interesting DEC Pro stuff on eBay

[Mike Katz]

Intel has never understood interrupts or good cpu architecture.

Look at the segment:offset architecture of the 8086 and of course it's 
single interrupt (without the separate interrupt controller chip) vs the 
68000 somewhat orthogonal 32 bit architecture and 256 interrupts with 8 
levels built into the chip.

I could spend pages just describing how the 68K chip just blows away the 
8086 considering they were both released at about the same time.

For crying out loud the 6809 even though it only addresses 64K is a more 
powerful processor than the 8086.  Even with the 8086 clocking faster 
than the 6809.

On 4/21/2022 1:51 PM, Paul Koning via cctalk wrote:
>
>> On Apr 21, 2022, at 2:47 PM, Dave Mitton via cctech <cctech at classiccmp.org> wrote:
>>
>> Date: Wed, 20 Apr 2022 14:03:48 -0400
>> From: Paul Koning <paulkoning at comcast.net>
>> To: Chris Zach <cz at alembic.crystel.com>, "cctalk at classiccmp.org"
>> 	<cctalk at classiccmp.org>
>> Subject: Re: interesting DEC Pro stuff on eBay
>> ….
>>
>> ➢ That said, you'd think that DMA would make a 1:1 interleave controller much more feasible.  And Bjoren also mentioned Ethernet.  The DECNA is not to horrible without DMA because you can use its on-board memory directly as host packet buffers, though CT bus based memory is as I recall slower than motherboard memory.  Still, one wonders why they didn't use a correctly designed Ethernet chip like LANCE, either with local memory or with DMA.
>>
>> 	paul
>>
>> You’d have to ask Bill Duane about that…  He found several dynamic problems with the Intel Ethernet chipset.  He was under NDA to them at one time.
>> I suspect that the Pro team didn’t ask us and just went with that chip for some other reason.   We obviously regretted that in the long run.
> I sometimes get the feeling that the Pro team grabbed every Intel chip they could possibly use, even though every single one of them is badly designed and causes piles of problems.  For example, the interrupt structure of the Pro is a nightmare, attributable directly to the fact that is what Intel came up with.  And the 82586 implements bugs that were recognized and fixed 20 years earlier.
>
> 	paul
>