16-bit & WinNT?
Richard Tsai
On a Pentium Pro-200, does 16-bit software run as fast under Win NT 4.0 as under Win95? Or is it faster/slower?
Thanks in advance.
Richard Tsai
Jason Johnston
Richard Tsai (ts...@medicine.wustl.edu) wrote:
: On a Pentium Pro-200, does 16-bit software run as fast under Win NT 4.0 as under Win95? Or is it faster/slower?
:
: Richard Tsai
Under WinNT 4.0 it is NOT backwards compatible.. It's FAT 32.. NOT FAT
16..
--
---
_ _/|
\'o.O" Jason M. Johnston | Internet: john...@visi.com
={____}= St. Paul, MN. 55102 | IRC: jmj
U | http://www.visi.com/~johnston
| ftp://ftp.visi.com/~johnston
Felix says, "Vector! The #1 Internet Provider in the Twin Cities Area!"
Andrew Ansell
Jason Johnston wrote:
>
> Richard Tsai (ts...@medicine.wustl.edu) wrote:
> : On a Pentium Pro-200, does 16-bit software run as fast under Win NT 4.0 as under Win95? Or is it faster/slower?
> : Thanks in advance.
> :
> : Richard Tsai
>
> Under WinNT 4.0 it is NOT backwards compatible.. It's FAT 32.. NOT FAT
> 16..
Which is one way of saying 16bit programs don't work under NT. 16bit
stuff under 95 is no faster on a p5 or p6
Andrew
Sawaddi
There are many variables that come into play but generally 16 bit apps run
faster under 95. See the Intel web site for the facts
Richard Tsai <ts...@medicine.wustl.edu> wrote in article
<32A994...@medicine.wustl.edu>...
Sawaddi
PS: That said, don't waste your money on a P5. Go with a P6
durgin
Richard Tsai wrote:
>
> On a Pentium Pro-200, does 16-bit software run as fast under Win NT 4.0 as under Win95? Or is it faster/slower?
> Thanks in advance.
>
> Richard Tsai
Well, based on Mr. Johnston's incorrect answer to this thread I guess I
have to provide a more accurate answer.
Firstoff, there are two factors to consider. Hardware-wise, the P-6 is
optimized for 32 bit code. It stumbles on 16 bit and boggs down. This
is regardless of which O/S you use. Despite this bumbling about, a
200MHz P-6 running NT with adequate RAM is probably faster than if
running Win 95. Not enough so's you'd really care, though.
Second, NT is designed to be hardware independant (the whole Government
GOSIP/POSIX deal). NT want you to write applications to the WIN32 API
as a hardware interface and is perturbed if you don't. Win95, on the
other hand, while supporting and largely implementing Win32 API, retains
backward compatibility with old 16 bit Windows. In this respect it is
more resiliant than NT, and will gracefully degrade and run 16 bit
software even if they do nasty things like trying to directly address
hardware features. Again, it's kind of a toss-up. Win95 will run more
of the antique games and programs, whereas NT will run the newer true 32
bit software faster.
Now as to the question of file systems, Mr. Johnston's one liner is
certainly confusing, and off-topic relative to the speed question which
was posed. But to clarify any confusion it might have raised, please
note the following:
FAT 16 is the old DOS & Win3.11 file system we've all grown to know and
become very aggravated with, in these days of multi-gigabyte drives.
The NT File System (hence NTFS) does not have the cluster size or drive
size limitations which existed with FAT16. Instead of a maximum
theoretical drive size of 2GB, NT can handle a volume of 16 petabytes
(well, a really big number anyway, equivalent to the GDP of a small 3d
world country). FAT32 is a new file system which is coming out now with
the latest release of Win95 as part of OSR2 (OEM Service Release 2). It
solves the problems with large drives that FAT16 had.
However. NTFS and FAT32 are not compatible; Win95 FAT32 partitions
can't be recognized by NTFS, and NTFS partitions can't be recognized by
FAT32. Further, I don't believe the older (i.e. current) Win95 with
only FAT16 can recognize a FAT32 partition without repartitioning the
drive. And of course MS-DOS and Win3.11 can't see any of this. Perhaps
dumb on MS's part, but that's the way it is.
If anyone would like to further clarify or correct, I'm sure everyone,
myself included, would benefit.
Power
Andrew Ansell wrote:
>
> Jason Johnston wrote:
> >
> > : On a Pentium Pro-200, does 16-bit software run as fast under Win NT 4.0 as under Win95? Or is it faster/slower?
> > : Thanks in advance.
> > :
> > : Richard Tsai
> >
> > 16..
>
> Which is one way of saying 16bit programs don't work under NT. 16bit
> stuff under 95 is no faster on a p5 or p6
> Andrew
I think older software will run under NT as long as it doesn't try and
access the hardware direct, which a lot of older DOS and Windows
software will do.
--
Power
.................................
This is only the opinion of Power.
Power does not bare any malice
towards any individual or corporation.
They do not represent the policy of
anybody else or any other corporation.
They do not represent the policy
of the Internet Service Provider.
If they do resemble the policy or
opinion of another individual
or corporation that is merely
coincidental.
Power
durgin wrote:
>
> Richard Tsai wrote:
> >
> > On a Pentium Pro-200, does 16-bit software run as fast under Win NT 4.0 as under Win95? Or is it faster/slower?
> > Thanks in advance.
> >
> > Richard Tsai
>
Wow, couldn't have said it better myself.
Glenn M
test
J. David Bryan
On Sat, 07 Dec 1996 17:02:41 +0000 in article <32A9A3...@cam.ac.uk>,
asa...@cam.ac.uk wrote...
>
>Which is one way of saying 16bit programs don't work under NT.
This is not correct. I run 16-bit programs (Windows 3.x and DOS) under NT
every day. The only requirement is that they not attempt to access the
hardware directly, but rather operate through the Win16 API or DOS Int21
interfaces.
Can't offer an opinion on the P6 question.
--
-- Dave Bryan
dbr...@mail.bcpl.lib.md.us
Bill Cohane
J. David Bryan wrote:
> I run 16-bit programs (Windows 3.x and DOS) under NT
> every day. The only requirement is that they not attempt to access the
> hardware directly, but rather operate through the Win16 API or DOS Int21
> interfaces.
> Can't offer an opinion on the P6 question.
The computer magazines are now saying that a P6-200 will outperform
a P5-200 for even 16 bit programs. I am not trying to start an
arguement, just reporting what I read in PC Magazine and PC World.
Robert Thorne
durgin wrote:
>
> If anyone would like to further clarify or correct, I'm sure everyone,
> myself included, would benefit.
In the interest of more confusion I offer the following. My wife has a
shareware recipe program that ran fine under Win3.11. Under Win95 it
blows up when trying to print. Under NT4.0 it runs fine. Go figure.
aur...@iamerica.net
Bill Cohane wrote:
>
> The computer magazines are now saying that a P6-200 will outperform
> a P5-200 for even 16 bit programs. I am not trying to start an
> arguement, just reporting what I read in PC Magazine and PC World.
The PC Magazines are very adept at promoting the next technology carrot
that that hangs in front of us. How much more power is really useable
for the average user 23 mips versus 56 mips, so what. As long as my
computer does what I want it to do in a reasonable time frame that's
fine for me. These orchestrated releases of newer and better computing
hardware is engineered to maximize profit taking, not technology.
IMHO
Vince
aur...@iamerica.net
Sorry for the typo, "that that" should have been "that Intel"
Vince
Michael Mattox
aur...@iamerica.net wrote:
>Bill Cohane wrote:
>>
>> The computer magazines are now saying that a P6-200 will outperform
>> a P5-200 for even 16 bit programs. I am not trying to start an
>> arguement, just reporting what I read in PC Magazine and PC World.
>The PC Magazines are very adept at promoting the next technology carrot
>that that hangs in front of us. How much more power is really useable
>for the average user 23 mips versus 56 mips, so what. As long as my
>computer does what I want it to do in a reasonable time frame that's
>fine for me. These orchestrated releases of newer and better computing
>hardware is engineered to maximize profit taking, not technology.
Using your PC for Quicken and Doom is fine, but some people use them
for high end uses like Multimedia, Video, Programming, etc. and not
only NEED every bit of speed they can get but need every bit they
are paying for!
Michael Mattox
mat...@advweb.com
Advanced Web Internet Adveritisng & Consulting
http://www.advweb.com
Steve Harkness
Power <powe...@micron.net> wrote in article <32A9BC...@micron.net>...
> durgin wrote:
> >
> > Richard Tsai wrote:
> > >
> > > On a Pentium Pro-200, does 16-bit software run as fast under Win NT
4.0 as under Win95? Or is it faster/slower?
> > > Thanks in advance.
<snip>
> > If anyone would like to further clarify or correct, I'm sure everyone,
> > myself included, would benefit.
>
> Wow, couldn't have said it better myself.
>
> --
> Power
Yes, great summary. I'd add that NT can run 16-bit programs in their own
address space so if you're used to running 3 or 4 older programs
simultaneously and alt-tabbing between them, NT will help in performance as
well as stability.
Bill Sabalburo
Robert Thorne wrote:
>
> durgin wrote:
> >
>
> > If anyone would like to further clarify or correct, I'm sure everyone,
> > myself included, would benefit.
>
> shareware recipe program that ran fine under Win3.11. Under Win95 it
> blows up when trying to print. Under NT4.0 it runs fine. Go figure.
The big question is: does your wife cook good?
The second question is: did you like her recipe?
just wondering....
Bill
Matt Smith
Wow, Isn't it cool to post a one line reply with an eleven line .sig
M
Philip MacCabe
Robert Thorne wrote:
>
> durgin wrote:
> >
>
> > If anyone would like to further clarify or correct, I'm sure everyone,
> > myself included, would benefit.
>
> In the interest of more confusion I offer the following. My wife has a
> shareware recipe program that ran fine under Win3.11. Under Win95 it
> blows up when trying to print. Under NT4.0 it runs fine. Go figure.
What recipes does it have in it??? hmmmmmm
maybe 95 is has a food critique feature too ... I wonder
durgin
Matt Smith wrote:
>
> Wow, Isn't it cool to post a one line reply with an eleven line .sig
>
> M
Actually, Power's sig is a reasonable ~400 bytes. Mine, on the other
hand, is 25 lines and 1700 bytes. I just remove it for prevention of
disease in the bandwidth impaired.
mailto:dur...@durgin.net
Tae Song
durgin wrote:
>
> Richard Tsai wrote:
> >
> > On a Pentium Pro-200, does 16-bit software run as fast under Win NT 4.0
> > as under Win95? Or is it faster/slower?
> > Thanks in advance.
> >
>
> Well, based on Mr. Johnston's incorrect answer to this thread I guess I
> have to provide a more accurate answer.
>
> Firstoff, there are two factors to consider. Hardware-wise, the P-6 is
> optimized for 32 bit code. It stumbles on 16 bit and boggs down. This
> is regardless of which O/S you use. Despite this bumbling about, a
> 200MHz P-6 running NT with adequate RAM is probably faster than if
> running Win 95. Not enough so's you'd really care, though.
>
> Second, NT is designed to be hardware independant (the whole Government
> GOSIP/POSIX deal). NT want you to write applications to the WIN32 API
> as a hardware interface and is perturbed if you don't. Win95, on the
> other hand, while supporting and largely implementing Win32 API, retains
> backward compatibility with old 16 bit Windows. In this respect it is
> more resiliant than NT, and will gracefully degrade and run 16 bit
> software even if they do nasty things like trying to directly address
> hardware features. Again, it's kind of a toss-up. Win95 will run more
> of the antique games and programs, whereas NT will run the newer true 32
> bit software faster.
I disagree. NT, even 4.0, still has a lot more overhead to deal with
than 95. Of course, ideally Windows 95 would need to be running all
32-bit apps and using 32-bit drivers, but it should run 32-bits apps
faster than NT.
Johannes Ullrich
Tae Song wrote:
>
>
> I disagree. NT, even 4.0, still has a lot more overhead to deal with
> than 95. Of course, ideally Windows 95 would need to be running all
> 32-bit apps and using 32-bit drivers, but it should run 32-bits apps
> faster than NT.
>
NT-Magazine (www.ntmagazine.com?), December issue.
I don't have the issue with me right now. Bat a quick summary:
They used a P-Pro 200 and ran a 32 bit Application based benchmark
(SysMark?). With 16 MB Ram, Win 95 wins. With 32 and more Win NT 4.0
wins.
Win NT 3.51 is slower than 95 at 16 MByte but faster at 32 MB.
EDO memory shows a 2 % (yes only 2%) performance increase.
For anything more than 32 MByte speed gains diminish, particularly for
Win 95.
--
Johannes Ullrich --- jull...@xos.com --- http://www.xos.com
------------
X-Ray Optical Systems Inc. | RR2 Box 409 B | work phone: (518) 442
3362
90 Fuller Rd. | Averill Park NY | home phone: (518) 674
0023
Albany NY 12205 | 12018 | FAX: (518) 442
5292
Edmond
Tae Song wrote:
>
> I disagree. NT, even 4.0, still has a lot more overhead to deal with
> than 95. Of course, ideally Windows 95 would need to be running all
> 32-bit apps and using 32-bit drivers, but it should run 32-bits apps
> faster than NT.
Win95 GDI is still 16-bit (last I heard). It must then thunk all those
32-bit calls down to their 16-bit equivalents. NT4, while it has some
overhead, generally has a more robust graphics subsystem capable of out
performing Win95. Win95 would probably do better with DirectDraw and
Direct3D I take it. For OpenGL based programs, there are no ICD or MCD
drivers for Win95 thus NT takes that one by a long shot (and the
software based rendering is 3X faster on NT). On 16 bit appliations, NT
actually thunks the Win16 API calls to 32 bits. That would mean a
performance improvement on a Pentium PRO (comparing 16-bit performance
on a Pentium PRO running NT against a Pentium PRO running Win95 is
probably a wash because of that). All in all, one would think that NT
would out perform Win95 running a fully Win32 application (no 16-bit
parts in it). On a Pentium PRO, that's a gimme. All this takes for
granted that you have enough available ram for each test.
--
Edmond Underwood
U Software
Bench32 1.21 - Now available at the Bench32 web site.
Bench32 web site: http://rainbow.rmii.com/~underwoe/bench32.html
Richard C. Rauch
aur...@iamerica.net wrote in article <32AB3F...@iamerica.net>...
> Bill Cohane wrote:
>
> >
> > The computer magazines are now saying that a P6-200 will outperform
> > a P5-200 for even 16 bit programs. I am not trying to start an
> > arguement, just reporting what I read in PC Magazine and PC World.
Would be great to hear if anyone knows in which issue of PCMag this
appeared; page # would be even better.
> The PC Magazines are very adept at promoting the next technology
> carrot that [Intel] hangs in front of us. How much more power is
really
> useable for the average user 23 mips versus 56 mips, so what. As long
> as my computer does what I want it to do in a reasonable time frame
> that's fine for me. These orchestrated releases of newer and better
> computing hardware is engineered to maximize profit taking, not
> technology.
>
> IMHO
> Vince
If _PC_Magazine_ (as opposed to "the PC magazines") says the P6-200
outperforms the P5-200, even with 16-bit apps, the claim is likely
backed up by benchmark testing, the results of which probably appear in
the article. Especially if said benchmark is their Winstone, I'm going
to take their word for it. As far as I know, nothing comes closer to
testing the "real world" performance of a PC. 'Course, I wouldn't mind
if they'd add a "games" suite to it. They're part of our real world,
too. ;-)
--
Rich Rauch http://users.aol.com/richrauch/
-=- Not affiliated with any entity or product mentioned above. -=-
==> Not a spokesman for Rockwell, its suppliers, or customers. <==
G Tarayan
machines. You will notice that Pros are at least same as Pentiums in Win95 but
mostly even better. Greg