By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Nintendo - AMD Hinting that they will get the win for Nintendo's next handheld?

phaedruss said:

Anyways, if you're telling me you can't get more out a chip by changing the design to be more efficient I think you're either lying or full of it.

I'm not saying that you can't get more out of an architecture, I'm just saying that the gains for having a better microarchitecture are minimal compared to chip scaling ... 



Around the Network
fatslob-:O said:
phaedruss said:

You just told me they can lol.

How so ? By "architecture" you are likely referring to the "instruction set architecture" or the "micro architecture" which is entirely different from a "transistor architecture". Having a more mature process node does not mean that you solve the leakage issues. 


By architecture I mean the design of the chip. By maturing I mean better manufacturing methods and possibly different materials if that is needed although I'm guessing those materials you listed would be very expensive at least at first.



fatslob-:O said:
phaedruss said:

Anyways, if you're telling me you can't get more out a chip by changing the design to be more efficient I think you're either lying or full of it.

I'm not saying that you can't get more out of an architecture, I'm just saying that the gains for having a better microarchitecture are minimal compared to chip scaling ... 


Again, I'm not sure that's true. I don't know all of the history, but I'm positive there have been substantial gains in performance with the same process node.



Soundwave said:

70 gflops is pretty good for a device made in 2011. By 2016, Nintendo should be able to get chips 3-4x more powerful I would imagine, which does put XBox 360-like graphics in the relam of possibility at a lower resolution without too much fuss IMO. 

I think they will go with a higher resolution screen, but such a device will be a multi purpose device and the Wii U/360 like graphic games won't neccessarily be the norm. So y'know there's nothing wrong with running a game that has Wii-level graphics at full 720p or something. Or for web browsing and other apps a higher res screen is beneficial and these days cheap. If a developer wants to make a game that runs at 480x270, there's nothing stopping them. 

The vita has waaay less than 70 Gflops. Try 35 Gflops ...

Your right that there may not be something wrong with running games running 720p at WII level graphics however the guys with the scaled up console versions are getting the short end of the stick as improvements would be minimal so I think it's best that Nintendo stick with a resolution of 480x270 with 4x MSAA to push for the best graphics quality plus non gaming tasks aren't important.



http://en.wikipedia.org/wiki/GeForce_600_series
http://en.wikipedia.org/wiki/GeForce_200_series

They were able to go from 1 gigaflop/ watt to 5 gigaflop/watt on 40 nm. They probably got higher on 40 nm but that's just looking at it quickly.



Around the Network
phaedruss said:

By architecture I mean the design of the chip. By maturing I mean better manufacturing methods and possibly different materials if that is needed although I'm guessing those materials you listed would be very expensive at least at first.

@Bold Which means you weren't referring to the design of the transistor themselves ...

That's not exactly how it sounded to me when you were referring to "maturing" cause you made it sound like as if it was about improving current manufacturing. 



phaedruss said:

Again, I'm not sure that's true. I don't know all of the history, but I'm positive there have been substantial gains in performance with the same process node.

Over the past 40 years in the chip manufacturing industry if we counted all the gains from better architectures it would amount to 50x better performance compared to chip scaling giving us 3500x better performance. 



fatslob-:O said:
phaedruss said:

By architecture I mean the design of the chip. By maturing I mean better manufacturing methods and possibly different materials if that is needed although I'm guessing those materials you listed would be very expensive at least at first.

@Bold Which means you weren't referring to the design of the transistor themselves ...

That's not exactly how it sounded to me when you were referring to "maturing" cause you made it sound like as if it was about improving current manufacturing. 

Which would include using different materials would it not? Chip design can increase performance though. I don't know what you mean by transistor design exactly. A transistor is a transistor is it not? I mean I could see how over time they get improved but that's a given.



fatslob-:O said:
phaedruss said:

Again, I'm not sure that's true. I don't know all of the history, but I'm positive there have been substantial gains in performance with the same process node.

Over the past 40 years in the chip manufacturing industry if we counted all the gains from better architectures it would amount to 50x better performance compared to chip scaling giving us 3500x better performance. 

I doubt it's only 50x but that's something.



phaedruss said:
http://en.wikipedia.org/wiki/GeForce_600_series
http://en.wikipedia.org/wiki/GeForce_200_series

They were able to go from 1 gigaflop/ watt to 5 gigaflop/watt on 40 nm. They probably got higher on 40 nm but that's just looking at it quickly.

The Geforce GTX 600 series is featured on 28nm whereas the Geforce GTX 200 series was featured on 55nm ...

#FAIL