Anyway I'll just cut and paste the info then, basically I was just clarifying the DC PowerVR uses 32x32 tiles, the Neon250 a derivative of the DC PowerVR with a higher clock but lower spec part uses 32x16 tiles.
Just to clarify some of the (key) differences between The PC and console versions of the series 2 HW,
CLX tile was 32x32, PC part was 32x16.
CLX support alpha test with HW front->back sorting to deliver massive effective fillrates when it was use, this was not available on the PC part.
CLX tiling was completely handled by HW, the PC part was 50:50 split between HW and SW
CLX include latency buffering that allowed VQ and palettised to run at full rate, this was removed from the PC part resulting in 50% performance for those formats.
CLX support 2 or 3 (think it was effectively 3 wasn't it Simon ?) external 32 bit memory busses (think in terms of much hyped NV/ATi crossbars), The PC part utilised a single 64 bit bus resulting much greate page break impact.
This all adds up to the PC part being between 50-75% of the performance of the console part inspite of the higher clock rate (125 vs 100 MHz)...
...The differences between CLX and PMX1/Neon250 were all to do with minimising chip area.
You will have to PM Baloo to get it approved, otherwise it will stay in limbo. Thanks for the info, I assumed the two systems would be roughly identical. Good to know.
I'm just glad MSR & Ecco exist since they totally look proper 6th gen. They really show the Dreamcast could compete if used well. It never got the chance to show it's true power, such a shame. A part of me is happy about it though since I really like the original Xbox and it probably would never have had a chance as a 4th competitor.
Last edited by Kamahl; 09-18-2012 at 03:33 PM.
This thread needs more... ENGINEERS
Something else I don't know how to calculate is the difference between frames per second and polygons per frame. We saw earlier that the Dreamcast would absolutely not have enough VRAM for 5 million polygons per second texture mapped and lit at 30 frames per second (FPS). But at 60FPS the same amount would leave over 2MB for textures. My experience is very few games this generation even achieved a solid 30FPS, and even fewer managed 60FPS, most of which are on the Dreamcast or are graphically unimpressive games.
So, had it lasted, would the Dreamcast have been able to push ~4 million polygons per
frameper second at 60FPS? I am not sure how people are arithmetically figuring how many polygons could be generated per frame versus per second. I would really keen on finding out how these numbers are arrived at.
Last edited by sheath; 09-18-2012 at 05:55 PM.
For polygons per frame if you have a number of polygons per second just divide that by your frame rate and that tells you how many polygons per frame. To get that number from a game you would need to use an emulator or something similar that can count give you a polygon count per frame.
4 Million polygons per frame would be completely unrealistic. It would be able to do it but at only 1 Frame Per second at best. 4 Million polygons per frame at 60 fps would be 240 Million Polygons per second. I don't even think the PS3 can do that.
PS2 has God Hand, so it automatically wins that generation.
Right, that wasn't what I was asking. Nothing does millions of polygons per frame as far as I know. The polygons per frame are found by dividing the polygons per second by the frames per second. What I want to know is how to look at the bandwidth or processors or RAM speed or whatever and figure out how many polygons per frame can actually be generated.
Oh well for that you simply multiply the polygons per frame by 40 to find out how many bytes that takes up, and then see how much space that takes up in VRAM.
So for example, 4 million polygons per frame comes out like this:
4 million * 40 Bytes = 160 Million Bytes
160 Million / 1024 = 156250 KB
156,250 / 1024 = ~152MB
So the Dreamcast doesn't even have the memory to render that many polygons in one frame.
With this info though some things concerning the Saturn do puzzle me somewhat. According to specs I've looked up the Model 2 board could do 300,000 polygons per second at 60fps. If Daytona USA was maxing that out that's around 5000 polygons per frame. Saturn Daytona runs at 60,000 polygons per second at 20fps, which is 3000 polygons per frame. Since the Saturn can do 200,000 textured polygons per second at apparently 60fps that's again 3000 polygons per frame. So why couldn't Saturn Daytona USA run at 60fps? Memory shouldn't be an issue as it's using the same amount of memory as it does at 20fps.
I didn't mention 4 million polygons per frame did I? It
might have been a typodoh definitely was a typo. 4 million polygons per second at 60FPS was Gran Turismo 4's supposed peak performance. The mass quote I made earlier showed that the Dreamcast has more VRAM to handle that than the PS2 does. That would be 66,667 polygons per frame. What I want to know is if the SH4+FPU could generate enough polygons per frame to sustain that. Increasing the framerate was no simple task, and still isn't as indicated by the top games of this generation.
Also, Sega GT is easily my favorite car RPG. The gameplay and music are great even though the graphics are merely adequate. I have Gran Turismo 2 Bleemcast! and Sega GT blows the balls off of that. I owned GT3 Aspec for a while and absolutely could not stand it, pretty or no and it looked no better than F355 Challenge.
Had Super GT been released on Dreamcast we would not even be having this conversation, and "9/11" wouldn't have happened.
Last edited by sheath; 09-18-2012 at 08:17 PM.
I don't know any accurate specs, I'm just going with the 150 million count they threw around when the machine came out. Assuming that as an imaginary peak, it still only needs to do 1/4th of it to reach 1 million polys per frame assuming 30 fps.
There are currently 1 users browsing this thread. (0 members and 1 guests)