[comp.sys.ibm.pc] V20 vs. 8088: Is there any useful difference?

jfy@phobos.cis.ksu.edu (Joseph F. Young S89) (07/16/89)

Looking through some magazine advertisements, I found
the NEC V20 chip which was described in the ad as
giving a 10 to 40% increase in performance as compared to
the Intel 8088.  If these claims are indeed true, I will
probably get one for my system.  
Here's my current system configuration:
       XT clone motherboard with 640K ram (Datatek)
       8088 and 8087 running at 4.77/10 MHz
       DTK/Erso Bios
       CGA/monochrome card
       I/O card with joystick port, serial port, clock, and
           floppy controller.
       Seagate 238R with RLL controller.

Now my question is: does anyone have any experience with the
V20 as far as software/hardware compatability is concerned?
Would it be worth my money to install one in my system?
Any comments, suggestions, etc. are welcome.
Please reply by email whereever possible.

--
Joseph F. Young  |Bitnet: cbm2064@ksuvm
                 |Inet  : cbm2064@ksuvm.ksu.edu / jfy@phobos.cis.ksu.edu
-----------------|UUCP  : ...!rutgers!ksuvax1!phobos.cis.ksu.edu!jfy 
      "I'm sorry, Master, but that WAS the backup system."--Slave        

ked@garnet.berkeley.edu (Earl H. Kinmonth) (07/16/89)

In article <2305@deimos.cis.ksu.edu> jfy@phobos.cis.ksu.edu (Joseph F. Young S89) writes:
>Looking through some magazine advertisements, I found
>the NEC V20 chip which was described in the ad as
>giving a 10 to 40% increase in performance as compared to
>the Intel 8088.  If these claims are indeed true, I will
>probably get one for my system.  

I cannot speak to the V20. I do have considerable experience with the
V30 (replacement for the 8086). First, the punch line: for a very minor
expenditure, you get a MAJOR improvement.

Now, the details. I was running SCO Xenix on an ATT 6300 (XT clone with
8086 processor). Changing to a V30 allowed at least one more process to
run without making the machine too sluggish to tolerate. Screen updates
were MUCH more rapid.

You will note that there is no V40 (or whatever) to replace the 80286.
This is a classic case of a company using intellectual property rights
to protect its incompetence. NEC was doing such a good job of
correcting the brain-dead features of Intel processors that Intel took
NEC to court and succeeded in stopping NEC innovations.

When you read all the hype and bull shit about the 80386 and 80486,
keep in mind that if Intel had not had heavier weight legal council
than NEC, some of the features of these chips might have been here
earlier, "Made in Japan."

Usually I am a defender of American firms, but as far as I am
concerned, Intel is the GM of the chip world -- using marketing and
legal power to make up for a desperate shortage of common sense.

jca@pnet01.cts.com (John C. Archambeau) (07/17/89)

Correction, there is a V40, it has an architecture similiar to the 80286, but
it is not machine language compatable with the 80286.  I don't know what NEC
had to gain by designing such a beast, but they did.  The V40 also has a real
V30 mode, etc.  If NEC did what I think they did, I believe you don't have to
reset the V40 to get it in and out of protect mode as you have to with a real
80286.  I would have loved to have seen the V40 as a 100% 80286 replacement,
but unfortunately it is not.

 /*--------------------------------------------------------------------------*
  * Flames: /dev/null (on my Minix partition)
  *--------------------------------------------------------------------------*
  * ARPA  : crash!pnet01!jca@nosc.mil
  * INET  : jca@pnet01.cts.com
  * UUCP  : {nosc ucsd hplabs!hd-sdd}!crash!pnet01!jca
  *--------------------------------------------------------------------------*/

chasm@attctc.DALLAS.TX.US (Charles Marslett) (07/17/89)

In article <2305@deimos.cis.ksu.edu>, jfy@phobos.cis.ksu.edu (Joseph F. Young S89) writes:
> Looking through some magazine advertisements, I found
> the NEC V20 chip which was described in the ad as
> giving a 10 to 40% increase in performance as compared to
> the Intel 8088.  If these claims are indeed true, I will
> probably get one for my system.  

VERY FEW programs will show a 10% improvement in speed -- expect a range of 2%
to 10%.  I would never get a V20 (or V30 if the box has an 8086) to speed it up
(unless you want to recode the stuff you are running).

> Here's my current system configuration:
>        XT clone motherboard with 640K ram (Datatek)
>        8088 and 8087 running at 4.77/10 MHz
>        DTK/Erso Bios
>        CGA/monochrome card
>        I/O card with joystick port, serial port, clock, and
>            floppy controller.
>        Seagate 238R with RLL controller.
> 
> Now my question is: does anyone have any experience with the
> V20 as far as software/hardware compatability is concerned?

I have had no trouble at all running a fairly varied selection of software
(before permanently installing it  I brought home the full compatibility
test library at work -- demos and eval copies of anything I might ever want
to use, and several that I have no interest in running again).  NOTHING failed.

As a side comment:  OS/2 will not run on a V20 (not too very suprisingly), but
Excel did run fine.  I decided not to buy a copy, not because it failed in my
tests, but because it is rather slow running Windows/286 and Excel on a 4.77
MHz box.  At 10 MHz with a fast RLL controller (any of the 1:1 types), you may
actually get useful work done.  I would suspect that any program that does
not REQUIRE extended memory, but requires a 286 will probably (emphasis here!)
run on a V20.  Certainly, you can compile code for a 286 using Microsoft
C, and it will compile and run with no problems.  That is the main reason
I changed over to the V20.

> Would it be worth my money to install one in my system?
> Any comments, suggestions, etc. are welcome.

As I said, real mode 286 code runs fine on a V20 -- that can be enough reason,
if it is not, I would forget the V20.

> Please reply by email whereever possible.
> --
> Joseph F. Young  |Bitnet: cbm2064@ksuvm
>                  |Inet  : cbm2064@ksuvm.ksu.edu / jfy@phobos.cis.ksu.edu
> -----------------|UUCP  : ...!rutgers!ksuvax1!phobos.cis.ksu.edu!jfy 
>       "I'm sorry, Master, but that WAS the backup system."--Slave        


Sorry, but I though this was a general enough question, and a short enough
answer, that it warranted posting.  (here come 'de flames... here come 'de
flames)

Charles Marslett
chasm@attctc.dallas.tx.us

ho@fergvax.unl.edu (Tiny Bubbles...) (07/17/89)

From article <26367@agate.BERKELEY.EDU>, by ked@garnet.berkeley.edu (Earl H. Kinmonth):
> In article <2305@deimos.cis.ksu.edu> jfy@phobos.cis.ksu.edu (Joseph F. Young S89) writes:
>>Looking through some magazine advertisements, I found
>>the NEC V20 chip which was described in the ad as
>>giving a 10 to 40% increase in performance as compared to
> I cannot speak to the V20. I do have considerable experience with the
> V30 (replacement for the 8086). First, the punch line: for a very minor
> expenditure, you get a MAJOR improvement.

I email'ed the original poster, but since that's the second email I've sent 
on the subject, I'll waste bandwidth.

I have had great luck with the V20.  It gives a 10-20% real improvement, and
really does wonders for graphics applications.

The V20 speeds the heck out of integer math and memory operations.  And it's
cheap.  I say, you can't beat it for the buck.
---
	... Michael Ho, University of Nebraska    <ho@fergvax.unl.edu>

plim@hpsgpa.HP.COM (Peter Lim) (07/17/89)

Well, I happened to have used a V20 for about a year 3 years ago.
I think the bottom line is that running normal software (i.e.
those not custom made to use V20 string handling instructions),
you get about 5% performance improvement. No big deal, but then
I was paying only A$25 for one (beggars can't complain, right ?).
I can't even remember the software I used to confirm the 5%
performance improvement (I think it was something written in
Turbo Pascal). From what I heard, the insignificant performance
improvement is due to the fact that the 8088 bus in the PC/XT is
already very congested (somewhere around 80-90 % duty cycle), so
you can't push much more data through. Not so for the V30 due
to the more under-utilized 8086 bus.

I hope the above helps. I would put a V20 in if I am using an XT.
The cost is almost zero and sloooooooowwwww XT could certainly
use any speeding up it can get (what happens to the good-old-days
when an XT with CGA is very good, and EGA is heaven sent ?  :-)).


Regards,
Peter Lim.
ICDS,  HP Singapore
(* EVERY WORD HERE HAS NOTHING TO DO WITH MY EMPLOYER *)

mbb@cbnewsh.ATT.COM (martin.b.brilliant) (07/17/89)

From article <8673@attctc.DALLAS.TX.US>, by chasm@attctc.DALLAS.TX.US (Charles Marslett):
> In article <2305@deimos.cis.ksu.edu>, jfy@phobos.cis.ksu.edu (Joseph F. Young S89) writes:
>> the NEC V20 chip .....
> 
> VERY FEW programs will show a 10% improvement in speed -- expect a range of 2%
> to 10%.....

I replied by mail, but my comment might be of more general interest:

How do you know you have a v20?  The itemized bill for my XTurbo clone
says I have an 8088, but Norton SI says I have a V20.  If I hadn't run
SI, I wouldn't know I had a V20, and then if I replaced it with a V20,
I wouldn't get the spectacular imporovement other users say they get....

> ....  (here come 'de flames... here come 'de > flames)

Well, how about cold water?  :-)

M. B. Brilliant					Marty
AT&T-BL HO 3D-520	(201) 949-1858
Holmdel, NJ 07733	att!hounx!marty1 or marty1@hounx.ATT.COM

Disclaimer: Opinions stated herein are mine unless and until my employer
	    explicitly claims them; then I lose all rights to them.

ray@ole.UUCP (Ray Berry) (07/17/89)

> Correction, there is a V40, it has an architecture similiar to the 80286, but
> it is not machine language compatable with the 80286.  I don't know what NEC
> had to gain by designing such a beast, but they did.  The V40 also has a real
> V30 mode, etc.  If NEC did what I think they did, I believe you don't have to
> reset the V40 to get it in and out of protect mode as you have to with a real
...
   There is no similarity between the V40 & 286 other than the instruction
set.  The V40 does NOT support protected mode nor address more than 1 meg. 
It is simply a V20 with a bunch of peripheral circuity built in.  It is 
similar in intent to the 80188 Intel part.   

-- 
Ray Berry  kb7ht uucp: ...{sumax|thebes}!ole!ray   COMPU$ERVE: 73407,3152 
Seattle Silicon Corp. 3075 112th Ave NE. Bellevue WA 98004 (206) 828 4422

lance@helios (Lance Bresee) (07/18/89)

I am curious...
can a V30 or V40 be used to replace 
   1: an 8088
   2: an 8086

leonard@bucket.UUCP (Leonard Erickson) (07/19/89)

The V20 typically gives only around a 5-10% increase in performance.
But in some applications it does a lot better. Considering that a
10 MHz V20 might set you back all of $10-15, I'd say it's worth it.

There are programs that use the V20's 8080 mode to run CP/M, and
other programs use the other "extra" instructions that it has.
(I don't have a manual for it, but I am told that it has most of
the 286 instructions available)

-- 
Leonard Erickson		...!tektronix!reed!percival!bucket!leonard
CIS: [70465,203]
"I'm all in favor of keeping dangerous weapons out of the hands of fools.
Let's start with typewriters." -- Solomon Short

ked@garnet.berkeley.edu (Earl H. Kinmonth) (07/19/89)

In article <8401@saturn.ucsc.edu> lance@helios.ucsc.edu (Lance Bresee) writes:
>
>
>I am curious...
>can a V30 or V40 be used to replace 
>   1: an 8088
>   2: an 8086

8088 ==> V20
8086 ==> V30

scott@hpcvca.CV.HP.COM (Scott Linn) (07/20/89)

/ hpcvca:comp.sys.ibm.pc / ho@fergvax.unl.edu (Tiny Bubbles...) /  4:56 pm  Jul 16, 1989 /

>I have had great luck with the V20.  It gives a 10-20% real improvement, and
>really does wonders for graphics applications.
>
>The V20 speeds the heck out of integer math and memory operations.  And it's
>cheap.  I say, you can't beat it for the buck.
>
>	... Michael Ho, University of Nebraska    <ho@fergvax.unl.edu>
----------

I agree.  I replaced the 8088 in my HP150 computer with a V20 (cost: $12).
I noticed quite a performance increase with microsoft windows, and
other graphics oriented programs (integer arithmetic?).

Overall, about 10-20%.

Scott Linn

pechter@scr1.UUCP (Bill Pechter) (07/20/89)

In article <8401@saturn.ucsc.edu> lance@helios.ucsc.edu (Lance Bresee) writes:
>
>
>I am curious...
>can a V30 or V40 be used to replace 
>   1: an 8088
>   2: an 8086

The V20 is the 8088 replacement, the V30 is the 8086 replacement.
The V30 runs fine in my AT&T PC6300 under both Dos and Xenix-86.
I have had no problems with the chip - and  the MS-Windows screen updates seem
noticeably faster. (About 10-15%).  My SI is now 4.2 (after putting in Qfresh
to tighten up the refresh times).



-- 
Bill Pechter -- Home - 103 Governors Road, Lakewood, NJ 08701 (201)370-0709
Work -- Concurrent Computer Corp., 2 Crescent Pl, MS 172, Oceanport,NJ 07757 
Phone -- (201)870-4780    Usenet  . . .  rutgers!pedsga!tsdiag!scr1!pechter
  **   MS-DOS is CP/M on steroids, bigger bulkier and not much better  ** 

mdfreed@ziebmef.uucp (Mark Freedman) (07/23/89)

(I don't believe you have to reset the V40 to get in and out of protected mode)
  
    That's certainly true. The V40 doesn't HAVE protected mode :-)
   
     I believe that it's the NEC-enhanced equivalent of an 80186. The V20 / V30
also implement some of the 186 / 188 instructions (in effect, 286 real-mode)
such as PUSHA and POPA  (push all registers / pop all registers), which allows
software which DOESN'T have to be portable across various platforms to be 
slightly more efficient (smaller, faster, etc.).
    I have a V20 in my original PC (had problems with the bug in the original
8088 mpu), and consider it $12 well spent.

hjg@amms4.UUCP (Harry Gross) (07/27/89)

In article <8401@saturn.ucsc.edu> lance@helios.ucsc.edu (Lance Bresee) writes:
>
>
>I am curious...
>can a V30 or V40 be used to replace 
>   1: an 8088
>   2: an 8086

I don't know about the V30 or V40, but when I replaced my 8088 with the V20, I
saw a very real improvement in performance.

While on the topic - I have heard that there are additional instructions
present on the V20 - anyone know anything about them?

Cheers,
-- 
		Harry Gross				 |  reserved for
							 |  something really
Internet: hjg@amms4.UUCP   (we're working on registering)|  clever - any
UUCP: {jyacc, rna, bklyncis}!amms4!hjg			 |  suggestions?

lance@helios (Lance Bresee) (07/27/89)

More curiosity....
does anyone know a source for a 10MHZ V20?
I can find 8MHZ versions for $8, from DIGIKEY..$12 seems a bit high.
I would prefer 10MHZ, and would need a chip data sheet as well..

Thanx muchly in advance for all replies.

lance%helios.UCSC.EDU

jca@pnet01.cts.com (John C. Archambeau) (07/30/89)

Try Jameco.  If anybody has it, they do.

 /*--------------------------------------------------------------------------*
  * Flames: /dev/null (on my Minix partition)
  *--------------------------------------------------------------------------*
  * ARPA  : crash!pnet01!jca@nosc.mil
  * INET  : jca@pnet01.cts.com
  * UUCP  : {nosc ucsd hplabs!hd-sdd}!crash!pnet01!jca
  *--------------------------------------------------------------------------*/

rwp@cup.portal.com (Roger William Preisendefer) (07/30/89)

JDR Microdevices has 10MHz V20's for 11.95 (I bought one for an XT clone.)

Roger Preisendefer