• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

New ATI Ruby Demo at SIGGRAPH running on X360...1M Polys per frame!!!

K

Kleegamefan

K. LEE GAIDEN
http://www.theinquirer.net/?article=25197

The guy in charge of the demo told us that the graphic part is much more powerful than even the upcoming R5XX[EDIT:Oh teh SHYTE!!], series and that ATI's desktop unit will match Xbox 360 graphics with a next generation scheduled for next year.
Click to expand...

The Rubi model was rendered with 70,000 polygons while some of its elements had as many as 120,000 polygons per character. The most demanding scene had the dreamlike number of one million polygons per scene.
Click to expand...

xboxwire.jpg


xboxwire1.jpg


Xenos/C1 is more powerfull than R520???

THAT is impressive, IMO...

Depending on the frame rate, 1M polys per frame puts it in the 20-60Mpps range....pretty impressive IMO....
 
If ATI continues with the R300 architecture and it takes the way of Nvidia about the number of Shaders (8+24) I believe that the R520 will be less powerful than C1.
 
So that's what, 30 million polys/sec at 30fps? Weren't we supposed to get that this gen? :P I'm not impressed by this number.
 
crunker99 said:
Well nothign is shaded, what do you expect. Ill bet the xbox1 could push a pretty amazing wireframe demo.
Click to expand...

The demo isn't wireframe, they just switched to wireframe to show off the polygons
 
*looks at pic and squints*
xboxwire1.jpg

Metroid Prime 3 on Xbox360! Nintendo = MS 2nd party!*





*
may or may not be true...
 
llTll said:
not that it matters but .. how many polys can the RSX pull?
Click to expand...


Do you really want to go there?

Nightbringer said:
If ATI continues with the R300 architecture and it takes the way of Nvidia about the number of Shaders (8+24) I believe that the R520 will be less powerful than C1.
Click to expand...

The ATI guy already saic C1>R520....

We won't get something as powerful as C1 till R600 and even then, it won't have access to a memory bus as fast as the X360 nor a CPU as powerful as XeCPU...

This thing is powerful...
 
llTll said:
dude., its a question. not a war. i really want to know if there is any info about RSX and how many poly it can push.
Click to expand...

then start your own thread on RSX or do a fucking search where that debate has already been done a thousand times. No need to turn this thread into a c1 vs rsx thread.
 
sangreal said:
then start your own thread on RSX or do a fucking search where that debate has already been done a thousand times. No need to turn this thread into a c1 vs rsx thread.
Click to expand...


uhh no? i dont want to start any thread. i just want to know. i promise lol :lol

EDIT:- i also dont want to debate too..

on the other hand. how many polys PS2 and Xbox can do?

wasnt the PS2 something like 60 million with no effects on or something???
 
op_ivy said:
isnt this the same e3 demo?
Click to expand...

yeah, its the same r520 demo they showed at E3. Maybe updated a little, who knows

Personally, I don't find the demo to be particularly impressive. The ATi comments are more interesting than the demo.
 
op_ivy said:
isnt this the same e3 demo?
Click to expand...

Yes it is, although this article is from Siggraph seemingly.

I do wonder about R520 =/ I'm not bang up to date with my rumours, but last I heard it may not be everything we thought it might. But then it also may not be so "conventional". I was convinced pre-E3 that the R520 should be at least better than Xenos on paper, in terms of raw shading power, but that may not even be the case anymore.
 
gofreak said:
Yes it is, although this article is from Siggraph seemingly.

I do wonder about R520 =/ I'm not bang up to date with my rumours, but last I heard it may not be everything we thought it might. But then it also may not be so "conventional". I was convinced pre-E3 that the R520 should be at least better than Xenos on paper, in terms of raw shading power, but that may not even be the case anymore.
Click to expand...


Yeah, latest rumors put R520 and G70 about equal performance wise...

R520 should have been out even before G70 but that project has been problematic for ATI to say the least...
 
R520 seesm to be in some sort of development hell. Also, I do question whatever The Inquirer states. They are like the PC version of s p o n g.
 
crunker99 said:
Well nothign is shaded, what do you expect. Ill bet the xbox1 could push a pretty amazing wireframe demo.
Click to expand...

:lol

you DO know that those wireframe shots are overlayed on top of the fully lit, shaded, highlighted, animated scene in realtime, right? I mean, cause, otherwise, a person could look like a total moron if . . . you know . . . they made a comment on it without realizing that . . .

Edit: Here's an example of the old X800 ruby demo (the first one) from a year or so ago.

atix800_screen002.jpg

Ruby8.jpg

Ruby2.jpg
 
crunker99 said:
Well nothign is shaded, what do you expect. Ill bet the xbox1 could push a pretty amazing wireframe demo.
Click to expand...

perhaps you should read the article:
"He was able to change to the wireframe model by pressing a button on a controller."
 
Depending on the frame rate, 1M polys per frame puts it in the 20-60Mpps range....pretty impressive IMO....
Click to expand...
At E3 it was running at 30FPS with no AA, so 30MPPS peak... That number alone is certainly not impressive when C1 is supposed to be able to render ~500Mpps with non trivial shaders. The whole thing (the demo) does look highly detailed and impressive though.
 
I would like to see that exact same demo again now knowing the numbers behind it.
 
I'm not impressed with 1m per scene. Thats not 1m per frame, as many polys will be behind the camera/obscured by other objects and culled. Even 1m per frame isn't *that* impressive.

What kind of figures do the tech guys expect from next gen? 5m per frame @ 60fps is 300m. Is that acheiveable? Would you even have the memory to store that number (probably around 10m per scene?)
 
n3mo_toad said:
you DO know that those wireframe shots are overlayed on top of the fully lit, shaded, highlighted, animated scene in realtime, right? I mean, cause, otherwise, a person could look like a total moron if . . . you know . . . they made a comment on it without realizing that . .
Click to expand...
:lol :lol
 
Ruby's model is more akin to what I was expecting out of Team Ninja for DoA4.

Sorry for bringing the subject, but I think she looks really great. That said, DoA4 does look quite good.
 
This is greater scene complexity than any game on the market...

Yes, next gen games should even have more complexity but 30-60Mpps is not bad...thats alot of detail....
 
Rogue Squadron 2, which came out 4 years ago and was made in 8 months had around 20-25 million polygons per second. This is not impressive.
 
littlewig said:
Rogue Squadron 2, which came out 4 years ago and was made in 8 months had around 20-25 million polygons per second. This is not impressive.
Click to expand...

More like 12-15 million from what I recall (I dunno, maybe you have figures somewhere), which puts it at about 250,000 polgons per frame...but yeah, I agree with you that that's not a very impressive increase.
 
Mama Smurf said:
More like 12-15 million from what I recall (I dunno, maybe you have figures somewhere), which puts it at about 250,000 polgons per frame...but yeah, I agree with you that that's not a very impressive increase.
Click to expand...


I think you're right, I might be confusing my number with Rogue Squadron 3. But these numbers can be achieved with current generation hardware, so it's not really impressive.
 
littlewig said:
Rogue Squadron 2, which came out 4 years ago and was made in 8 months had around 20-25 million polygons per second. This is not impressive.
Click to expand...

LINK

Newduck: To satisfy all the forum spec and number adorers, how many million polygons does Rogue Leader tend to push in a standard situation per second, and how much does it tend to vary? and has there been any performances increases since previous showings

RLT: If we would start counting the polygons now the game wouldn't get done, but we estimtae most scenes at between 12 and 15 million polygons per second. The version being show in Europe has quite a performance increase in the Hoth level compared to what was shown previously.
Click to expand...

Let's keep it honest...
 
littlewig said:
I think you're right, I might be confusing my number with Rogue Squadron 3. But these numbers can be achieved with current generation hardware, so it's not really impressive.
Click to expand...


30-60Mpps on current hardware??

I don't think so....


This demo is doing that with a *lot* of shader effects too...

Yeah, its not mind blowing but current gen it is not...



There were no quote I can remember concering 20-25Mpps with RS3 and I follow this shit alot..

The highest current gen pps number I have heard was from a (sadly) canceled Xbox racer being programmed by ERP, who posts @ B3D and used to post here too..

That was 30Mpps (an amazing number, I admit) but I don't think it was that pixel shader intensive of a game, either..

This Ruby demo has tons of shader effects (bump, specular,diffuse, Anisotropic filtering, etc.) along with that 1M number...

I have the movie of it on my HDD, trust me, current gen couldn't touch it....shit...it is one of the better Next gen demos I've seen...
 
Pseudo judo said:
LINK



Let's keep it honest...
Click to expand...


littlewig said:
I think you're right, I might be confusing my number with Rogue Squadron 3. But these numbers can be achieved with current generation hardware, so it's not really impressive.
Click to expand...

Let's keep up to date, tool. Ingore List you go.
 
Kleegamefan said:
30-60Mpps on current hardware??

I don't think so....


This demo is doing that with a *lot* of shader effects too...

Yeah, its not mind blowing but current gen it is not...
Click to expand...

The polygon count can ceetainly be matched, but the effects certainly can't. I'm just pointing out that the polygon is not what is impressive, it's the effects and texture work that also hogs a lot of resources that make it impressive.
 
thanx for the heads up. Kleegamefan


but I agree that 1M polygons/scene is not very impressive .... at max, that would be, as you say '30-60Mpps' .... but, AFAIK, there are already Xbox games that push ~20 million pps .. granted, these Xbox games are definitely running less complex shaders than this Ruby demo is (and as you mention, a lot of other effects, as well), but still I also agree that this 1M polygons/scene is not very impressive..







as has been stated (in this thread: http://www.ga-forum.com/showthread.php?t=57200&page=3 ) someone claimed that in PGR3, the "Manhattan Bridge" is "1 million polygons" .. which is more impressive, IMO :D
 
¿The inquirer says xenos is more powerful than R5xx? I bet R5xx is more powerful than xenos then X-D

I still remember... two years back in time i read an article that stated that by 2006 intel would release its monocore processors at 6.4GHz, and those would be running at 11GHz by 2007 :lol :lol :lol
 
Wunderchu said:
thanx for the heads up. Kleegamefan


but I agree that 1M polygons/scene is not very impressive .... at max, that would be, as you say '30-60Mpps' .... but, AFAIK, there are already Xbox games that push ~20 million pps .. granted, these Xbox games are definitely running less complex shaders than this Ruby demo is (and as you mention, a lot of other effects, as well), but still I also agree that this 1M polygons/scene is not very impressive..







as has been stated (in this thread: http://www.ga-forum.com/showthread.php?t=57200&page=3 ) someone claimed that in PGR3, the "Manhattan Bridge" is "1 million polygons" .. which is more impressive, IMO :D
Click to expand...


Yeah, but unlike what mrklaw said, I believe this is 1M per frame (yes, that probably inclues occuled polys in the scene too) and not just 1M in the entire level, if I am explaining myself correctly...

Yeah, PGR3 will probably have similar levels of scene complexity too, but did you see the movie of the ATI demo??

Current gen, my ass...
 
Kleegamefan said:
Yeah, but unlike what mrklaw said, I believe this is 1M per frame (yes, that probably inclues occuled polys in the scene too) and not just 1M in the entire level, if I am explaining myself correctly...

Yeah, PGR3 will probably have similar levels of scene complexity too, but did you see the movie of the ATI demo??

Current gen, my ass...
Click to expand...
yes, I have seen the footage of the demo.. yes it looks very nice :D

this new R520 Ruby demo is about the level of graphics that I expect from next gen. games ....... the only thing not so impressive about this this new R520 Ruby demo, IMO, is the number of characters on screen (even the original R420 Ruby demo has more characters on screen) ... however, we all know from the footage of N3: Ninety Nine Nights , Heavenly Sword , Kameo , etc. that the next gen consoles will have no problem with many characters on screen, so this point is kinda' moot :D
 
Marconelly said:
At E3 it was running at 30FPS with no AA, so 30MPPS peak... That number alone is certainly not impressive when C1 is supposed to be able to render ~500Mpps with non trivial shaders..
Click to expand...

It will not be rendering 500m pps with "next-gen" level shaders. What's your definition of "non-trivial"? I've seen a dev (gu)estimate that it could maybe sustain that kind of polygon rendering with Xbox1-level shaders...but I doubt too many next-gen games will be in a position to go to that level.

The idea of that kind of setup rate, imo, is that it's more than you could need such that it isn't a bottleneck.
 
Kleegamefan said:
Yeah, but unlike what mrklaw said, I believe this is 1M per frame (yes, that probably inclues occuled polys in the scene too) and not just 1M in the entire level, if I am explaining myself correctly...

er

The Rubi model was rendered with 70,000 polygons while some of its elements had as many as 120,000 polygons per character. The most demanding scene had the dreamlike number of one million polygons per scene.
Click to expand...

thats the quote from your first post, which is why I mentioned it. If its 1m per frame, thats still not impressive, but better than 1m per scene.


Still, can someone like Faf work out the approximate storage required for a scene of say, 5m polys - just the vertex data. I bet its a lot.
Click to expand...
Click to expand...
 
You must log in or register to reply here.

Similar threads

No Rest for the Wicked has sold 1M copies in early access
2
65
292
ClanOfNone replied
Good news everyone! Futurama Hit & Run mod demo released.
8
286
Cakeboxer replied
A new Xbox-era Blizzard aims for more one or two major game launches per year
News 
20
224
March Climber replied
I am fking jonesin' for the Steam Frame
24
169
Dorago replied
Konami producer Motoi Okamoto wants to release around one new Silent Hill per year
Opinion Game Dev  2
79
513
Cyberpunkd replied
Top Bottom