Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple's A6 Chip Confirmed to Carry Triple-Core Graphics
- Thread starter MacRumors
- Start date
- Sort by reaction score
Wii: 729MHz IBM PowerPC CPU (1 core), 243MHz "Hollywood" GPU
PS3: 3.2GHz CPU (1 core I think), 550MHz NVIDIA GPU
XBOX 360: 3.2GHz Tri-Core PowerPC CPU, 500MHz ATI GPU
I think the GPU in the iPhone 5 might be better than these, but I hardly know anything about GPUs. The PowerPC processor in the 360 is interesting since you might be able to run Mac OS on it
Xbox 360 development was actually done on G5s. The Wii is basically a G3 and something along the lines of a Radeon R400/R500.
Running Mac OS wouldn't happen though because there's no support for the northbridge/southbridge components. Somebody'd have to do a ton of coding to get that sorted out.
Doing something because it's possible is an awful idea especially when it's all downside and no upside.
+1 core gets you:
1) higher power consumption.
2) larger die (less chips are made per wafer)
3) lower chip yields (more chips manufactured arrive dead on arrival)
4) harder to layout (takes longer to get to market)
5) more performance.
The only thing positive is more performance. Which, for a screen of that size gets you:
1) bragging rights
2) even more ridiculously high framerates that no LCD can display.
I'm guessing everybody else here will agree with me when I say that bragging rights is not a good reason to make things more expensive, delay production even more, make design harder, and have worse battery life.
The performance target for a system like this should never be benchmarks. It should be, "at what point do I reasonably max out all the other components in the pipeline."
----------
I see your point. I was off topic. I was referring to the CPU which contains the GPU Cores.
SoC.
The SoC contains the GPU and CPU. The CPU doesn't contain the GPU.
Nit-picky, I know. But I gotta say it for the other hundreds of people who get it wrong too.
I read somewhere that they have actually found the a6 to range between 800-1200mhz, not just stay at 1ghz. I guess whatever the phone was doing at the time has something to do with it ( Im really no expert but I found the article interesting)
Looks like I was on to something when I said that the MP4 in the A5X is CPU limited... There is no reason why the MP3 should be faster than the MP4 in the off screen 1080p/720p test of GLBenchmark.
Peak memory bandwidth will limit the GPU fill rate in unified memory system architectures. The iPhone 5 appears to have significantly more peak memory bandwidth.
Honestly would never have thought Apple would end up developing top of the line processors.
4 years ago if someone were to tell me this I'd think they're crazy.
Now we need to get Apple to create processors for desktop/laptop computers.
4 years ago if someone were to tell me this I'd think they're crazy.
Now we need to get Apple to create processors for desktop/laptop computers.
TBDR's may require a certain amount of scene depth complexity to gain any significant advantages under constrained memory bandwidth limits.
Well, Apple was one of the partners that developed the PowerPC architecture. But it's a long way from being able to compete with Intel on laptop or desktop class processors. Which is probably why Apple made the massive switch from PowerPC to Intel.
True, I always thought that lots of overdraw would help show the strength of TDBR. Most major engine nowadays seems to reduce the amount of overdraw being performed. I remember this being a big deal due to how draw calls were handled with DX. I don't remember it being that big of a deal with OpenGL (or maybe I have the two reversed).
EDIT: Based on Anandtechs Performance Preview it appears that The A6 should be bandwidth constrained versus the A5X. Yet the A6 is posting faster/identical scores (in some of the offscreen tests).
Last edited: