3DO: How many polygons per second ?
Dustin Cushman
of?
Dustin--
||
Dustin Cushman ||
Rcus...@vms1.gmu.edu __//__ I'd rather be excavating
Geroge Mason University \ /
\ /
\/
Jim Totaro
said they had the technical manuals... can anyone
confirm this?
Jim
Czeslaw Szymanski
: Does anyone know how many polygons per second the 3DO m1 is capble
: of?
According to DHGF, it can do 40,000 polys per sec. This sounds
reasonable, but this might be flat-shaded polygons, or it might be just
plain wrong. It's the only number I've heard.
--
Marcin Szymanski
mar...@execpc.com
alex werner
current systems like the PSX, the M2 (of course), and maybe even the
jag, does not really have hardware for doing polygons. Instead, it has
the cel engine, which basically can write images to the screen warped
any way you want really really fast. The cel engine makes the texture
mapping pretty easy, but there's no built in support for 3-d
transformations or anything. So figuring out where all the polygons go
on the screen is hard. Actually putting them there, since they're all
just warped cels, is really fast.
In any case, I don't know an accurate figure off the top off my head.
Suffice it to say that the M1 can do enough polys per second to run NFS
and WC3, and what more could we want? :)
Shawn Rader
: >I heard 25,000 texture mapped PPS... from someone who
: >confirm this?
: Hmmm, I think you are missing a zero there, no? -- 250,000 t-mapped pps?
No way! That would be more than the PSX, and Saturn! The
highest rate shown on the PSX to date is Tekken at 120,000 pps. 25,000pps
sounds right for texture mapped polys, and 40,000 sounds about right for
flat unshaded polys. The new Nintendo ads say their FX2 chip can do
20,000 per second which is comparible to a 3DO, so if Nintendo isn't
exagerating too much the polygon count would be around 25,000pps on the 3DO.
--
-Shawn Rader
Quentin Mohos
>I heard 25,000 texture mapped PPS... from someone who
>said they had the technical manuals... can anyone
>confirm this?
>
>Jim
Hmmm, I think you are missing a zero there, no? -- 250,000 t-mapped pps?
--
Quentin Mohos mo...@remus.rutgers.edu
The "Q" stands for Quality.
David Nagy
: Jim Totaro <alt...@peak.org> writes:
: >I heard 25,000 texture mapped PPS... from someone who
: >said they had the technical manuals... can anyone
: >confirm this?
: Hmmm, I think you are missing a zero there, no? -- 250,000 t-mapped pps?
In your (and our) dreams! :) This the M1 we're talking about, perhaps
you missed the beginning of the thread.
I'll have to sit down and do a more careful job sometime, but I
guessimate that NFS does something less than 15,000 polys per second.
I'd guess that Crash and Burn does the most although I haven't played
much of Total Eclipse, and it's in the running.
I think 25k polys/sec is a reasonable "max" rate.
Dave Nagy
Jim Coyle
> Does anyone know how many polygons per second the 3DO m1 is capble
> of?
I wouldn't use polygon per second rates to judge which system has more
powerful. A company can twist that figure around anyway they'd like.
The 3DO-M1 is slightly less powerful than the Saturn.
Quentin Mohos
> No way! That would be more than the PSX, and Saturn! The
>highest rate shown on the PSX to date is Tekken at 120,000 pps. 25,000pps
>sounds right for texture mapped polys, and 40,000 sounds about right for
>flat unshaded polys. The new Nintendo ads say their FX2 chip can do
>20,000 per second which is comparible to a 3DO, so if Nintendo isn't
>exagerating too much the polygon count would be around 25,000pps on the 3DO.
Really? Oops, sorry. I'd thought though that M2 was supposed to be doing
something like 750,000 polys persecond, and the psx ~500,000. so that's
why I thought that 25,000 figure for M1 must have been really low. Did I
mistakenly add and extra zero to all of the above figures then? (hhmmm..
or maybe I was thinking about the pixels per second thing...?)
David Nagy
: On 24 Jun 1995, Dustin Cushman wrote:
: > Does anyone know how many polygons per second the 3DO m1 is capble
: > of?
: I wouldn't use polygon per second rates to judge which system has more
: powerful. A company can twist that figure around anyway they'd like.
It's a handy yardstick for making rough comparisons. (Certainly better
than how many "bits" it has.) :)
Luckily, most companies "twist" this figure as you say, to give the
greatest possible poly/sec number. Since all companies bloat their
figures semi-equally (a lot), one CAN make some worthwhile conclusions
from them.
You can not however translate these figures directly into the kind of
"game performance" you should be able to expect. I guess you could if you
knew the "theorectical to actual" conversion factor. :) I tend to chop
the benchmark in half, and then work from there.
The situation is muddied by the fact that an "M2 polygon" will look
better than a "PSX polygon". (Perhaps better texturing, shading,
light-sourcing, etc...)
: The 3DO-M1 is slightly less powerful than the Saturn.
I'd guess perhaps half as powerful, but that's just a guess. Compared to
the M2, all current systems are practically identical in performance.
Wow, what a concept. :)
DaveNagy
Chris Curry
mo...@remus.rutgers.edu (Quentin Mohos) wrote:
|>Really? Oops, sorry. I'd thought though that M2 was supposed to be doing
|>something like 750,000 polys persecond, and the psx ~500,000. so that's
|>why I thought that 25,000 figure for M1 must have been really low. Did I
|>mistakenly add and extra zero to all of the above figures then? (hhmmm..
|>or maybe I was thinking about the pixels per second thing...?)
The PSX does not do 500,000 polys per sec I think its more like 350,000
and those are just plain flat shaded polygons. The M2 will be able to
so over a million of these type of polygons and when you add effects and
so forth (like texture mapping, etc) it goes down to 750,000. The PSX
I think goes down to around 250,000 when you add effects and even that may
be too high of a figure.
I think 3DO is around 50,000 polys/sec and the Saturn is around 90,000
From last I heard.
Chris C
------------------------------------------------+
In order to understand one must learn how others
see the world and then learn how not to see the
world how one wants it be (S)
------------------------------------------------+
kryp...@iwaynet.net
I think you need to bolt your head down straight, dude. If 3do could do 250,000 textured pps the psx and Saturn wouldn't compare(wi=
th a mere 360 textured polies).
Argonaut software sez Nintendo's FX2 chip can do 20,000 pps which is comparable to 3do(of course, the dummies didn't factor in CD te=
chnology and textures, which make 3do WAY better). I'd say 25-30,000 pps, but that's plenty for FIFA soccer to run at 60fps(does it=
? It should).
Of course, I'm not a programmer so if you are you can step in and tell me what 3do can really do, but I think a game that can do 25=
,000pps+ dolby surround is worth buying.
MAC 3do blaster?
David Nagy
: Really? Oops, sorry. I'd thought though that M2 was supposed to be doing
: something like 750,000 polys persecond, and the psx ~500,000. so that's
: why I thought that 25,000 figure for M1 must have been really low. Did I
: mistakenly add and extra zero to all of the above figures then? (hhmmm..
: or maybe I was thinking about the pixels per second thing...?)
No, I think the above zeros are all in the right places. (Although that
PSX figure sounds a tad too high.) If one went by all the hype about the
M2 being "7 times faster" that the PSX, and one used the M2's untextured
poly rate, (over a million) you'd get a figure for the PSX of about 150k
per second.
That's pretty bass-ackwards way of arriving at a PSX performance figure,
so how about this: The 3DO can do about 30K (in a game), the Saturn
appears to be perhaps 3 times faster. That gives us about 100k per
second. The PSX is reportedly quite a bit faster than that. Perhaps
200-300k per second?
The world may never know...
Dave Nagy
Wayne Magor
mo...@remus.rutgers.edu (Quentin Mohos) writes:
>sha...@teleport.com (Shawn Rader) writes:
>
>> No way! That would be more than the PSX, and Saturn! The
>>highest rate shown on the PSX to date is Tekken at 120,000 pps. 25,000pps
>>sounds right for texture mapped polys, and 40,000 sounds about right for
>>flat unshaded polys. The new Nintendo ads say their FX2 chip can do
>>20,000 per second which is comparible to a 3DO, so if Nintendo isn't
>>exagerating too much the polygon count would be around 25,000pps on the 3DO.
>
>something like 750,000 polys persecond, and the psx ~500,000. so that's
>why I thought that 25,000 figure for M1 must have been really low. Did I
>mistakenly add and extra zero to all of the above figures then? (hhmmm..
>or maybe I was thinking about the pixels per second thing...?)
That PSX number is totally wrong. Developers I talked to at the E3 show
were saying they were actually seeing 90,000 polygons per second on the
PSX and were very happy with that.
Wayne.
Mike Sone
Dustin Cushman <rcus...@vms1.gmu.edu> wrote:
>Does anyone know how many polygons per second the 3DO m1 is capble
>of?
>
>
> ||
> Dustin Cushman ||
> Rcus...@vms1.gmu.edu __//__ I'd rather be excavating
>Geroge Mason University \ /
> \ /
> \/
Dustin,
The great polygon mystery strikes again. Why are you wondering about the
number of polygons per second the M1 is capable of doing?
First off polygons per second is a relative term used by developers to
determine the limits for their realtime rendering. Systems like the 3DO
M1 is quite capable of handling nearly as many polygons as the other systems.
For example, PF Magic used nearly 80,000 polygons per second in the game
Paatank. These polygons were moved in the X, Y and Z axis and included
those used for the live video on the screens. These polygons are texture
mapped and use alot of colors.
Other systems like the Sony Playstation been doing about 90,000 texture
mapped and Gouraud shaded polygons per second. Without Gouraud shading
the Playstation will do alot more. However, the norm for the Playstation
seems to be around 90,000 polygon per second for polygons in the x,y and
z axis.
From what I know from various developer, the 3DO is easily capable of
50,000 polygon per second and can be pushed to 80,000 polygons per second.
Aloha,
Mike Sone
Mike Sone
David Nagy <dave...@netcom.com> wrote:
>I'd guess that Crash and Burn does the most although I haven't played
>much of Total Eclipse, and it's in the running.
>
>I think 25k polys/sec is a reasonable "max" rate.
>
> Dave Nagy
Hey Dave,
This is no flame, but how did you arrive at this approximation? I'm just
curious cause the 3DO development kit of old more than doubles your figure.
At 24 frames per second that puts about 1,000 or so on screen at one time.
Well in any case the games will show the just how limited the count is or
if having more will make for a better game. NFS seemed to be doing alot
as well as Road Rash.
Aloha,
Mike Sone
Brian Yee
: >I heard 25,000 texture mapped PPS... from someone who
: >said they had the technical manuals... can anyone
: >confirm this?
: >
: >Jim
: Hmmm, I think you are missing a zero there, no? -- 250,000 t-mapped pps?
: --
Brian Yee
: mo...@remus.rutgers.edu (Quentin Mohos) writes:
: >sha...@teleport.com (Shawn Rader) writes:
: >
: >> No way! That would be more than the PSX, and Saturn! The
: >>highest rate shown on the PSX to date is Tekken at 120,000 pps. 25,000pps
: >>sounds right for texture mapped polys, and 40,000 sounds about right for
: >>flat unshaded polys. The new Nintendo ads say their FX2 chip can do
: >>20,000 per second which is comparible to a 3DO, so if Nintendo isn't
: >>exagerating too much the polygon count would be around 25,000pps on the 3DO.
: >
: >Really? Oops, sorry. I'd thought though that M2 was supposed to be doing
: >something like 750,000 polys persecond, and the psx ~500,000. so that's
: >why I thought that 25,000 figure for M1 must have been really low. Did I
: >mistakenly add and extra zero to all of the above figures then? (hhmmm..
: >or maybe I was thinking about the pixels per second thing...?)
AHem 1 million for M2
Paul Rance
:
:
: I second that. I talked with developers at E3 as well and repeatedly
: heard the 90,000 number. It seems the theoretical number is 300,000
: (or at least that is what Sony is saying) but most developers were
: seeing there actual games only move 90,000. This could go up with time
: and more learning but applying the same logic to M2's 1 million or
: 750,000 then we should still see a quantum leap above the PS even at
: the same 30% reduction in true to theoretical power.
:
: Erin
: --
: Erin A. Fritz
: email: fr...@iastate.edu
I wonder if the fact that the PSX doesnt have Z-Buffering has reduced
the amount of polygons in games, without this the machine must surely
be having to use more processing power to move polygons around. Any experts
know if this is the case?
Paul
--
.-----------------------------------------------------------------.
!Email pa...@rance.demon.co.uk 2:254/516.2@Fidonet !
! !
! WWW page is http://metro.turnpike.net/P/paulr/index.html !
`-----------------------------------------------------------------'
Erin A Fritz
>were saying they were actually seeing 90,000 polygons per second on the
>PSX and were very happy with that.
>
>Wayne.
I second that. I talked with developers at E3 as well and repeatedly
heard the 90,000 number. It seems the theoretical number is 300,000
(or at least that is what Sony is saying) but most developers were
seeing there actual games only move 90,000. This could go up with time
and more learning but applying the same logic to M2's 1 million or
750,000 then we should still see a quantum leap above the PS even at
the same 30% reduction in true to theoretical power.
Erin
--
Erin A. Fritz
email: fr...@iastate.edu
"We've got 200 woody-inducing titles to play on it." 3DO
____DISCLAIMER______________________________________________________
|The opinions expressed above are mine and in no way represent those|
|of my employer, my school, my country, my wife, my political party,|
|myself, or my cat. |
|___________________________________________________________________|
David Nagy
: David Nagy <dave...@netcom.com> wrote:
: This is no flame, but how did you arrive at this approximation? I'm just
: curious cause the 3DO development kit of old more than doubles your figure.
: At 24 frames per second that puts about 1,000 or so on screen at one time.
This was just a wild-ass guess. :) Basically, I approximated the max
number of polys on screen at once, (500-1000?) and multiplied by the
framerate. (In C & B, about 17 FPS.)
I might be underestimating the poly-count, but I don't think I'm TOO far
off. (It's possible thought) :)
Someday... I'll actually count those suckers. :)
I take those "factory benchmark" numbers with a grain of salt. MY PC
will do almost 100k polys a second... in a graphics demo. I've never
seen a game that even approaches this.
(My 486'66 seems pretty comparable to my 3DO in terms of polygon
performance in games. Pretty impressive considering the price and CPU
differences...)
Dave Nagy
David Nagy
: I second that. I talked with developers at E3 as well and repeatedly
: heard the 90,000 number. It seems the theoretical number is 300,000
: (or at least that is what Sony is saying) but most developers were
: seeing there actual games only move 90,000. This could go up with time
: and more learning but applying the same logic to M2's 1 million or
: 750,000 then we should still see a quantum leap above the PS even at
: the same 30% reduction in true to theoretical power.
Uh... WHAT 30% reduction in true-to-theoretical power? I usually figure
50% of the "quoted" speed should be possible. According to your above
figures the Sony is suffering a 70% drop from its "hoped for" numbers.
If the M2 suffered a similar fate, it would be able to do "only" 225,000
polys/sec. (fully textured, etc.) Still, this is an order of magnitude
faster than the M1--with much better quality. And 2.5 times the speed the
PSX has (so far) achieved.
My hunch is that the M2 will fulfill more of its potential than that though.
Dave Nagy
Erin A Fritz
Yeah, me too. What I was saying (and I wasn't clear enough, I admit) was
that even if the REAL number of polys the M2 moves drops to a similar
30% reduction of claimed specs we should still see a clear difference
in the M2 over the PS. I, however, don't beleive the number of polys
will go down that much from the 750,000 number. I believe that is why
3DO gave two numbers: 1mil and 750,000. The 1 mil number is like Sony's
300,000 to 360,000 number, theoretical simple polys. But the 750,000
is more like Sony's 90,000, real life polys in the game with textures.
This is what I think. What I was saying above was just speculation that
M2 would still kick PS's ass even if 3DO would have fudged on the number
like Sony did.
John Capra
numbre of polygons with all of the additional hardware features on them
(textured, shading,alpha channel, etc.). If you want to talk straight
virt. fighter one plain polygons, the number is one million. The M2 is
going to be simply incredible. I have a feeling that it will take time
for developers' imaginations to catch up with the possibilities of this
system, the initial games will be only shadow of what's actually possible.
-
JOHN CAPRA VJH...@prodigy.com
Albert Abramson
I.e., three decades out, we can now see that every vendor seemed to be
exaggerating their poly figures. We now know that:
system textured/shaded flat-shaded
PS1 160k (*120k actual) 360k
Saturn 180k(<100k actual) 500k (using the VLIW-like DSP)
3DO 20k (50k actual?) 25k
3DO M2 700k 1 million
Jaguar 10k
Jaguar 2 900k 2.5 million
32X 180k 50k
Sega CD 3,000
Sega SVP 9,000 20,000
Nintendo 64 180k 500k
Voodoo 1.5 million
Voodoo 2 3 million
Dreamcast 3-6 million
PS2 75 million (3-20 million actual games)
PS3 3,800 million (est.)
*Note: Due to poor software libraries, most PS1 launch era titles averaged
5,000-20,000 pps.
Sad that the Jaguar 2 (nearly completed when this thread was running) was
never released. Backward compatible with the 50 or so Jag titles (plus about
50 more that were never completed), ATARI might've had a shot at a real
comeback.