Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I'm also thinking the Geekbench will be around the 8000 mark.

Keep in mind though that the iPad Pro is expected to come with a quad core processor compared to the MacBook Air's dual core processor.
No, it's not. The die shots already show a tri-core design just like A8X. Go to beyond3d if you want to verify. I don't know why I keep reading people repeating this info, it appears to have derived from intel fans who seem unable to comprehend that Apple can outperform them on perf/w and on a per-core basis.
 
No, it's not. The die shots already show a tri-core design just like A8X. Go to beyond3d if you want to verify. I don't know why I keep reading people repeating this info, it appears to have derived from intel fans who seem unable to comprehend that Apple can outperform them on perf/w and on a per-core basis.

I'm certainly not an Intel fan (although I would always go for a PC with one of their processors over AMD) and so I'm in fact even more impressed if Apple can outperform them. I haven't heard anything about it being a tri-core design. Have you got a link?
 
I'm certainly not an Intel fan (although I would always go for a PC with one of their processors over AMD) and so I'm in fact even more impressed if Apple can outperform them. I haven't heard anything about it being a tri-core design. Have you got a link?
Here is the discussion. It gets rather technical but somewhere in that thread a guy points out the L2$ on A9X and the corresponding cores (there are 3).

https://forum.beyond3d.com/threads/apple-a9-soc.57221/
 
If that's true, then it's good news for iPad Air 2 owners as apps still won't be programmed to need more than three cores. I'm hoping that that proves correct.
 
If that's true, then it's good news for iPad Air 2 owners as apps still won't be programmed to need more than three cores. I'm hoping that that proves correct.
Why is that a good thing? The tri core in the A8X an be optimised for at the same time the quad core could be.
 
Can you link directly to the post? I don't remember any confirmations on a die shot or cache amount for the A9X in that thread.
I'm not going to read a 19 page thread all over again to find a single quote just to prove something that will be proven for me in a couple days when this thing is released and geekbench shows 3 cores. It will literally be hours before you know for a fact it's tri-core. If I'm wrong it will only be a very short time for me to be proven so (which won't happen) and then you'll have plenty of time to gloat.


Or you could just read the thread an analyze the die shots yourself.
 
I'm not going to read a 19 page thread all over again to find a single quote just to prove something that will be proven for me in a couple days when this thing is released and geekbench shows 3 cores. It will literally be hours before you know for a fact it's tri-core. If I'm wrong it will only be a very short time for me to be proven so (which won't happen) and then you'll have plenty of time to gloat.


Or you could just read the thread an analyze the die shots yourself.
I doubt they'd go quad core as you've said as well. They're focusing on beefy GPUs AND they have most likely increased the clock speed along with many other changes to the architecture. Only thing that interests me is if they use TSMC exclusively.
 
If they stick to a tri-core design (which I think they will) and the A9X jumps from the A8X as much as the A9 did relative to the A8, then we can expect GeekBench scores of around 3,000 for single-core and 7,500 for multi-core. That is crazy!

I don't expect those kinds of scores for the A10, though. But they can get close, much like the A9 outclassed the A8X's single-core score and almost reached the A8X's multi-core score with one less core.
 
I'm not going to read a 19 page thread all over again to find a single quote just to prove something that will be proven for me in a couple days when this thing is released and geekbench shows 3 cores. It will literally be hours before you know for a fact it's tri-core. If I'm wrong it will only be a very short time for me to be proven so (which won't happen) and then you'll have plenty of time to gloat.


Or you could just read the thread an analyze the die shots yourself.
In fact, I've followed that thread since its beginning, and I don't remember seeing any die shot or confirmed information about the L2 cache of the A9X in the thread.

Also, the core count of the A9X and the presence of "proof" in the Beyond3D thread are two separate statements, and you can be right in one and wrong in the other. We are discussing the latter claim and your (current) lack of evidence for it.
 
Last edited:
Ah... 10 clusters, eh? Still mighty powerful considering the architecture tweaks and probable clock speed bump.
 
It's Dual core. 10 cluster GPU. Weird, I thought for sure it would be tri-core. Still right about it being < 4 cores, makes intel look really really bad though. Apple is outperforming 3Ghz 2C/4T Skylake with 2Ghz A9X.

http://forums.anandtech.com/showthread.php?t=2454159

So we're both wrong about how many cores there were. I totally agree with how bad this will make Intel look but it's good for the market and will encourage even more competition.
 
  • Like
Reactions: Codeseven
Quite underwhelmed with the overall Geekbench multi core score of the iPad Pro which comes out to around 5500. I was expecting scores more of 7000+. Nevertheless that is a great score for a dual core processor and will definitely allow the iPad Pro to last for years.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.