PS3 to have 50x graphics power of PS2

Shifty Geezer

uber-Troll!
Moderator
Legend
From CnVG
Speaking at a Silicon Valley press conference, nVidia president Jen-Hsun Huang confirmed that the company has already been working on the chip for an extended time and suggested that it would be in the region of 50 times more powerful than the PS2's graphics chip.
How can this be quantified when we know PS2's graphical performance was hampered by simplicity and lack of high-end features?
How does a contemporary high-end card compare? 50x PS2 doesn't seema lot to me, but with figures like 1000x being bantered around I guess I'm desensitized :oops:

Are there any Linux benchmarks that can be run on PS2 and PC to compare graphics performance?
 
Well 50 x 66 million polygons/sec = 3.3 billion polygons/sec

1920 x 1080(res) x 4(micropolys/pixel) x 60(fps) x 4 overdraw = ~1.99 billion.

Seems like doing the job to me, and that is with 1080p60(which is unlikely because of memory)
 
There a term for that PR bluster... just can think of it, oh yeah b@~@*$~S.

Like talk in game number shall we? 66 million when in reality PS2 game are probably between 10-15 million at a stretch.
 
keep in mind that he is being quoted saying PS2 graphics chip. Would this indicate again that the workload between CELL and Nvidia's part is similar to how the work is being split between the EE and GS in PS2?

50 x GS can mean many things of course... if the workload is similar and Nvidia's part will be a more complex rasterizer... Maybe it's really a reference to rendering performance? Though how indicative of the systems performance (PS3) can it be if Nvidia is just talking about their unit within the system?
 
Pugger said:
There a term for that PR bluster... just can think of it, oh yeah b@~@*$~S.

Like talk in game number shall we? 66 million when in reality PS2 game are probably between 10-15 million at a stretch.

Of course this is just PR talk, but the figure sounds reasonable... 15 million polys with PS2 * 50 = 750 million polys /sec. That is entirely possible.

A 0.065 micron part with some EDRAM and you have that performance.

Afterall Playstation 2 is rather slow/feature poor when it comes gfx. At least with today's standards.
 
so 50x10 = 500 million, and we'll be lucky to get that in game too :D
1920 x 1080(res) x 4(micropolys/pixel) x 60(fps) x 4 overdraw = ~1.99 billion.
you cut the resolution and fps in half, and you get that... there you go, that might be optimistic realism.
 
This could mean so many things, and the many things it could mean could be all wrong anyway so...

50x fillrate? 50x polygons? 50x pixel shading? Hold on, PS2 had zero pixel shading and multiplying a number to zero is zero so... TAHT MUST MEEN PSX3 HAS NO PIXELZ SHADIN'!!!111!11!!11












[Sorry, got carried away in the Trashing of Useless Thread(TM)]
 
Shifty Geezer said:
From CnVG
Speaking at a Silicon Valley press conference, nVidia president Jen-Hsun Huang confirmed that the company has already been working on the chip for an extended time and suggested that it would be in the region of 50 times more powerful than the PS2's graphics chip.
How can this be quantified when we know PS2's graphical performance was hampered by simplicity and lack of high-end features?
How does a contemporary high-end card compare? 50x PS2 doesn't seema lot to me, but with figures like 1000x being bantered around I guess I'm desensitized :oops:

Are there any Linux benchmarks that can be run on PS2 and PC to compare graphics performance?

Need directions back to the pc forums? You seem lost.
 
It's a pretty meaningless number without any context.

I'm surprised no other outlets picked up on that comment though.

But to add to the silliness, I'll add this: if you were to break the gap between PS2 and PS3 up into 6 month cycles, and if we take it that graphics cards double in power every 6 months, then PS3 should be over 1000x more powerful than PS2!!!1!one :rolleyes: ;)
 
For fun, lets Put four Graphics Synthesizers on a chip and run it at 1Ghz.

If we take the GS (1.2 Gpixels/sec textured fillrate - one texture) and multiply the number of pixel pipelines by four, then give each pipeline a proper TMU and run this hypothetical chip at 950Mhz, we would get a textured fillrate of over 60 Million pixels per second.

Thats 50 times the textured fillrate of the GS.

60/1.2=50

This satisfies the "tens of Gpixels" hopes.
According to Moore's Law, we could put 16 GSs on a chip for PS3 :rolleyes: .


I know this is a pointless calculation, but this thread is kinda strange anyway isn't it?
 
Titanio said:
It's a pretty meaningless number without any context.

I'm surprised no other outlets picked up on that comment though.

But to add to the silliness, I'll add this: if you were to break the gap between PS2 and PS3 up into 6 month cycles, and if we take it that graphics cards double in power every 6 months, then PS3 should be over 1000x more powerful than PS2!!!1!one :rolleyes: ;)


1024 times actually, if we assume a 5 year difference. ;) !!!onehundredeleven!!11!
 
No Nvidia + Sony = 10.
See you add up all the letters and you get ten.
Try to get the full effect of the PR will you. :p
 
Xenus said:
No Nvidia + Sony = 10.
See you add up all the letters and you get ten.
Try to get the full effect of the PR will you. :p

Oh they'll have lots of material.
"MS will NV Sony..." for example... :LOL:
 
How does a contemporary high-end card compare? 50x PS2 doesn't seema lot to me
Well you can spin it many ways.
For example:
50xGS = 120GPixel/s.
Current PC cards are ~6-8GPixel

*gives ERP a naughty wink* :LOL:
 
Back
Top