> I just received the engineering review from Maxim.
> If I only take a look at the pictures,
> we should immediately stop using PICs and AVRs and go for
>Maxim. Judge yourself:
Received that the other day, and had a quick glance through it, but hadn't
appreciated what they were peddling. Have now had another look after this
series of emails.
There claims on the strengths and weaknesses of the PIC16 series seem a bit
hit and miss though.
Weakness - pipelined instruction fetch
- cannot direct load W (accumulator) (what the heck does the MOVLW
do??)
- RETLW for code memory reads conveniently forgets that most recent
additions can do direct reads
- indirect data access requires INDF,FSR registers - well to do an
indirect access you need a register with a value to use as an address
anyway.
And then in the MAXQ strengths, they list "prefixable opcode allows a simple
means for instruction set extensions or enhancements". Anyone recall the
8086 with the FPU and 8089 enhancements to the instruction set?
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
> And then in the MAXQ strengths, they list "prefixable opcode allows a simple
> means for instruction set extensions or enhancements". Anyone recall the
> 8086 with the FPU and 8089 enhancements to the instruction set?
Gads, you mean there is someone besides me who even REMEMBERS the 8089?
Bill {getting creaky...}
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
On Thu, 26 Feb 2004 18:45:00 -0500, Bill Couture wrote:
>On Thu, 26 Feb 2004, Alan B. Pearce wrote:
>
>> And then in the MAXQ strengths, they list "prefixable opcode allows a simple
>> means for instruction set extensions or enhancements". Anyone recall the
>> 8086 with the FPU and 8089 enhancements to the instruction set?
>
>Gads, you mean there is someone besides me who even REMEMBERS the 8089?
>
>Bill {getting creaky...}
Wasn't it the 8087 FPU?
Bob {been creaky awhile...}
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
At 18.45 26/02/2004 -0500, you wrote:
>On Thu, 26 Feb 2004, Alan B. Pearce wrote:
>
>> And then in the MAXQ strengths, they list "prefixable opcode allows a simple
>> means for instruction set extensions or enhancements". Anyone recall the
>> 8086 with the FPU and 8089 enhancements to the instruction set?
>
>Gads, you mean there is someone besides me who even REMEMBERS the 8089?
Wasn't the 8089 an I/O chip?
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
At 12:55 AM 2/27/2004 +0100, you wrote:
>At 18.45 26/02/2004 -0500, you wrote:
> >On Thu, 26 Feb 2004, Alan B. Pearce wrote:
> >
> >> And then in the MAXQ strengths, they list "prefixable opcode allows a
> simple
> >> means for instruction set extensions or enhancements". Anyone recall the
> >> 8086 with the FPU and 8089 enhancements to the instruction set?
> >
> >Gads, you mean there is someone besides me who even REMEMBERS the 8089?
>
>Wasn't the 8089 an I/O chip?
The i8087 was the math co-processor for the 8086/8.
IIRC, the 80486DX (not the 80486SX) was the first Intel PC processor with
the math co-processor on-chip.
At 01:12 AM 2/27/2004 +0100, you wrote:
> > Wasn't the 8089 an I/O chip?
> >
>
>The 8089 was a smart DMA Controller which was able to follow programmed
>instruction sequences. So you could see it as a very specialized processor.
> On Thu, 26 Feb 2004 18:45:00 -0500, Bill Couture wrote:
>
> >On Thu, 26 Feb 2004, Alan B. Pearce wrote:
> >
> >> And then in the MAXQ strengths, they list "prefixable opcode allows a simple
> >> means for instruction set extensions or enhancements". Anyone recall the
> >> 8086 with the FPU and 8089 enhancements to the instruction set?
> >
> >Gads, you mean there is someone besides me who even REMEMBERS the 8089?
> >
> >Bill {getting creaky...}
>
> Wasn't it the 8087 FPU?
>
> Bob {been creaky awhile...}
No, the 8087 (80287, 80387, and clones) were the FPU. The 8089 was an
I/O chip.
Bill
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
On Thursday, Feb 26, 2004, at 15:45 US/Pacific, Bill Couture wrote:
>> Anyone recall the
>> 8086 with the FPU and 8089 enhancements to the instruction set?
>
> you mean there is someone besides me who even REMEMBERS the 8089?
>
I still have one. :-)
Intel pretty much invented the co-processor. I thought it was
pretty brilliant, actually.
BillW
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
On Thursday, Feb 26, 2004, at 16:12 US/Pacific, 0xDEADBEEF wrote:
>> Wasn't the 8089 an I/O chip?
>>
>
> The 8089 was a smart DMA Controller which was able to follow programmed
> instruction sequences. So you could see it as a very specialized
> processor.
>
It claimed to be an "IO co-processor" - used the same co-processor hooks
as the (more or less concurrent) 8087, but was focused on doing IO.
BillW
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
> At 12:55 AM 2/27/2004 +0100, you wrote:
> >At 18.45 26/02/2004 -0500, you wrote:
> > >On Thu, 26 Feb 2004, Alan B. Pearce wrote:
> > >
> > >> And then in the MAXQ strengths, they list "prefixable opcode allows a
> > simple
> > >> means for instruction set extensions or enhancements".
> Anyone recall the
> > >> 8086 with the FPU and 8089 enhancements to the instruction set?
> > >
> > >Gads, you mean there is someone besides me who even REMEMBERS the 8089?
> >
> >Wasn't the 8089 an I/O chip?
>
> The i8087 was the math co-processor for the 8086/8.
>
> IIRC, the 80486DX (not the 80486SX) was the first Intel PC processor with
> the math co-processor on-chip.
Maybe this is an urban legend or something but I seem to remember that the
486SX DID have the co-processor on die, it just wasn't enabled (that or it
failed qual and was disabled and sold as an SX). TTYL
Spehro Pefhany wrote:
> At 01:12 AM 2/27/2004 +0100, you wrote:
> > > Wasn't the 8089 an I/O chip?
> > >
> >
> >The 8089 was a smart DMA Controller which was able to follow programmed
> >instruction sequences. So you could see it as a very specialized processor.
>
> What was it used for, typically?
It was an I/O coprocessor designed for use with the 8086.
The Apricot (PC built around 1984) had one designed in but did not ship with one
installed. The story goes that when the Apricot went into production they
couldn't get a fraction of the 8089s they required so they patched it out using
a piece of wire and an inverter. Ironically the patched machine ran faster.
I think the story might be true, I got it for the guy that designed the
motherboard.
>The Apricot (PC built around 1984) had one designed in but did
>not ship with one installed. The story goes that when the Apricot
>went into production they couldn't get a fraction of the 8089s
>they required so they patched it out using a piece of wire and
>an inverter. Ironically the patched machine ran faster.
Could be the case, but I believe the 8089 was also horribly expensive for
what it did. No figures to back this up, it may just be that I was in a
backwater they did not want to supply with these things, or it may be that
this was how they determined who got supplied with the low production
numbers.
Offloading instruction sequences in in a lot of processors as a good idea that the design team put in silicon for future expansion but rarely used. I have only seen it actually used in a couple of 8 bit processors (Both 8051 based) It is in the silicon of
a couple more 8 implementations with no way to use the feature without custom bondout.
w..
"Alan B. Pearce" wrote:
>
> >Gads, you mean there is someone besides me who even REMEMBERS the 8089?
>
> Yeah, but never used one. Was always intrigued with the concept of
> offloading to a co-processor (as distinct from an FPU).
>
> --
> http://www.piclist.com hint: The list server can filter out subtopics
> (like ads or off topics) for you. See http://www.piclist.com/#topics
interlog.com Info for manufacturers: http://www.trexon.com
