes back.

Oh, I thought you were making a joke. My mistake. My recollection of the Charles H. Moore saga involves him trying desperately trying to make his OKAD tool
work with the foundry-supplied transistor models. These models were bad, and he made no secret of his use of GlobalFoundries as his supplier. The resulting corrections to that fab's "secret sauce" would be reasonably assumed to have important downstream ramifications.
Of course, if it hadn't been "our Chuck" it would have been someone with real money so the problem would have been fixed, after the crucial market window.

Other than that, my post is more to clarify my own thoughts on the metaphysics of sustainably sourced computing. Way too Zen. How this plays out in the hardware world is anyone's guess. Google is pioneering a new model that allows software guys to make real chips on real foundries. Yup, there are all kinds of languages and all kinds of tools. Did you write them? Did someone who loves you write them? What were their motivations?

Are those motivations something you want to buy into? Google's FOSS commitment seemed to be a showstopper for the fabs, but they found a way. Their way is something I can buy into.

I own an Apple smartphone and sure wish I didn't. Yup, I think the old duderino is about to go Android. I would love Apple if they weren't so cut-throat gangster. Apple is the new Philip Morris. Not something anyone should buy into, but that's me.

You seem to share the same view of historical Forth as Charles H. Moore. It's an interesting footnote, but it's not going anywhere. Forth is words and stack. What we build around that is up to us. That's why the multicore Forth chip - because they can. C is also an interesting historical footnote. But as you can see, it's still around. Why? Libraries. You know, the things that tie us to our past. The things Chuck tried to warn us about. Isn't the point of computing to move beyond the structures of the past? Isn't that why he left Forth Inc? Everyone wanted to canonize their past work, just like the C guys, instead of leaving Forth open to endless conceptualization. Here's your virtual machine, this is what Forth is. No it isn't, it's only one embodiment of an underlying informatics based on stacks.

I can see how reinventing the wheel doesn't make business sense. But what if reinventing the wheel is the point? That's what makes computer programming a kind of Zen practice. Keep taking away the old, not continue to build upon it. Libraries make us slaves to our past. Standing on the shoulders of giants is only as good as the giants themselves. If you're in for the hero's journey, Forth is probably more your thing. Yoda and floating rocks sold separately.

Re: 6 GHz stack machine

<875yvzh4ut.fsf@nightsong.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14285&group=comp.lang.forth#14285

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: no.em...@nospam.invalid (Paul Rubin)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Fri, 20 Aug 2021 20:08:26 -0700
Organization: A noiseless patient Spider
Lines: 18
Message-ID: <875yvzh4ut.fsf@nightsong.com>
References: <sbmuh0$l0m$1@dont-email.me>
<544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me>
<a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com>
<47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
<87eeaogd2y.fsf@nightsong.com>
<d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com>
<87a6lbhmyr.fsf@nightsong.com>
<c9a84734-9ac0-4b68-962b-0d213338e9ebn@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: reader02.eternal-september.org; posting-host="5c5a6d242f37741257e0717fad018708";
logging-data="2581"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+u5D5Ll+prQPgUrFfaQ+mA"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
Cancel-Lock: sha1:Qim81R2bclQZRJPJu37//Cp9QUE=
sha1:d4sqJnjs1bzXVrIeQ0w9n79daYI=
 by: Paul Rubin - Sat, 21 Aug 2021 03:08 UTC

Brad Eckert <hwfwguy@gmail.com> writes:
> C is also an interesting historical footnote. But as you can see, it's
> still around. Why? Libraries. You know, the things that tie us to our
> past. The things Chuck tried to warn us about.

C is still relevant because of the humongous amount of historical
programs written in it, by which I mean large programs like the Linux
kernel, not libraries in particular. There really isn't that much
library code for C compared with other languages. You have to do more
yourself. C is historical in that relatively few people today, when
they decide to embark on a large new software project, choose to write
it in C. C programming today is either small embedded applications, or
maintenance programming of older projects.

Library ecosystems spring up around new languages relatively quickly:
look at Go, Rust, Ruby, etc., all of which are much newer than C and
have plenty of libraries. This is especially true since you can usually
call C code from other languages.

Re: 6 GHz stack machine

<2021Aug21.094356@mips.complang.tuwien.ac.at>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14290&group=comp.lang.forth#14290

 copy link   Newsgroups: comp.lang.forth
Path: i2pn2.org!i2pn.org!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: ant...@mips.complang.tuwien.ac.at (Anton Ertl)
Newsgroups: comp.lang.forth
Subject: Re: 6 GHz stack machine
Date: Sat, 21 Aug 2021 07:43:56 GMT
Organization: Institut fuer Computersprachen, Technische Universitaet Wien
Lines: 33
Message-ID: <2021Aug21.094356@mips.complang.tuwien.ac.at>
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com> <sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com> <782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com> <47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com> <87eeaogd2y.fsf@nightsong.com> <d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com> <87a6lbhmyr.fsf@nightsong.com>
Injection-Info: reader02.eternal-september.org; posting-host="1a058b2af70ed0f49c1fa9b430f27a44";
logging-data="27238"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1/bGyu1+tQEmyYIU9bC9rpX"
Cancel-Lock: sha1:GOJ5d8DS+Vkgfgpcsz6SFinIrPw=
X-newsreader: xrn 10.00-beta-3
 by: Anton Ertl - Sat, 21 Aug 2021 07:43 UTC

Paul Rubin <no.email@nospam.invalid> writes:
>Your entire post is completely confusing, but particularly: are you
>saying that Charles H. Moore (the inventor of Forth) had something to do
>with GlobalFoundries? Or are you thinking of Charles R. Moore, a
>different person who was an architect at AMD, which later spun off its
>fab division to become GlobalFoundries?

Looking at the wikipedia page of Charles R. Moore, he was a computer
architect, and while that is adjacent to circuit design, I doubt that
he made substantial contributions to circuit design.

Concerning Charles H. Moore, he went from programming down to
programming language design (Forth), to computer architecture (Novix
ff.), and finally to circuit design (OKAD).

However, he has always used outdated processes (and I remember the
name MOSIS, but not Globalfoundries), so it's not very likely that he
finds problems and solutions that make a difference for a lot of other
products.

He also uses the processes in unusual ways: At EuroForth he mentioned
that he had a problem with one particular design rule (basically his
design was not dense enough), and Bernd Paysan commented that usual
circuits don't have a problem with that rule; so it can easily be that
he discovered things about the process that are not in the usual
models, but it's less clear that this discovery helps other circuits.

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: http://www.forth200x.org/forth200x.html
EuroForth 2021: https://euro.theforth.net/2021

Re: 6 GHz stack machine

<3f221fb6-80e5-4e95-9cae-bd4e4abcd97dn@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14291&group=comp.lang.forth#14291

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:a05:6214:570:: with SMTP id cj16mr24308557qvb.41.1629539122532;
Sat, 21 Aug 2021 02:45:22 -0700 (PDT)
X-Received: by 2002:ac8:43d6:: with SMTP id w22mr21289500qtn.92.1629539122305;
Sat, 21 Aug 2021 02:45:22 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Sat, 21 Aug 2021 02:45:22 -0700 (PDT)
In-Reply-To: <2021Aug21.094356@mips.complang.tuwien.ac.at>
Injection-Info: google-groups.googlegroups.com; posting-host=85.210.153.81; posting-account=eAOrwQkAAABheFES5y-02sBOFdTlBRio
NNTP-Posting-Host: 85.210.153.81
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com> <47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
<87eeaogd2y.fsf@nightsong.com> <d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com>
<87a6lbhmyr.fsf@nightsong.com> <2021Aug21.094356@mips.complang.tuwien.ac.at>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <3f221fb6-80e5-4e95-9cae-bd4e4abcd97dn@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: jpita...@gmail.com (Jurgen Pitaske)
Injection-Date: Sat, 21 Aug 2021 09:45:22 +0000
Content-Type: text/plain; charset="UTF-8"
 by: Jurgen Pitaske - Sat, 21 Aug 2021 09:45 UTC

On Saturday, 21 August 2021 at 09:14:25 UTC+1, Anton Ertl wrote:
> Paul Rubin <no.e...@nospam.invalid> writes:
> >Your entire post is completely confusing, but particularly: are you
> >saying that Charles H. Moore (the inventor of Forth) had something to do
> >with GlobalFoundries? Or are you thinking of Charles R. Moore, a
> >different person who was an architect at AMD, which later spun off its
> >fab division to become GlobalFoundries?
> Looking at the wikipedia page of Charles R. Moore, he was a computer
> architect, and while that is adjacent to circuit design, I doubt that
> he made substantial contributions to circuit design.
>
> Concerning Charles H. Moore, he went from programming down to
> programming language design (Forth), to computer architecture (Novix
> ff.), and finally to circuit design (OKAD).
>
> However, he has always used outdated processes (and I remember the
> name MOSIS, but not Globalfoundries), so it's not very likely that he
> finds problems and solutions that make a difference for a lot of other
> products.
>
> He also uses the processes in unusual ways: At EuroForth he mentioned
> that he had a problem with one particular design rule (basically his
> design was not dense enough), and Bernd Paysan commented that usual
> circuits don't have a problem with that rule; so it can easily be that
> he discovered things about the process that are not in the usual
> models, but it's less clear that this discovery helps other circuits.
>
> - anton
> --
> M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
> comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
> New standard: http://www.forth200x.org/forth200x.html
> EuroForth 2021: https://euro.theforth.net/2021

A slight correction:

> However, he has always used outdated processes (and I remember the
> name MOSIS, but not Globalfoundries), so it's not very likely that he
> finds problems and solutions that make a difference for a lot of other
> products.

These were not outdated processes of MOSIS at the time.
Chuck had to use processes that he could afford
and that were available for the way he generated the GDSII.

He had simulated what comes out,
so, the process selected by Chuck was suitable and good enough for his prototypes.

See Wikipedia https://en.wikipedia.org/wiki/MOSIS
It seems that the MOSIS Website is dead now,
So any new Greenarrays chips would need a new set of masks to be paid for at a new suitable fab.

Hopefully, GA still has sufficient wafers and dies to bridge the gap until the new version comes out - whenever this might be.
No dates for availability on the GA website.

Re: 6 GHz stack machine

<2ab78b8c-944b-46a8-822c-e7f88c2ba5fan@googlegroups.com>

 copy mid

https://www.novabbs.com/devel/article-flat.php?id=14292&group=comp.lang.forth#14292

 copy link   Newsgroups: comp.lang.forth
X-Received: by 2002:ad4:57b3:: with SMTP id g19mr24309765qvx.38.1629540211892;
Sat, 21 Aug 2021 03:03:31 -0700 (PDT)
X-Received: by 2002:ac8:4e44:: with SMTP id e4mr21981807qtw.252.1629540211748;
Sat, 21 Aug 2021 03:03:31 -0700 (PDT)
Path: i2pn2.org!i2pn.org!aioe.org!news.uzoreto.com!news-out.netnews.com!news.alt.net!fdc2.netnews.com!peer01.ams1!peer.ams1.xlned.com!news.xlned.com!peer01.iad!feed-me.highwinds-media.com!news.highwinds-media.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.forth
Date: Sat, 21 Aug 2021 03:03:31 -0700 (PDT)
In-Reply-To: <3f221fb6-80e5-4e95-9cae-bd4e4abcd97dn@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=85.210.153.81; posting-account=eAOrwQkAAABheFES5y-02sBOFdTlBRio
NNTP-Posting-Host: 85.210.153.81
References: <sbmuh0$l0m$1@dont-email.me> <544d104c-7315-4184-81d9-eb33ea2fcfb6n@googlegroups.com>
<sbt5dt$goo$1@dont-email.me> <a75d937d-3003-4d0d-8444-684a0b03f1b2n@googlegroups.com>
<782743b2-0f55-41c2-8c35-e91cb13d292en@googlegroups.com> <47830d23-afd5-4b57-a5d8-0b17daa84f9dn@googlegroups.com>
<87eeaogd2y.fsf@nightsong.com> <d7c78aad-5c13-4e4c-8269-9551dcecac9dn@googlegroups.com>
<87a6lbhmyr.fsf@nightsong.com> <2021Aug21.094356@mips.complang.tuwien.ac.at> <3f221fb6-80e5-4e95-9cae-bd4e4abcd97dn@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <2ab78b8c-944b-46a8-822c-e7f88c2ba5fan@googlegroups.com>
Subject: Re: 6 GHz stack machine
From: jpita...@gmail.com (Jurgen Pitaske)
Injection-Date: Sat, 21 Aug 2021 10:03:31 +0000
Content-Type: text/plain; charset="UTF-8"
X-Received-Bytes: 4642
 by: Jurgen Pitaske - Sat, 21 Aug 2021 10:03 UTC

On Saturday, 21 August 2021 at 10:45:23 UTC+1, Jurgen Pitaske wrote:
> On Saturday, 21 August 2021 at 09:14:25 UTC+1, Anton Ertl wrote:
> > Paul Rubin <no.e...@nospam.invalid> writes:
> > >Your entire post is completely confusing, but particularly: are you
> > >saying that Charles H. Moore (the inventor of Forth) had something to do
> > >with GlobalFoundries? Or are you thinking of Charles R. Moore, a
> > >different person who was an architect at AMD, which later spun off its
> > >fab division to become GlobalFoundries?
> > Looking at the wikipedia page of Charles R. Moore, he was a computer
> > architect, and while that is adjacent to circuit design, I doubt that
> > he made substantial contributions to circuit design.
> >
> > Concerning Charles H. Moore, he went from programming down to
> > programming language design (Forth), to computer architecture (Novix
> > ff.), and finally to circuit design (OKAD).
> >
> > However, he has always used outdated processes (and I remember the
> > name MOSIS, but not Globalfoundries), so it's not very likely that he
> > finds problems and solutions that make a difference for a lot of other
> > products.
> >
> > He also uses the processes in unusual ways: At EuroForth he mentioned
> > that he had a problem with one particular design rule (basically his
> > design was not dense enough), and Bernd Paysan commented that usual
> > circuits don't have a problem with that rule; so it can easily be that
> > he discovered things about the process that are not in the usual
> > models, but it's less clear that this discovery helps other circuits.
> >
> > - anton
> > --
> > M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
> > comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
> > New standard: http://www.forth200x.org/forth200x.html
> > EuroForth 2021: https://euro.theforth.net/2021
> A slight correction:
> > However, he has always used outdated processes (and I remember the
> > name MOSIS, but not Globalfoundries), so it's not very likely that he
> > finds problems and solutions that make a difference for a lot of other
> > products.
> These were not outdated processes of MOSIS at the time.
> Chuck had to use processes that he could afford
> and that were available for the way he generated the GDSII.
>
> He had simulated what comes out,
> so, the process selected by Chuck was suitable and good enough for his prototypes.
>
> See Wikipedia https://en.wikipedia.org/wiki/MOSIS
> It seems that the MOSIS Website is dead now,
> So any new Greenarrays chips would need a new set of masks to be paid for at a new suitable fab.
>
> Hopefully, GA still has sufficient wafers and dies to bridge the gap until the new version comes out - whenever this might be.
> No dates for availability on the GA website.

Just to add to the timeline:
https://news.ycombinator.com/item?id=3267428
GA144, available since 2011,
and not many applications known unfortunately over the last 10 years.

Pages:123456789
server_pubkey.txt

rocksolid light 0.9.7
clearnet tor