9 Dec
2005
9 Dec
'05
6:35 p.m.
Subject: Re: Legacy apps in Windows/OS X was Re: Old MS-DOS & Win Software
From: "Chuck Guzis" <cclist at sydex.com>
Date: Fri, 09 Dec 2005 09:54:34 -0800
To: cctalk at classiccmp.org
On 12/9/2005 at 7:35 AM Allison wrote:
I added why it was faster. For the V20 in a XT clone it was easily 10% and
for the V30 is was more.. Really nice parts. I'm generally not into 8088
but I've kept some of the older and newer V20s that have crossed my path
for "maybe one day". The faster parts could scream. I kept a Tandy
1000hx because it was odd and used the V20.
The other differences were internal effective
address calculation was
handled
differntly from 8088 which shaved a few cycles of the execution. For the
same clock V series were between 5-10% faster. Also the V20/V30 had 8080
emulation.
Hmmm, thought I just said that. One of the more highly-touted performance
improvements in the V20/V30 was the implemenation of dual internal data
busses. In practice, I don't know if it made a huge difference in speed,
but NEC claimed up to a 30% improvement for som operations.. FWIW, my reference manual for the 70108/70116 is dated
August 1985 and
identifies both as "Low Power CMOS"; no mention otherwise of any other fab
technology. NEC did later offer the -H series variants (V20H, V30H, etc.)
which featured fully static operation and lower power consumption than the
V20/V30--and you could get them in clock speeds of up to 16 MHz.
Those were the HCMOS process the change was mostly for speed.
I've got a reasonably thick folder in my files
titled "Natick" with a bunch
of correspondence with the NEC folks there on the whole subject. The V40,
which was sort of an 80188 in CMOS (PGA, QFP and PLCC pacakges) also
claimed to emulate the 8080 instruction set, but I never got hold of one to
try it out (or I didn't care to).
Likely from my old office and boss. By 85 it was CMOS before that it was
NMOS. Even the first NMOS parts were really hybrid and much lower in power
than the HMOSII from intel.
Between 82 and 85 NEC was working on CMOS designs real hard and the first
non-propritory 8 bit parts were 80C35. I have a few of the first samples
with the page boundary bug, fixed in production.
The V40/50 parts were still futures in 83 when I left. But looked interesting.
I thought it interesting that the 8080 emulation of the
V20/V30 emulated
the Intel 8080A and not tne NEC 8080A instruction set.
Cheers,
Chuck
Well the Intel 8080A and the NEC8080AF are the same. NEC discontinued the
incompatable 8080 version back around '78, they got burnt bad. The V20
implemented the 8080AF as bad memories were not lost. I have both
versions. Taught them a lesson in compatability that wasn't forgotten.
Even the 8085 and D780(Z80) execute the same undocumented instructions.
On the other hand there is a sorta improved Z80 like thing NEC did, the
ucom7800 series and they are interesting. Looks Z80 like but the
instruction set is anything but. I have a bunch of the PIGGYback
(78PG11) and romless 7800 parts.
Allison