JRPGfan said: "The Rv770 has 180 lots of 5 shaders giving a total of 900 shaders of which only 800 were enabled on the full fat die as the other 100 were spares to increase yields. Looking at the latte die it appears to have 32 lots of 5 shaders which is 160 cores meaning 176 GLOPS is the correct answer." - Marlenus
160 x 2 x 550 mhz = 176000 flops (176 Gflops)
Everything else beside what I just wrote above is wrong. Now you know, stop spreading wrong information. |
You are entirely correct.
The Wii U is based on the VLIW 5 architecture. (Very Long Instruction Word - 5 Way.)
But because the GPU is far more efficient, it is also far more capable than what the Gflop number implies, the Wii U can do more modern effects and lots more of them at any one time than the Playstation 3 and Xbox 360... And the games back that up.
teigaga said: No one actually cares about specs, they care about games and NX games will look near identical to PS4 games unless they got a decent CPU to double the FPS which would likely negate their price advantage over the PS4. |
If the NX is using Tegra... Then it is likely using Denver. And possibly Denver 2. The NX could give the Playstation 4/Xbox One a good run for it's money in CPU performance... As the latest Tegra uses a Big.Super CPU design. - The real kicker is if Nintendo has deviated from that.
We just have to wait and see.
shikamaru317 said: That would put it at at least Xbox One level, possibly between Xbox One and PS4, since Nvidia flops are better than AMD flops usually. 352 x 3 = 1056 gflops, 352 x 4 = 1,408 gflops. |
No. nVidia's flops aren't any better than AMD's flops.
There is just so much more to a GPU's performance that people are associating a chips complete performance to flops and then thinking that nVidia's flops are magically better than AMD's, when that is entirely inaccurate.
shikamaru317 said: I highly doubt Wii U is 176 gflops, it wouldn't be able to top 360 and PS3 graphics if it was 176, because 360 and PS3 were both 240 gflops. Multiple sources have it listed at 352 gflops and actual game performance seems to back that up. |
You can doubt it all you wan't, it won't change anything. There is more to performance than flops.
JRPGfan said: Newer architecture. Flops =/= flops. Also bottlenecks, maybe the Wii U has area's its stronger in than the PS3 or Xb360, which means its 176 Gflops make its equal to the other two. |
The Wii U has a ton of cache and memory to hide bandwidth and latency deficits.
Plus one of the largest issues developers had last generation was the lack of memory, especially towards the end of the generation.
Ck1x said: Where they go wrong in their assumptions is all of the unaccounted for parts of the chip which no one on NeoGaf could tell what they did, besides being fixed function. If you believe them to be them to be fixed functions then how do you just throw those processing abilities away and not include them in the theoretical number? I believe they said 35%- 40% of the chip they can't iron down exactly what it does, but they know the latte is more efficient than the previous 7th generation consoles... This was the main reason why the 176 Gflops was always being challenged. |
Even so, the unaccounted parts of the chip doesn't change what we do know about the chip.
The unacconuted parts do not have any parts resembling other parts of the chip, so it's likely logic, fixed function and specialized units and spare area.
--::{PC Gaming Master Race}::--