Printable View
Great job, TrekkiesUnite118! Thanks for sharing those numbers with us, looks like they are floating well beyond the sheath's predictions...
Do you know which build and/or plugin of PCSX2 provides polygon counts? I tried to find it recently but failed.
I think 0.9.6 or lower will do it. But be warned though it's very buggy. Many games like Gran Turismo 4 won't even run. I was able to get GT3 to boot, but there were a lot of glitches going on. The numbers appear in the window border information. I forget the exact order but I think it's one of the middle numbers.Quote:
Originally Posted by Barone
Some of the newer builds might still feature it, I just can't figure out how to get it to display it.
Thanks! I'll check that now.
Just confirmed it is 0.9.6. Here's a picture with what I believe is the correct number pointed out:Quote:
Originally Posted by Barone
http://i102.photobucket.com/albums/m...ps23ff7900.png
I'm pretty sure this is the right number aswhen you pause the game it immediately freezes. Sorry all the numbers freeze when you pause. Though I still think that's the correct number as it's the most reactive to what's on screen. A lot of the other numbers don't fluctuate that much, even when it's obvious there's a ton of crap on screen.
For what it's worth, I would say that MGS2 looks to be within the Dreamcast's abilities as far as polygons are concerned from the numbers I'm seeing.
Thanks for the heads up.
I have a long list of games that I'd like to test but stuff like MGS2 or GT3 aren't my priorities.
If Transformers is one you want to test I can already tell you not to bother. It took 20 minutes for my 4.2GHz 8 Core CPU to load to the "Would you like to create a save file?" screen. I'm pretty sure it crashed after that.Quote:
Originally Posted by Barone
Good to know that. Eventually I'd pick that one and waste my time...
I think I'll be checking open world 3D platformers and racers first. Most of the unsourced counts and polarizing claims in the forums I've read are located in those two genres.
Yeah, a 30 fps version of F-Zero GX would really kill the comparison. No chance of a slightly more optimistic version with native pvr2 effects instead of flipper effects eh? Oh what a small world we live in. A PS2 version with lower resolution everything would be just fine though, just a sloppy port by a bad developer who didn't "get" the brilliance of all that under used potential.
The 60fps is part of what gives the game it's insane sense of speed. So yes, limiting that to 30fps would be a big deal. And you still don't get it. The polygon counts alone in F-Zero GX, not even counting transparency and lighting effects (or flipper effects as you're calling them), are more than the Dreamcast is realistically capable of. Even at 30fps we're hitting right against the upper limit of what the Dreamcast can do.Quote:
Originally Posted by sheath
While were at it, looking at nullDC footage Daytona USA appears to be hovering between 400k - 1.2 Million Polygons per second. That's 6000-20,000 polygons per frame. It's not even close to F-Zero GX.
We're not talking about a PS2 port though. I have doubts that the PS2 could handle it as well. Yeah it might do the polygons and possibly even the framerate fine, but the lighting and transparency effects would probably take a hit and it probably wouldn't run in 480p anamorphic widescreen either.Quote:
Originally Posted by sheath
Also I love how like clock work you posted right after I closed my Sega-16 tab and was no longer visible as logged on.
Screenshot: https://drive.google.com/file/d/0B0c...it?usp=sharing
IDK how accurate those numbers are given the obvious graphical errors but they seem to vary consistently with what the game is rendering.
Oh, and my old Dell laptop runs the emulator very slowly.
Time to do the math:
Burnout (2001) (Criterion Games) (PS2)
In game polygon count (by PCSX2 0.9.6): 78699 polygons per frame
About the scene: Single race at the Interstate track showing the usual traffic cars and some buildings, without the loads of special effects used in collisions and burnout/nitro mode.
Frame rate of the game: 60 fps.
Result: 60 x 78699 = 4,721,940 polygons per second ~ 4.7 million polygons per second.
sheath says:
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
The entire generation was "floating around" the 1 million mark.
@Trekkies
Maybe it could be interesting to load GC's Burnout in Dolphin emulator to double-check the polygon counts.
I gotta say that's a lot higher than I thought it would be. Might be worth throwing the Gamecube version at Dolphin and seeing how it performs. It might also help us figure out how accurate those PCSX2 numbers are if we can't see a noticeable difference in geometry.
You kinda read my thoughts... Double-checking is always good.
I had check "The Bouncer" before Burnout and the numbers were much lower though.
Supposedly, Alex Ward stated that the first Burnout game peaked at 13 million polygons per second but the only source linked to it is pretty dead and Internet archive doesn't have a copy of it.
http://www.sega-16.com/forum/attachm...4&d=1395905062Quote:
Originally Posted by Barone
Screenshot: https://drive.google.com/file/d/0B0c...it?usp=sharing
Time to do the math:
Burnout (2002) (Criterion Games) (GC)
In game polygon count (by Dolphin 4.0): 92564 polygons per frame
About the scene: Single race at the Interstate track showing the usual traffic cars and some buildings, without the loads of special effects used in collisions and burnout/nitro mode.
Frame rate of the game: 60 fps.
Result: 60 x 92564 = 5,553,840 polygons per second ~ 5.6 million polygons per second.
I didn't notice any obvious increase in the geometry complexity of the cars and buildings in this version but I have played it very briefly and some reviews of the time claimed that the car models were improved. The draw distance is better but the 3D models are the same IMO.
So I guess it's safe to say that the PCSX2 numbers are fairly reliableand maybe they were also lowered a bit more due to the emulation errors - since it's easy to notice that it was not drawing all the polygons that it should be drawing(it seems that even when it doesn't texture it right the polygons are still there, even in games with far more graphical errors).
Anyway, what did sheath say about my posts creating errors in this thread? lol.
http://i.imgur.com/a6kmM7Q.gif
So basically if the Dreamcast wanted to run Burnout at the very least it would either need to drop the frame rate to 30fps or lower the geometric detail.
I'm interested to look at Resident Evil 4 now, to see if sheath's 900k Polygons per second claim for the PS2 and 3 Million Polygons per second for the Gamecube holds any water.
To not talk about the AI cars which deal with heavy traffic and the collision model which probably would eat some extra CPU time when compared to most of the DC racers.Quote:
Originally Posted by TrekkiesUnite118
I'll try Resident Evil 4 ASAP but, you know, the success rate of this PS2 emulator is very low. Several games already failed with me.Quote:
Originally Posted by TrekkiesUnite118
Dolphin seems to be both much more compatible and way more optimized.