Re: i860: Re: modern stuff

[Chris Hanson via cctalk]

The i860 seemed to be everywhere in high end graphics for a brief period of 
time; it seems like everyone whose graphics had been several ganged Weitek 
units and their own execution engine to feed them switched to one or more i860 
chips at once. (Wasn’t RealityEngine also i860?)

Did Intel offer some sort of incentive to do so? Was the chip really all that 
for its day, as the contemporary deep dive in BYTE seemed to make it out to be? 
Or was it just an attempt to hedge and use something developers hoped would 
become a commodity with successive backwards-compatible generations like 
Intel’s CPUs?

  — Chris

Sent from my iPad

> On Nov 2, 2018, at 1:53 PM, Marc Howard via cctalk  
> wrote:
> 
> I designed the video section of that board set (VX/MVX).  The VX had an
> i860 + a very large 32 bit frame buffer.  It also had and 2nd 8 bit frame
> buffer based two custom Sun chips that was used for the window system.  The
> video could switch between the two frame buffers on a per-pixel basis.  The
> output format of the larger frame buffer was micro-programmable; some VXs
> were used by Sarnoff Labs in early development of the HDTV standard.
> 
> The MVX had four i860s and a very wide (256 bits?) high speed connection to
> the VX.
> 
> Oh, and the guys that developed the chip set for 2D graphics?  They left
> and founded a little company called Nvidia.  Sigh.
> 
> Marc

Re: i860: Re: modern stuff

[Marc Howard via cctalk]

I designed the video section of that board set (VX/MVX).  The VX had an
i860 + a very large 32 bit frame buffer.  It also had and 2nd 8 bit frame
buffer based two custom Sun chips that was used for the window system.  The
video could switch between the two frame buffers on a per-pixel basis.  The
output format of the larger frame buffer was micro-programmable; some VXs
were used by Sarnoff Labs in early development of the HDTV standard.

The MVX had four i860s and a very wide (256 bits?) high speed connection to
the VX.

Oh, and the guys that developed the chip set for 2D graphics?  They left
and founded a little company called Nvidia.  Sigh.

Marc

On Fri, Nov 2, 2018 at 5:34 AM Michael Thompson via cctalk <
cctalk@classiccmp.org> wrote:

> >
> > Date: Wed, 31 Oct 2018 14:29:18 -0700
> > From: Eric Korpela 
> > Subject: Re: i860: Re: modern stuff
> >
> > A Google search on Skybolt i860 produces interesting results.
> > >Additional realtime signal processing
> > > capability is provided by four Skybolt i860-based VMEbus single-board
> > > computers with 240 MFLOPS peak combined capacity.
> > > --
> > > Remember when 240 MFLOPS was a lot?
> >
>
> That's the board that I have.
>
> Quad i860 on a 9Ux400 VME board.
>
> Its in a Sun 4/280 development system.
>
>
> --
> Michael Thompson
>

Re: i860: Re: modern stuff

[Guy Sotomayor Jr via cctalk]

> On Nov 4, 2018, at 9:37 AM, Todd Goodman via cctalk  
> wrote:
> 
> Yes, a company I was working for OEMed what because IBM's X25Net software and 
> it was ported to their RTIC i960 cards from our own homegrown i960 cards.
> 
> The IBM group we worked with was in La Gaude France but we heard the RTIC 
> cards were developed in Boca Raton, FL.
> 

Yes, the RTIC cards were developed in Boca Raton.  I had a number of friends 
that worked in that group over the years.  They were in a building off of the 
main site that was the last to be closed down when IBM decided to close the 
Boca Raton site down.

TTFN - Guy

Re: i860: Re: modern stuff

[Todd Goodman via cctalk]

Yes, a company I was working for OEMed what because IBM's X25Net 
software and it was ported to their RTIC i960 cards from our own 
homegrown i960 cards.


The IBM group we worked with was in La Gaude France but we heard the 
RTIC cards were developed in Boca Raton, FL.


We ran VxWorks on them.

Todd


On 11/4/2018 10:42 AM, Carlos E Murillo-Sanchez via cctalk wrote:

William Donzelli wrote:

So, what is this i960-based card for?

They were the routers. At the core nodes of the network, there would
be a big RS/6000s (very early POWER1 types) that would each do about
4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
was one of these cards, so each of the big RS/6000s would have about
4-5 of these cards.

IBM tried to commercialize the design, but it was doomed - the routing
engines were very fast, but the internet quickly outgrew the
architecture of the engines, and they apparently needed a complete
redesign to compete. IBM did release very few of these RS/6000s to the
public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
I have only seen one of these routers in the wild, but most of the
real NSFnet ones (I was decommissioning them, one time with a Sawzall
because of some live tangled cables).


Could it be related to what you
say in your post?

https://imgur.com/NIvQPBv

Possibly related, but that card is not one of the NSFnet ones.

--
Will
After searching the web for a while, I finally discovered what this 
is:  the key is that it is a "2-O" mca adapter, and it is a V.35 
communications adapter.  But I also learned that IBM produced a series 
of adapters hosting an i960 consisting of a processor card and a 
daughter card; the daughter card would have the specifics for the kind 
of interface that was implemented (rs232, rs422, X.25, etc).  These 
adapters were called "ARTIC960 coprocessors".  They were first 
produced for microchannel, later for PCI.  You could develop code for 
it in an rs/6000 system, and then load on the adapter and run it:


http://ohlandl.ipv7.net/communications/aa6proggde.pdf

carlos.

Re: i860: Re: modern stuff

[Carlos E Murillo-Sanchez via cctalk]

William Donzelli wrote:

So, what is this i960-based card for?

They were the routers. At the core nodes of the network, there would
be a big RS/6000s (very early POWER1 types) that would each do about
4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
was one of these cards, so each of the big RS/6000s would have about
4-5 of these cards.

IBM tried to commercialize the design, but it was doomed - the routing
engines were very fast, but the internet quickly outgrew the
architecture of the engines, and they apparently needed a complete
redesign to compete. IBM did release very few of these RS/6000s to the
public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
I have only seen one of these routers in the wild, but most of the
real NSFnet ones (I was decommissioning them, one time with a Sawzall
because of some live tangled cables).


Could it be related to what you
say in your post?

https://imgur.com/NIvQPBv

Possibly related, but that card is not one of the NSFnet ones.

--
Will
After searching the web for a while, I finally discovered what this is:  
the key is that it is a "2-O" mca adapter, and it is a V.35 
communications adapter.  But I also learned that IBM produced a series 
of adapters hosting an i960 consisting of a processor card and a 
daughter card; the daughter card would have the specifics for the kind 
of interface that was implemented (rs232, rs422, X.25, etc).  These 
adapters were called "ARTIC960 coprocessors".  They were first produced 
for microchannel, later for PCI.  You could develop code for it in an 
rs/6000 system, and then load on the adapter and run it:


http://ohlandl.ipv7.net/communications/aa6proggde.pdf

carlos.

Re: i860: Re: modern stuff

[Kevin Bowling via cctalk]

That’s interesting stuff, get it done! :)

On Sat, Nov 3, 2018 at 3:01 PM William Donzelli  wrote:

> I have quite  few NSFnet backup tapes that need to go to Al at some
> point. They likely have good stuff on them.
>
> --
> Will
> On Sat, Nov 3, 2018 at 5:03 AM Kevin Bowling 
> wrote:
> >
> > Do you have software or docs for any of this stuff? I run
> ps-2.kev009.com
> >
> > On Thu, Nov 1, 2018 at 3:41 PM William Donzelli via cctalk <
> cctalk@classiccmp.org> wrote:
> >>
> >> OK, I assumed the 6611s used the NSFnet type cards. Artic960s are
> >> different animals - but probably very similar in idea.
> >>
> >> My memory is hazy, but I think the NSFnet cards were referred to as
> Hawthornes.
> >>
> >> Somewhere around here I have one of the really early 386 based routing
> >> cards - a weird double height Microchannel card (the RS/6000s were
> >> RPQ'd with extra tall chassis to accommodate them).
> >>
> >> Anyway, I would like to get a 6611, but I do not think very many were
> >> made at all.
> >>
> >> --
> >> Will
> >> On Thu, Nov 1, 2018 at 6:19 PM Paul Berger via cctalk
> >>  wrote:
> >> >
> >> > The machine type was 6611 and there where three model, the smallest
> was
> >> > based on a 7011 the mid size one was based on a 7012 and the largest
> was
> >> > based on a 7013.
> >> >
> >> > The base card is an Artic 960 card which is just a processor card with
> >> > some memory that gets an application loaded on the fly.  The top
> >> > interface card has a lot to do with determining what the function of
> the
> >> > card sandwich is, there should be a X-Y type code on the back of the
> >> > card that would define the interface.  They where used for all kinds
> of
> >> > things like Synchronous communications, X25 and network accelerators.
> >> > Some of the interfaces cards used in the 6611 where unique to it and
> >> > never made it to the "standard" RS/6000 line.  There was also a PCI
> >> > version of the Artic 960 but by the time it came along the 6611 was
> long
> >> > gone.
> >> >
> >> > Paul.
> >> >
> >> >
> >> > On 2018-11-01 1:15 PM, William Donzelli via cctalk wrote:
> >> > >> So, what is this i960-based card for?
> >> > > They were the routers. At the core nodes of the network, there would
> >> > > be a big RS/6000s (very early POWER1 types) that would each do about
> >> > > 4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
> >> > > was one of these cards, so each of the big RS/6000s would have about
> >> > > 4-5 of these cards.
> >> > >
> >> > > IBM tried to commercialize the design, but it was doomed - the
> routing
> >> > > engines were very fast, but the internet quickly outgrew the
> >> > > architecture of the engines, and they apparently needed a complete
> >> > > redesign to compete. IBM did release very few of these RS/6000s to
> the
> >> > > public (I think RS/6000-320Hs with a fancy tag - machine type
> 6767?).
> >> > > I have only seen one of these routers in the wild, but most of the
> >> > > real NSFnet ones (I was decommissioning them, one time with a
> Sawzall
> >> > > because of some live tangled cables).
> >> > >
> >> > >> Could it be related to what you
> >> > >> say in your post?
> >> > >>
> >> > >> https://imgur.com/NIvQPBv
> >> > > Possibly related, but that card is not one of the NSFnet ones.
> >> > >
> >> > > --
> >> > > Will
> >> >
>

Re: i860: Re: modern stuff

[William Donzelli via cctalk]

I have quite  few NSFnet backup tapes that need to go to Al at some
point. They likely have good stuff on them.

--
Will
On Sat, Nov 3, 2018 at 5:03 AM Kevin Bowling  wrote:
>
> Do you have software or docs for any of this stuff? I run ps-2.kev009.com
>
> On Thu, Nov 1, 2018 at 3:41 PM William Donzelli via cctalk 
>  wrote:
>>
>> OK, I assumed the 6611s used the NSFnet type cards. Artic960s are
>> different animals - but probably very similar in idea.
>>
>> My memory is hazy, but I think the NSFnet cards were referred to as 
>> Hawthornes.
>>
>> Somewhere around here I have one of the really early 386 based routing
>> cards - a weird double height Microchannel card (the RS/6000s were
>> RPQ'd with extra tall chassis to accommodate them).
>>
>> Anyway, I would like to get a 6611, but I do not think very many were
>> made at all.
>>
>> --
>> Will
>> On Thu, Nov 1, 2018 at 6:19 PM Paul Berger via cctalk
>>  wrote:
>> >
>> > The machine type was 6611 and there where three model, the smallest was
>> > based on a 7011 the mid size one was based on a 7012 and the largest was
>> > based on a 7013.
>> >
>> > The base card is an Artic 960 card which is just a processor card with
>> > some memory that gets an application loaded on the fly.  The top
>> > interface card has a lot to do with determining what the function of the
>> > card sandwich is, there should be a X-Y type code on the back of the
>> > card that would define the interface.  They where used for all kinds of
>> > things like Synchronous communications, X25 and network accelerators.
>> > Some of the interfaces cards used in the 6611 where unique to it and
>> > never made it to the "standard" RS/6000 line.  There was also a PCI
>> > version of the Artic 960 but by the time it came along the 6611 was long
>> > gone.
>> >
>> > Paul.
>> >
>> >
>> > On 2018-11-01 1:15 PM, William Donzelli via cctalk wrote:
>> > >> So, what is this i960-based card for?
>> > > They were the routers. At the core nodes of the network, there would
>> > > be a big RS/6000s (very early POWER1 types) that would each do about
>> > > 4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
>> > > was one of these cards, so each of the big RS/6000s would have about
>> > > 4-5 of these cards.
>> > >
>> > > IBM tried to commercialize the design, but it was doomed - the routing
>> > > engines were very fast, but the internet quickly outgrew the
>> > > architecture of the engines, and they apparently needed a complete
>> > > redesign to compete. IBM did release very few of these RS/6000s to the
>> > > public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
>> > > I have only seen one of these routers in the wild, but most of the
>> > > real NSFnet ones (I was decommissioning them, one time with a Sawzall
>> > > because of some live tangled cables).
>> > >
>> > >> Could it be related to what you
>> > >> say in your post?
>> > >>
>> > >> https://imgur.com/NIvQPBv
>> > > Possibly related, but that card is not one of the NSFnet ones.
>> > >
>> > > --
>> > > Will
>> >

Re: i860: Re: modern stuff

[Kevin Bowling via cctalk]

Do you have software or docs for any of this stuff? I run ps-2.kev009.com

On Thu, Nov 1, 2018 at 3:41 PM William Donzelli via cctalk <
cctalk@classiccmp.org> wrote:

> OK, I assumed the 6611s used the NSFnet type cards. Artic960s are
> different animals - but probably very similar in idea.
>
> My memory is hazy, but I think the NSFnet cards were referred to as
> Hawthornes.
>
> Somewhere around here I have one of the really early 386 based routing
> cards - a weird double height Microchannel card (the RS/6000s were
> RPQ'd with extra tall chassis to accommodate them).
>
> Anyway, I would like to get a 6611, but I do not think very many were
> made at all.
>
> --
> Will
> On Thu, Nov 1, 2018 at 6:19 PM Paul Berger via cctalk
>  wrote:
> >
> > The machine type was 6611 and there where three model, the smallest was
> > based on a 7011 the mid size one was based on a 7012 and the largest was
> > based on a 7013.
> >
> > The base card is an Artic 960 card which is just a processor card with
> > some memory that gets an application loaded on the fly.  The top
> > interface card has a lot to do with determining what the function of the
> > card sandwich is, there should be a X-Y type code on the back of the
> > card that would define the interface.  They where used for all kinds of
> > things like Synchronous communications, X25 and network accelerators.
> > Some of the interfaces cards used in the 6611 where unique to it and
> > never made it to the "standard" RS/6000 line.  There was also a PCI
> > version of the Artic 960 but by the time it came along the 6611 was long
> > gone.
> >
> > Paul.
> >
> >
> > On 2018-11-01 1:15 PM, William Donzelli via cctalk wrote:
> > >> So, what is this i960-based card for?
> > > They were the routers. At the core nodes of the network, there would
> > > be a big RS/6000s (very early POWER1 types) that would each do about
> > > 4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
> > > was one of these cards, so each of the big RS/6000s would have about
> > > 4-5 of these cards.
> > >
> > > IBM tried to commercialize the design, but it was doomed - the routing
> > > engines were very fast, but the internet quickly outgrew the
> > > architecture of the engines, and they apparently needed a complete
> > > redesign to compete. IBM did release very few of these RS/6000s to the
> > > public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
> > > I have only seen one of these routers in the wild, but most of the
> > > real NSFnet ones (I was decommissioning them, one time with a Sawzall
> > > because of some live tangled cables).
> > >
> > >> Could it be related to what you
> > >> say in your post?
> > >>
> > >> https://imgur.com/NIvQPBv
> > > Possibly related, but that card is not one of the NSFnet ones.
> > >
> > > --
> > > Will
> >
>

Re: i860: Re: modern stuff

[Eric Korpela via cctalk]

On Wed, Oct 31, 2018 at 2:27 PM Eric Korpela 
wrote:

>
> I also seem to recall that the SERENDIP III SETI spectrometer used i860
> and Austek A41102 FFT processors.  I'm pretty sure SERENDIP IV used i960
> and Xylinx FPGAs to do the FFTs.  I'll look at the boards tomorrow.
>

I was wrong on both counts.  SERENDIP III didn't have a CPU on the board,
but had a state machine on one of the FPGAs to control storage into and
retrieval from the memory.  And SERENDIP IV did use the A41102 FFT
processors.  Apparently, SERENDIP V.v was the first that did its FFTs in
the FPGA.  Since then (SERENDIP VI) we've moved to GPUs, but that's less
than 10 years old so I shouldn't even mention it.

Here's the obligatory board porn.  (Yes, SERENDIP III was wire wrap).
https://purcell.ssl.berkeley.edu/~korpela/gif/SERENDIPIII.jpg  (that's the
whole thing)
https://purcell.ssl.berkeley.edu/~korpela/gif/SERENDIPIV.jpg  (that's 1 of
20 boards)

Re: i860: Re: modern stuff

[Eric Korpela via cctalk]

>
>
> Was the 1983-84 year multibus sky floating point card the first
> offering from Sky Computers ?
>
> Did anyone use those in an embedded and online floating-point realtime
> type of setting ? Or was they only used for off-line number-crunching ?
>

Sky made the math coprocessor in early multibus and VME Sun 2 systems as
well.  I'm not aware of any embedded use of Sky FFPs

-- 
Eric Korpela
korp...@ssl.berkeley.edu
AST:7731^29u18e3

Re: i860: Re: modern stuff

[Stefan Skoglund via cctalk]

ons 2018-10-31 klockan 14:27 -0700 skrev Eric Korpela via cctalk:
> The i860 did find some use in the radio astronomy world.
> 
> Here's an excerpt from the 1998 annual report for the Arecibo
> Observatory...
> --
> Telescope pointing and realtime data acquisition are controlled using
> a
> network of VMEbus single-board computers running the VxWorks
> operating
> system kernel. Custom-built data acquisition devices (‘‘backends’’)
> include
> (1) a general purpose A/D system capable of sampling four analog
> channels
> at up to 10-MHz rates with programmable resolutions of 1 to 12 bits
> per
> sample per channel, (2) an ~interim! 50-MHz, 4096-lag Spectral Line
> Correlator with programmable bandwidth from 195 kHz to 50 MHz, (3) a
> 50-MHz
> Radar Decoder, ~4! a 100-MHz Spectral Line Correlator being
> developed, (5)
> a 10-MHz bandwidth Pulsar Search/Timing Machine with up to 256
> channels,
> and (6) a wideband continuum/polarimetry instrument being developed.
> An S2
> VLBI system is also available. Additional realtime signal processing
> capability is provided by four Skybolt i860-based VMEbus single-board
> computers with 240 MFLOPS peak combined capacity.
> --
> Remember when 240 MFLOPS was a lot?
> 

Was the 1983-84 year multibus sky floating point card the first
offering from Sky Computers ?

Did anyone use those in an embedded and online floating-point realtime
type of setting ? Or was they only used for off-line number-crunching ?

Hrrm, i now know that SKY computers had a dual-port memory system for
DEC LSI-11 computers (good if you have for example a really fast
accessory) so the FFP wasnt the first thing for them.

Re: i860: Re: modern stuff

[Michael Thompson via cctalk]

>
> Date: Wed, 31 Oct 2018 14:29:18 -0700
> From: Eric Korpela 
> Subject: Re: i860: Re: modern stuff
>
> A Google search on Skybolt i860 produces interesting results.
> >Additional realtime signal processing
> > capability is provided by four Skybolt i860-based VMEbus single-board
> > computers with 240 MFLOPS peak combined capacity.
> > --
> > Remember when 240 MFLOPS was a lot?
>

That's the board that I have.

Quad i860 on a 9Ux400 VME board.

Its in a Sun 4/280 development system.


-- 
Michael Thompson

Re: i860: Re: modern stuff

[William Donzelli via cctalk]

OK, I assumed the 6611s used the NSFnet type cards. Artic960s are
different animals - but probably very similar in idea.

My memory is hazy, but I think the NSFnet cards were referred to as Hawthornes.

Somewhere around here I have one of the really early 386 based routing
cards - a weird double height Microchannel card (the RS/6000s were
RPQ'd with extra tall chassis to accommodate them).

Anyway, I would like to get a 6611, but I do not think very many were
made at all.

--
Will
On Thu, Nov 1, 2018 at 6:19 PM Paul Berger via cctalk
 wrote:
>
> The machine type was 6611 and there where three model, the smallest was
> based on a 7011 the mid size one was based on a 7012 and the largest was
> based on a 7013.
>
> The base card is an Artic 960 card which is just a processor card with
> some memory that gets an application loaded on the fly.  The top
> interface card has a lot to do with determining what the function of the
> card sandwich is, there should be a X-Y type code on the back of the
> card that would define the interface.  They where used for all kinds of
> things like Synchronous communications, X25 and network accelerators.
> Some of the interfaces cards used in the 6611 where unique to it and
> never made it to the "standard" RS/6000 line.  There was also a PCI
> version of the Artic 960 but by the time it came along the 6611 was long
> gone.
>
> Paul.
>
>
> On 2018-11-01 1:15 PM, William Donzelli via cctalk wrote:
> >> So, what is this i960-based card for?
> > They were the routers. At the core nodes of the network, there would
> > be a big RS/6000s (very early POWER1 types) that would each do about
> > 4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
> > was one of these cards, so each of the big RS/6000s would have about
> > 4-5 of these cards.
> >
> > IBM tried to commercialize the design, but it was doomed - the routing
> > engines were very fast, but the internet quickly outgrew the
> > architecture of the engines, and they apparently needed a complete
> > redesign to compete. IBM did release very few of these RS/6000s to the
> > public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
> > I have only seen one of these routers in the wild, but most of the
> > real NSFnet ones (I was decommissioning them, one time with a Sawzall
> > because of some live tangled cables).
> >
> >> Could it be related to what you
> >> say in your post?
> >>
> >> https://imgur.com/NIvQPBv
> > Possibly related, but that card is not one of the NSFnet ones.
> >
> > --
> > Will
>

Re: i860: Re: modern stuff

[Paul Berger via cctalk]

The machine type was 6611 and there where three model, the smallest was 
based on a 7011 the mid size one was based on a 7012 and the largest was 
based on a 7013.


The base card is an Artic 960 card which is just a processor card with 
some memory that gets an application loaded on the fly.  The top 
interface card has a lot to do with determining what the function of the 
card sandwich is, there should be a X-Y type code on the back of the 
card that would define the interface.  They where used for all kinds of 
things like Synchronous communications, X25 and network accelerators.  
Some of the interfaces cards used in the 6611 where unique to it and 
never made it to the "standard" RS/6000 line.  There was also a PCI 
version of the Artic 960 but by the time it came along the 6611 was long 
gone.


Paul.


On 2018-11-01 1:15 PM, William Donzelli via cctalk wrote:

So, what is this i960-based card for?

They were the routers. At the core nodes of the network, there would
be a big RS/6000s (very early POWER1 types) that would each do about
4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
was one of these cards, so each of the big RS/6000s would have about
4-5 of these cards.

IBM tried to commercialize the design, but it was doomed - the routing
engines were very fast, but the internet quickly outgrew the
architecture of the engines, and they apparently needed a complete
redesign to compete. IBM did release very few of these RS/6000s to the
public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
I have only seen one of these routers in the wild, but most of the
real NSFnet ones (I was decommissioning them, one time with a Sawzall
because of some live tangled cables).


Could it be related to what you
say in your post?

https://imgur.com/NIvQPBv

Possibly related, but that card is not one of the NSFnet ones.

--
Will

Re: i860: Re: modern stuff

[Kevin Bowling via cctalk]

Yes, they are. There are reference to those machines in the various
nsfnet written histories but not cross linkage to those great
pictures.
On Thu, Nov 1, 2018 at 2:57 PM William Donzelli  wrote:
>
> Right, thanks. 6611 is correct. I do not think the FDDI or HSSI cards
> made it into those.
>
> The RCS/RI twitter feed has some pictures of NSFnet racks and a F960
> FDDI card. Those were from the GNJ node in Greensboro Junction, NC.
> Were those the pictures?
>
> https://twitter.com/RetroCompSocRI
>
> --
> Will
> On Thu, Nov 1, 2018 at 3:22 PM Kevin Bowling  wrote:
> >
> > 6611 was the commercialized version.  One early model was a standard 7012 
> > desktop with the special cards. A later cost optimized version had a custom 
> > PowerPC backplane.
> >
> > There were some good pics of the nsfnet T3 racks I linked onto nekochan 
> > forums but that site is gone. Wish people would migrate back to Usenet.
> >
> > On Thu, Nov 1, 2018 at 9:15 AM William Donzelli via cctalk 
> >  wrote:
> >>
> >> > So, what is this i960-based card for?
> >>
> >> They were the routers. At the core nodes of the network, there would
> >> be a big RS/6000s (very early POWER1 types) that would each do about
> >> 4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
> >> was one of these cards, so each of the big RS/6000s would have about
> >> 4-5 of these cards.
> >>
> >> IBM tried to commercialize the design, but it was doomed - the routing
> >> engines were very fast, but the internet quickly outgrew the
> >> architecture of the engines, and they apparently needed a complete
> >> redesign to compete. IBM did release very few of these RS/6000s to the
> >> public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
> >> I have only seen one of these routers in the wild, but most of the
> >> real NSFnet ones (I was decommissioning them, one time with a Sawzall
> >> because of some live tangled cables).
> >>
> >> > Could it be related to what you
> >> > say in your post?
> >> >
> >> > https://imgur.com/NIvQPBv
> >>
> >> Possibly related, but that card is not one of the NSFnet ones.
> >>
> >> --
> >> Will

Re: i860: Re: modern stuff

[William Donzelli via cctalk]

Right, thanks. 6611 is correct. I do not think the FDDI or HSSI cards
made it into those.

The RCS/RI twitter feed has some pictures of NSFnet racks and a F960
FDDI card. Those were from the GNJ node in Greensboro Junction, NC.
Were those the pictures?

https://twitter.com/RetroCompSocRI

--
Will
On Thu, Nov 1, 2018 at 3:22 PM Kevin Bowling  wrote:
>
> 6611 was the commercialized version.  One early model was a standard 7012 
> desktop with the special cards. A later cost optimized version had a custom 
> PowerPC backplane.
>
> There were some good pics of the nsfnet T3 racks I linked onto nekochan 
> forums but that site is gone. Wish people would migrate back to Usenet.
>
> On Thu, Nov 1, 2018 at 9:15 AM William Donzelli via cctalk 
>  wrote:
>>
>> > So, what is this i960-based card for?
>>
>> They were the routers. At the core nodes of the network, there would
>> be a big RS/6000s (very early POWER1 types) that would each do about
>> 4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
>> was one of these cards, so each of the big RS/6000s would have about
>> 4-5 of these cards.
>>
>> IBM tried to commercialize the design, but it was doomed - the routing
>> engines were very fast, but the internet quickly outgrew the
>> architecture of the engines, and they apparently needed a complete
>> redesign to compete. IBM did release very few of these RS/6000s to the
>> public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
>> I have only seen one of these routers in the wild, but most of the
>> real NSFnet ones (I was decommissioning them, one time with a Sawzall
>> because of some live tangled cables).
>>
>> > Could it be related to what you
>> > say in your post?
>> >
>> > https://imgur.com/NIvQPBv
>>
>> Possibly related, but that card is not one of the NSFnet ones.
>>
>> --
>> Will

Whither Google Groups - was Re: i860: Re: modern stuff

[Toby Thain via cctalk]

On 2018-11-01 5:06 p.m., Al Kossow via cctalk wrote:
> 
> 
> On 11/1/18 12:22 PM, Kevin Bowling via cctalk wrote:
> 
>> There were some good pics of the nsfnet T3 racks I linked onto nekochan
>> forums but that site is gone. Wish people would migrate back to Usenet.
> 
> Community fragmentation and reliance on unarchived forums is a Bad Thing.
> I wonder how much of value will be lost when Yahoo's unmigrated forums 
> finally collapse
> or when G gets bored with all of the 'Usenet' groups they created and 
> silently dump them.
> 
> 

Groups is getting more and more buried in the UI. You have to really dig
to browse there from the home page now.

--T

Re: i860: Re: modern stuff

[Al Kossow via cctalk]

On 11/1/18 12:22 PM, Kevin Bowling via cctalk wrote:

> There were some good pics of the nsfnet T3 racks I linked onto nekochan
> forums but that site is gone. Wish people would migrate back to Usenet.

Community fragmentation and reliance on unarchived forums is a Bad Thing.
I wonder how much of value will be lost when Yahoo's unmigrated forums finally 
collapse
or when G gets bored with all of the 'Usenet' groups they created and silently 
dump them.

Re: i860: Re: modern stuff

[Kevin Bowling via cctalk]

6611 was the commercialized version.  One early model was a standard 7012
desktop with the special cards. A later cost optimized version had a custom
PowerPC backplane.

There were some good pics of the nsfnet T3 racks I linked onto nekochan
forums but that site is gone. Wish people would migrate back to Usenet.

On Thu, Nov 1, 2018 at 9:15 AM William Donzelli via cctalk <
cctalk@classiccmp.org> wrote:

> > So, what is this i960-based card for?
>
> They were the routers. At the core nodes of the network, there would
> be a big RS/6000s (very early POWER1 types) that would each do about
> 4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
> was one of these cards, so each of the big RS/6000s would have about
> 4-5 of these cards.
>
> IBM tried to commercialize the design, but it was doomed - the routing
> engines were very fast, but the internet quickly outgrew the
> architecture of the engines, and they apparently needed a complete
> redesign to compete. IBM did release very few of these RS/6000s to the
> public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
> I have only seen one of these routers in the wild, but most of the
> real NSFnet ones (I was decommissioning them, one time with a Sawzall
> because of some live tangled cables).
>
> > Could it be related to what you
> > say in your post?
> >
> > https://imgur.com/NIvQPBv
>
> Possibly related, but that card is not one of the NSFnet ones.
>
> --
> Will
>

Re: i860: Re: modern stuff

[Grant Taylor via cctalk]

On 11/01/2018 10:15 AM, William Donzelli via cctalk wrote:

I have only seen one of these routers in the wild


I worked at a company years ago that ran a pair of IBM RS/6000 (the 
small beige desktop models, maybe a 43).  They were running an IBM 
Firewall software product that I don't remember the name of.  I think it 
ran on top of AIX 4..




--
Grant. . . .
unix || die

Re: i860: Re: modern stuff

[William Donzelli via cctalk]

> So, what is this i960-based card for?

They were the routers. At the core nodes of the network, there would
be a big RS/6000s (very early POWER1 types) that would each do about
4-5 high speed interfaces (FDDI, HSSI, and 10base2). Each interface
was one of these cards, so each of the big RS/6000s would have about
4-5 of these cards.

IBM tried to commercialize the design, but it was doomed - the routing
engines were very fast, but the internet quickly outgrew the
architecture of the engines, and they apparently needed a complete
redesign to compete. IBM did release very few of these RS/6000s to the
public (I think RS/6000-320Hs with a fancy tag - machine type 6767?).
I have only seen one of these routers in the wild, but most of the
real NSFnet ones (I was decommissioning them, one time with a Sawzall
because of some live tangled cables).

> Could it be related to what you
> say in your post?
>
> https://imgur.com/NIvQPBv

Possibly related, but that card is not one of the NSFnet ones.

--
Will

Re: i860: Re: modern stuff

[Eric Korpela via cctalk]

A Google search on Skybolt i860 produces interesting results.

On Wed, Oct 31, 2018 at 2:27 PM Eric Korpela 
wrote:

> The i860 did find some use in the radio astronomy world.
>
> Here's an excerpt from the 1998 annual report for the Arecibo
> Observatory...
> --
> Telescope pointing and realtime data acquisition are controlled using a
> network of VMEbus single-board computers running the VxWorks operating
> system kernel. Custom-built data acquisition devices (‘‘backends’’) include
> (1) a general purpose A/D system capable of sampling four analog channels
> at up to 10-MHz rates with programmable resolutions of 1 to 12 bits per
> sample per channel, (2) an ~interim! 50-MHz, 4096-lag Spectral Line
> Correlator with programmable bandwidth from 195 kHz to 50 MHz, (3) a 50-MHz
> Radar Decoder, ~4! a 100-MHz Spectral Line Correlator being developed, (5)
> a 10-MHz bandwidth Pulsar Search/Timing Machine with up to 256 channels,
> and (6) a wideband continuum/polarimetry instrument being developed. An S2
> VLBI system is also available. Additional realtime signal processing
> capability is provided by four Skybolt i860-based VMEbus single-board
> computers with 240 MFLOPS peak combined capacity.
> --
> Remember when 240 MFLOPS was a lot?
>
> I also seem to recall that the SERENDIP III SETI spectrometer used i860
> and Austek A41102 FFT processors.  I'm pretty sure SERENDIP IV used i960
> and Xylinx FPGAs to do the FFTs.  I'll look at the boards tomorrow.
>
>
>
> On Mon, Oct 29, 2018 at 9:56 AM Ken Seefried via cctalk <
> cctalk@classiccmp.org> wrote:
>
>> >the i860 found at least a little niche on graphics boards, so somehow
>> >not a complete failure ;-)
>>
>> I'd be mildly surprised if Intel ever made enough from selling i860s
>> as GPUs to cover the cost of developing and marketing them.  At the
>> time, Intel was pushing them as their RISC processor, and put a lot
>> into the program.  Going to take over the world and all that.  Maybe
>> not a 'complete' failure...just mostly.
>>
>> From: Chuck Guzis 
>> >On 10/26/18 6:10 AM, Gordon Henderson via cctalk wrote:
>> >
>> >> However it was a royal PITA to code for although a 32-bit CPU, it would
>> >> read memory 64 bits at a time (actually 128 IIRC to satisfy the cache),
>> >> with half that 64-bit word being an instruction for the integer unit
>> and
>> >> half for the floating point unit, so you effectively had to build a
>> >> floating point pipeline by hand coded instructions, so 8 (I think)
>> >> instructions to load the pipeline, then each subsequent instruction
>> >> would feed another value into the pipe, then another 8 at the end to
>> >> empty it. Great for big matrix operations, rubbish for a single add of
>> 2
>> >> FP numbers.
>> >
>> >My impression of the i860 was that it might have been fun for about 2
>> >weeks for which to code assembly, but after that, you'd really start
>> >looking hard for an HLL to do the dirty work for you.  While there's a
>> >sense of accomplishment over looking at a page of painfully
>> >hand-optimized code that manages to keep everything busy with no
>> >"bubbles", you begin to wonder if there isn't a better way to spend your
>> >life.
>>
>> It wasn't fun for the whole 2 weeks.  And the i860 is Yet Another
>> example of Intel claiming their compilers were going to be so smart
>> that all the architectural complexity/warts will never be noticed.
>> Wrong, and they didn't learn and said the same thing about Itanium.
>> The interrupt stall issue that Gordon pointed out was so bad they were
>> basically relegated to single-task software in the end.
>>
>> KJ
>>
>
>
> --
> Eric Korpela
> korp...@ssl.berkeley.edu
> AST:7731^29u18e3
>


-- 
Eric Korpela
korp...@ssl.berkeley.edu
AST:7731^29u18e3

Re: i860: Re: modern stuff

[Eric Korpela via cctalk]

The i860 did find some use in the radio astronomy world.

Here's an excerpt from the 1998 annual report for the Arecibo Observatory...
--
Telescope pointing and realtime data acquisition are controlled using a
network of VMEbus single-board computers running the VxWorks operating
system kernel. Custom-built data acquisition devices (‘‘backends’’) include
(1) a general purpose A/D system capable of sampling four analog channels
at up to 10-MHz rates with programmable resolutions of 1 to 12 bits per
sample per channel, (2) an ~interim! 50-MHz, 4096-lag Spectral Line
Correlator with programmable bandwidth from 195 kHz to 50 MHz, (3) a 50-MHz
Radar Decoder, ~4! a 100-MHz Spectral Line Correlator being developed, (5)
a 10-MHz bandwidth Pulsar Search/Timing Machine with up to 256 channels,
and (6) a wideband continuum/polarimetry instrument being developed. An S2
VLBI system is also available. Additional realtime signal processing
capability is provided by four Skybolt i860-based VMEbus single-board
computers with 240 MFLOPS peak combined capacity.
--
Remember when 240 MFLOPS was a lot?

I also seem to recall that the SERENDIP III SETI spectrometer used i860 and
Austek A41102 FFT processors.  I'm pretty sure SERENDIP IV used i960 and
Xylinx FPGAs to do the FFTs.  I'll look at the boards tomorrow.



On Mon, Oct 29, 2018 at 9:56 AM Ken Seefried via cctalk <
cctalk@classiccmp.org> wrote:

> >the i860 found at least a little niche on graphics boards, so somehow
> >not a complete failure ;-)
>
> I'd be mildly surprised if Intel ever made enough from selling i860s
> as GPUs to cover the cost of developing and marketing them.  At the
> time, Intel was pushing them as their RISC processor, and put a lot
> into the program.  Going to take over the world and all that.  Maybe
> not a 'complete' failure...just mostly.
>
> From: Chuck Guzis 
> >On 10/26/18 6:10 AM, Gordon Henderson via cctalk wrote:
> >
> >> However it was a royal PITA to code for although a 32-bit CPU, it would
> >> read memory 64 bits at a time (actually 128 IIRC to satisfy the cache),
> >> with half that 64-bit word being an instruction for the integer unit and
> >> half for the floating point unit, so you effectively had to build a
> >> floating point pipeline by hand coded instructions, so 8 (I think)
> >> instructions to load the pipeline, then each subsequent instruction
> >> would feed another value into the pipe, then another 8 at the end to
> >> empty it. Great for big matrix operations, rubbish for a single add of 2
> >> FP numbers.
> >
> >My impression of the i860 was that it might have been fun for about 2
> >weeks for which to code assembly, but after that, you'd really start
> >looking hard for an HLL to do the dirty work for you.  While there's a
> >sense of accomplishment over looking at a page of painfully
> >hand-optimized code that manages to keep everything busy with no
> >"bubbles", you begin to wonder if there isn't a better way to spend your
> >life.
>
> It wasn't fun for the whole 2 weeks.  And the i860 is Yet Another
> example of Intel claiming their compilers were going to be so smart
> that all the architectural complexity/warts will never be noticed.
> Wrong, and they didn't learn and said the same thing about Itanium.
> The interrupt stall issue that Gordon pointed out was so bad they were
> basically relegated to single-task software in the end.
>
> KJ
>


-- 
Eric Korpela
korp...@ssl.berkeley.edu
AST:7731^29u18e3

Re: i860: Re: modern stuff

[Chris Hanson via cctalk]

On Oct 30, 2018, at 6:48 AM, Carlos E Murillo-Sanchez via cctalk 
 wrote:
> 
> So, what is this i960-based card for?  Could it be related to what you say in 
> your post?
> 
> https://imgur.com/NIvQPBv 
> https://imgur.com/hsF0jO2 
> https://imgur.com/7f6sxDj 
> https://imgur.com/w96cLhT 
Just from the wording on the label it seems like some sort of telecom card, but 
it could also be a development board for the software that would run on telecom 
systems.

  -- Chris

Re: i860: Re: modern stuff

[Al Kossow via cctalk]

On 10/30/18 4:27 PM, Michael Thompson via cctech wrote:

> I have a Quad-i860 VME board in one of my Sun systems.

Do you have any of the software for it?

Re: i860: Re: modern stuff

[Michael Thompson via cctalk]

>
> From: Ken Seefried 
> Subject: i860: Re: modern stuff
> >the i860 found at least a little niche on graphics boards, so somehow
> >not a complete failure ;-)
>
>
I have a Quad-i860 VME board in one of my Sun systems.

Michael Thompson

Re: i860: Re: modern stuff

[Carlos E Murillo-Sanchez via cctalk]

William Donzelli via cctalk wrote:

AIX was ported in very cut down manner and used on the f960 and h960
routing cards used on the early T3 based NSFnet. F960 was FDDI and H960 was
HSSI. Come think of it, I think the v.25 and ether net cards also used
i960, just a smaller version.

--
Will

So, what is this i960-based card for?  Could it be related to what you 
say in your post?


https://imgur.com/NIvQPBv
https://imgur.com/hsF0jO2
https://imgur.com/7f6sxDj
https://imgur.com/w96cLhT

Carlos.

Re: i860: Re: modern stuff

[Rob Doyle via cctalk]

I'm aware of some airborne (avionics) computers that used i960s.  There 
were Mil-spec versions available.


I believe the i960 really only found a niche in embedded applications. 
If I recall correctly, the i960 became available at about the same time 
as the 80386 but was less expensive.  At the time, you couldn't afford 
to put a 80386 in an embedded application because Intel was getting 
artificially high prices due to PC-based pricing.


Intel later tried to address the embedded market pricing issues by 
releasing the 80376 and later the 80386EX.  Both those products were 
munged so they couldn't run DOS and that kept their pricing model intact.


The i960 did have a user mode and supervisor mode - so it could have 
supported a 'real' OS.


Rob.

On 10/29/2018 11:12 AM, alan--- via cctalk wrote:


I know i960 is a very different beast, but was there ever any high level 
OSs that ran on it?  Or was it pidgin-holed as a high speed embedded 
processor for storage controllers and NICs?


I picked up a cache of i960 CPUs a couple years ago and they speak to me 
in tongues every time I pass by the shelf.


-Alan


On 2018-10-29 12:56, Ken Seefried via cctalk wrote:

the i860 found at least a little niche on graphics boards, so somehow
not a complete failure ;-)


I'd be mildly surprised if Intel ever made enough from selling i860s
as GPUs to cover the cost of developing and marketing them.  At the
time, Intel was pushing them as their RISC processor, and put a lot
into the program.  Going to take over the world and all that.  Maybe
not a 'complete' failure...just mostly.

From: Chuck Guzis 

On 10/26/18 6:10 AM, Gordon Henderson via cctalk wrote:


However it was a royal PITA to code for although a 32-bit CPU, it would
read memory 64 bits at a time (actually 128 IIRC to satisfy the cache),
with half that 64-bit word being an instruction for the integer unit 
and

half for the floating point unit, so you effectively had to build a
floating point pipeline by hand coded instructions, so 8 (I think)
instructions to load the pipeline, then each subsequent instruction
would feed another value into the pipe, then another 8 at the end to
empty it. Great for big matrix operations, rubbish for a single add 
of 2

FP numbers.


My impression of the i860 was that it might have been fun for about 2
weeks for which to code assembly, but after that, you'd really start
looking hard for an HLL to do the dirty work for you.  While there's a
sense of accomplishment over looking at a page of painfully
hand-optimized code that manages to keep everything busy with no
"bubbles", you begin to wonder if there isn't a better way to spend your
life.


It wasn't fun for the whole 2 weeks.  And the i860 is Yet Another
example of Intel claiming their compilers were going to be so smart
that all the architectural complexity/warts will never be noticed.
Wrong, and they didn't learn and said the same thing about Itanium.
The interrupt stall issue that Gordon pointed out was so bad they were
basically relegated to single-task software in the end.

KJ

Re: i860: Re: modern stuff

[William Donzelli via cctalk]

AIX was ported in very cut down manner and used on the f960 and h960
routing cards used on the early T3 based NSFnet. F960 was FDDI and H960 was
HSSI. Come think of it, I think the v.25 and ether net cards also used
i960, just a smaller version.

--
Will

On Oct 29, 2018 12:13 PM, "alan--- via cctalk" 
wrote:


I know i960 is a very different beast, but was there ever any high level
OSs that ran on it?  Or was it pidgin-holed as a high speed embedded
processor for storage controllers and NICs?

I picked up a cache of i960 CPUs a couple years ago and they speak to me in
tongues every time I pass by the shelf.

-Alan



On 2018-10-29 12:56, Ken Seefried via cctalk wrote:

> the i860 found at least a little niche on graphics boards, so somehow
>> not a complete failure ;-)
>>
>
> I'd be mildly surprised if Intel ever made enough from selling i860s
> as GPUs to cover the cost of developing and marketing them.  At the
> time, Intel was pushing them as their RISC processor, and put a lot
> into the program.  Going to take over the world and all that.  Maybe
> not a 'complete' failure...just mostly.
>
> From: Chuck Guzis 
>
>> On 10/26/18 6:10 AM, Gordon Henderson via cctalk wrote:
>>
>> However it was a royal PITA to code for although a 32-bit CPU, it would
>>> read memory 64 bits at a time (actually 128 IIRC to satisfy the cache),
>>> with half that 64-bit word being an instruction for the integer unit and
>>> half for the floating point unit, so you effectively had to build a
>>> floating point pipeline by hand coded instructions, so 8 (I think)
>>> instructions to load the pipeline, then each subsequent instruction
>>> would feed another value into the pipe, then another 8 at the end to
>>> empty it. Great for big matrix operations, rubbish for a single add of 2
>>> FP numbers.
>>>
>>
>> My impression of the i860 was that it might have been fun for about 2
>> weeks for which to code assembly, but after that, you'd really start
>> looking hard for an HLL to do the dirty work for you.  While there's a
>> sense of accomplishment over looking at a page of painfully
>> hand-optimized code that manages to keep everything busy with no
>> "bubbles", you begin to wonder if there isn't a better way to spend your
>> life.
>>
>
> It wasn't fun for the whole 2 weeks.  And the i860 is Yet Another
> example of Intel claiming their compilers were going to be so smart
> that all the architectural complexity/warts will never be noticed.
> Wrong, and they didn't learn and said the same thing about Itanium.
> The interrupt stall issue that Gordon pointed out was so bad they were
> basically relegated to single-task software in the end.
>
> KJ
>

Re: i860: Re: modern stuff

[Paul Koning via cctalk]

> On Oct 29, 2018, at 5:12 PM, Eric Smith via cctalk  
> wrote:
> 
> The i960 was how Intel repositioned it to try to salvage as much as
> possible. Most i960 variants either don't have the tag bit hardware and
> object-oriented "microcode" that was used for BiiN; it is only officially
> present in the i960MX, but might also be in the i960MC. I think only the MX
> and MC have an MMU.

I used an i960 at Chipcom.  At this point, the only thing I remember is a very 
awful I/O architecture.  It felt a bit like the seriously broken architecture 
of the 82586 Ethernet chip.  (Yes, it's possible to design a queue based I/O 
architecture that is correct; Dijkstra did so in the Electrologica X8.  But 
Intel was clearly clueless about making distributed algorithms work.)

paul

Re: i860: Re: modern stuff

[Eric Smith via cctalk]

On Mon, Oct 29, 2018 at 12:13 PM alan--- via cctalk 
wrote:

> I know i960 is a very different beast, but was there ever any high level
> OSs that ran on it?


It was originally the BiiN processor, and ran the Osiris operating system.
However, few if any were sold, and it disappeared without a trace, leading
to BiiN being retronym'd "Billions invested in Nothing".

The i960 was how Intel repositioned it to try to salvage as much as
possible. Most i960 variants either don't have the tag bit hardware and
object-oriented "microcode" that was used for BiiN; it is only officially
present in the i960MX, but might also be in the i960MC. I think only the MX
and MC have an MMU.

It should be possible to port {Unix,xBSD,Linux} to the i960MC, but I don't
think anyone has.

Re: i860: Re: modern stuff

[alan--- via cctalk]

I know i960 is a very different beast, but was there ever any high level 
OSs that ran on it?  Or was it pidgin-holed as a high speed embedded 
processor for storage controllers and NICs?


I picked up a cache of i960 CPUs a couple years ago and they speak to me 
in tongues every time I pass by the shelf.


-Alan


On 2018-10-29 12:56, Ken Seefried via cctalk wrote:

the i860 found at least a little niche on graphics boards, so somehow
not a complete failure ;-)


I'd be mildly surprised if Intel ever made enough from selling i860s
as GPUs to cover the cost of developing and marketing them.  At the
time, Intel was pushing them as their RISC processor, and put a lot
into the program.  Going to take over the world and all that.  Maybe
not a 'complete' failure...just mostly.

From: Chuck Guzis 

On 10/26/18 6:10 AM, Gordon Henderson via cctalk wrote:

However it was a royal PITA to code for although a 32-bit CPU, it 
would
read memory 64 bits at a time (actually 128 IIRC to satisfy the 
cache),
with half that 64-bit word being an instruction for the integer unit 
and

half for the floating point unit, so you effectively had to build a
floating point pipeline by hand coded instructions, so 8 (I think)
instructions to load the pipeline, then each subsequent instruction
would feed another value into the pipe, then another 8 at the end to
empty it. Great for big matrix operations, rubbish for a single add 
of 2

FP numbers.


My impression of the i860 was that it might have been fun for about 2
weeks for which to code assembly, but after that, you'd really start
looking hard for an HLL to do the dirty work for you.  While there's a
sense of accomplishment over looking at a page of painfully
hand-optimized code that manages to keep everything busy with no
"bubbles", you begin to wonder if there isn't a better way to spend 
your

life.


It wasn't fun for the whole 2 weeks.  And the i860 is Yet Another
example of Intel claiming their compilers were going to be so smart
that all the architectural complexity/warts will never be noticed.
Wrong, and they didn't learn and said the same thing about Itanium.
The interrupt stall issue that Gordon pointed out was so bad they were
basically relegated to single-task software in the end.

KJ

i860: Re: modern stuff

[Ken Seefried via cctalk]

>the i860 found at least a little niche on graphics boards, so somehow
>not a complete failure ;-)

I'd be mildly surprised if Intel ever made enough from selling i860s
as GPUs to cover the cost of developing and marketing them.  At the
time, Intel was pushing them as their RISC processor, and put a lot
into the program.  Going to take over the world and all that.  Maybe
not a 'complete' failure...just mostly.

From: Chuck Guzis 
>On 10/26/18 6:10 AM, Gordon Henderson via cctalk wrote:
>
>> However it was a royal PITA to code for although a 32-bit CPU, it would
>> read memory 64 bits at a time (actually 128 IIRC to satisfy the cache),
>> with half that 64-bit word being an instruction for the integer unit and
>> half for the floating point unit, so you effectively had to build a
>> floating point pipeline by hand coded instructions, so 8 (I think)
>> instructions to load the pipeline, then each subsequent instruction
>> would feed another value into the pipe, then another 8 at the end to
>> empty it. Great for big matrix operations, rubbish for a single add of 2
>> FP numbers.
>
>My impression of the i860 was that it might have been fun for about 2
>weeks for which to code assembly, but after that, you'd really start
>looking hard for an HLL to do the dirty work for you.  While there's a
>sense of accomplishment over looking at a page of painfully
>hand-optimized code that manages to keep everything busy with no
>"bubbles", you begin to wonder if there isn't a better way to spend your
>life.

It wasn't fun for the whole 2 weeks.  And the i860 is Yet Another
example of Intel claiming their compilers were going to be so smart
that all the architectural complexity/warts will never be noticed.
Wrong, and they didn't learn and said the same thing about Itanium.
The interrupt stall issue that Gordon pointed out was so bad they were
basically relegated to single-task software in the end.

KJ