Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Danny82

macrumors member
Original poster
Jul 1, 2020
50
25
late to the news but apparently there are leaks for the Apple A14 in geekbench 5.. just sharing with those who have not seen it to create some hype as I am really excited :D imagine that becoming A14m with rumored 8p core and 4e core.. thats pretty amazing šŸ˜³ that should beat out intel i7-10875H..?

 
  • Like
Reactions: souko and Mojo1019

leman

macrumors Core
Oct 14, 2008
19,516
19,664
If these are accurate, then itā€™s much better than what I would have expected... if an iPhone A14 can score this, then the Mac will set new records in single-core performance.
 

burgerrecords

macrumors regular
Jun 21, 2020
222
106
If these are accurate, then itā€™s much better than what I would have expected... if an iPhone A14 can score this, then the Mac will set new records in single-core performance.

Iā€™m not certain thatā€™s a forgone conclusion. The March iPad Pro has lower clock speeds than the iPhone 11 I believe. More cores equals more heat than canā€™t be dissipated?
 

Danny82

macrumors member
Original poster
Jul 1, 2020
50
25
How do we know this isnā€™t the A14 in a Mac?
Because it writes 6 core which only only make sense if it is for iphone as 2p core and 4e core..?

If these are accurate, then itā€™s much better than what I would have expected... if an iPhone A14 can score this, then the Mac will set new records in single-core performance.
agree, if these are accurate I will be really happy. Not even considering if macbook clip are boosted 0.2 mhz more to kill the single core score.. just base only current leak, not only we see a good boost in cpu (just checked and single core score beating intel i9-10900k and multi core score close to beating amd 3950x with my not accurate counting by doing 1658 x 8 core), it does seems that the gpu is getting a good boost as well assuming A14 in iphone is the same as in A13 using only 4 GPU core.. straight line calculations will will put a 16 apple gpu score of around 33000.. hmmmm.. i was hoping gpu score to be close to 40000 in geekbench :( using razer blade 13 as benchmark.. bummer..
 
Last edited:
  • Like
Reactions: Woochoo

ek9max

macrumors regular
Feb 22, 2011
224
38
Exciting stuff for sure. I'm more curious how a A14 spec'd macbook would perform in GPU tasks also.
 

MandiMac

macrumors 65816
Feb 25, 2012
1,433
883
Is it just me or does AppleInsiders screenshot look photoshopped? If you look on the left, it seems legit with the text and all, and the right seems slightly warped...
 

leman

macrumors Core
Oct 14, 2008
19,516
19,664
Is it just me or does AppleInsiders screenshot look photoshopped? If you look on the left, it seems legit with the text and all, and the right seems slightly warped...

Yeah, and itā€™s also too good to be true... Iā€™m honestly skeptical.
 

MikhailT

macrumors 601
Nov 12, 2007
4,583
1,327
Iā€™m not certain thatā€™s a forgone conclusion. The March iPad Pro has lower clock speeds than the iPhone 11 I believe. More cores equals more heat than canā€™t be dissipated?

Nothing is a forgone conclusion, correct.

However, a few things of note;

1. iPad Pro is using two year old CPU design, A12. A12Z is just enabling an extra GPU core that was disabled before in A12X. However, it is the exact same as A12, which also means, it is fabbed on the original 7nm process, not the improved 7nm+ process that A13 is fabbed on. A14 is on the new 5nm process, we don't know the impact of it is yet, it can be faster or it can be more efficient, Apple chooses how to balance them per SoC.
2. iPhone 11 is using A13, which is already 15-25% improvement across the board compared to previous A12 chip in iPhone. Read all of the changes between A12 > A13 here: https://www.anandtech.com/show/14892/the-apple-iphone-11-pro-and-max-review/2
3. In the same review, Anandtech mentioned that A13 is actually more power hungry than A12 but not so much to harm it because they also fixed the sustained performance issue while keeping their devices much cooler than any other devices. So, in other words, more core does not mean more heat either, the higher frequency can cause more heat than just having another core. That's why Apple design SoC with various combinations of multiple performance and efficiency cores, instead of being consistent across all of them.
1. Macs will be actively cooled, that has a huge impact of how well it'll work as well. Apple can push the frequency to the max for a longer period of time, which will speed up the single core perf. That's what Intel's Turbo Boost is about. Again, A13 was designed for sustained pref in mind rather than being much faster but they were able to do both.

What's key is the IPC gain, how much work it can do per clock and perf per watt. In this case, A13 has ~15-20% IPC improvement over A12. A14 already is likely to match that. In other words, expect iPad Pro's A14X to be a monster (I can see 30-40% gain over A12X (we already know it'll be a min of 15-20% from A13)) and if that's going to be a monster, imagine what the Macs will be like with active cooling.

Of course, there's a chance Apple will reduce the pref intentionally to boost the GPU and other type of cores instead; which means they may intentionally reduce/limit the frequency/cores in order to focus on the GPU cores instead. Given how much of the OS is now GPU accelerated and how much Metal is improving their software ecosystem, that's a strong possibility.

So, no one will know for sure until it is out.
 
Last edited:

awesomedeluxe

macrumors 6502
Jun 29, 2009
262
105
What on earth. Clock speeds are looking at a boost of around 7-8% from N5... I would've expected overall single core performance in the ballpark of around 10%. But this is more like 25%.
 
  • Like
Reactions: burgerrecords

leman

macrumors Core
Oct 14, 2008
19,516
19,664
What on earth. Clock speeds are looking at a boost of around 7-8% from N5... I would've expected overall single core performance in the ballpark of around 10%. But this is more like 25%.

It is not impossible if they made the architecture even wider. But then a single a Apple core would be twice as wide than Sunny Cove, and I have my doubts about that...
 

lJoSquaredl

macrumors 6502a
Mar 26, 2012
522
227
I wonder how they'll differentiate the different lines. They use the same chips in the iPhone, iPad, etc so I wonder if they'll do that between the different laptop models as well, just turning off cores based on what you purchase maybe? Either way if I can efficiently edit video on a 13 inch MBP as well as a 15 inch now that'd be cool, depending on how all this GPU stuff works with their ARM based stuff.
 

leman

macrumors Core
Oct 14, 2008
19,516
19,664
I wonder how they'll differentiate the different lines. They use the same chips in the iPhone, iPad, etc so I wonder if they'll do that between the different laptop models as well, just turning off cores based on what you purchase maybe? Either way if I can efficiently edit video on a 13 inch MBP as well as a 15 inch now that'd be cool, depending on how all this GPU stuff works with their ARM based stuff.

They donā€™t use the same chip. The number of cores, clocks, cache sizes are different. The laptops might feature different RAM speeds and different size GPUs among others. Now that Apple is ā€œfreeā€ from the third party IHV, they can differentiate more meaningfully and more flexibly that earlier.
 

awesomedeluxe

macrumors 6502
Jun 29, 2009
262
105
It is not impossible if they made the architecture even wider. But then a single a Apple core would be twice as wide than Sunny Cove, and I have my doubts about that...
Faster memory is probably part of the equation too. Maybe this means the LPDDR5 dream is real. The multicore score increases by a much larger margin, and multicore results should see more benefit from faster DRAM.

Edit: I think Sunny Cove is actually 10 wide? While the A13 has a 6 ALU backend, and it's not clear to me if all of those ALUs are equally competent. So I think we can go wider. More ALUs, more better. That's the word folks. Tell your friends you're getting the processor with the most ALUs and see how jealous they get
 
Last edited:

leman

macrumors Core
Oct 14, 2008
19,516
19,664
Faster memory is probably part of the equation too. Maybe this means the LPDDR5 dream is real. The multicore score increases by a much larger margin, and multicore results should see more benefit from faster DRAM.
(Though honestly that multicore score is really high. Changes to the Icestorm cores?)

LPDDR5 on the new iPhone is almost a given, seeing how there are phones already shipping with it. Iā€™m more curious whether we will see it in a desktop.

Anyway, I would take these benchmarks with a big grain of salt for now.
 

TylerL

macrumors regular
Jan 2, 2002
207
291
That purported Geekbench single-core score is more than just exceptional.
It essentially matches the absolute best existing Geekbench scores attained through extreme overlocking.
If this is true, Apple will be effortlessly reaching a level of performance that only a handful of experts could hope to sustain for mere minutes on x86 hardware. ...And this is on a mobile device!
 

Danny82

macrumors member
Original poster
Jul 1, 2020
50
25
As mention, please take this with a super huge amount of salt and not just a grain.. not a technical person but because I'm excited so I keep finding sources to read.. so just sharing what I found and hope to hear more knowledgable discussion from u guys.

Something else I found from macworld on his prediction:

[automerge]1594703997[/automerge]
Oh and one more thing, from this macworld article it seems that SK-Hynix is Apple supplier for RAM, and SK-Hynix has begin mass production of HBM2e.. does it means anything..!? Omg.. please Apple please.. don't be cheap on us and give us HBM2e on MBP..!!! And I'm really surprise the performance leap for the GPU from the A12 to A13.. ?

And just to be clear, even though I am sharing what seems like iphone articles, am mainly just sharing what A14 could be for Apple silicon on mac..
 
Last edited:

brzy25

macrumors newbie
Jul 21, 2020
6
1
I wonder how they'll differentiate the different lines. They use the same chips in the iPhone, iPad, etc so I wonder if they'll do that between the different laptop models as well, just turning off cores based on what you purchase maybe? Either way if I can efficiently edit video on a 13 inch MBP as well as a 15 inch now that'd be cool, depending on how all this GPU stuff works with their ARM based stuff.

I'm hoping being free of Intel's menagerie of sku's will allow them to simplify the laptop range like their phones. Just a MBP 13 and MBP 14+16 with 3 tiers of SSD sizes, 8gb ram for the 13 and 16gb for the Pros.
 
  • Like
Reactions: awesomedeluxe

lJoSquaredl

macrumors 6502a
Mar 26, 2012
522
227
I'm hoping being free of Intel's menagerie of sku's will allow them to simplify the laptop range like their phones. Just a MBP 13 and MBP 14+16 with 3 tiers of SSD sizes, 8gb ram for the 13 and 16gb for the Pros.

They should jump to 16gb stock for all devices with the word "Pro" in it imo. But yeah I do find a lot of variations annoying, if they can slim down the lines a bit that'd be nice.
 
  • Like
Reactions: Jimmy James

Kostask

macrumors regular
Jul 4, 2020
230
104
Calgary, Alberta, Canada
The minimum memory for any AS powered Macs will be 16GB. Anything less will create a real issue. While posters can go on and on about CISC vs. RISC, how RISC is or isn't inherently "better", whether ARM can be a high performance architecture, etc. ad for forever and a day, the truth is, RISC architectures are less instruction dense than CISC architectures. What that comes down to is that you need more instructions to perform a task than CISC architectures do because the instructions are simpler. That drives the requirement for more RAM on a RISC machine, and why I am saying that 16GB will be the minimum for AS Macs, as it currently is for the DTK Mac Mini.

Think of it another way: Would Apple be sending out a system to developers in which the developers port, cross compile, and test their software (or at least verify their software works under Rosetta 2) with 16GB, and then ship consumer AS Macs with half that amount of RAM? Possible, but at least to me, highly unlikely.
 

leman

macrumors Core
Oct 14, 2008
19,516
19,664
While posters can go on and on about CISC vs. RISC, how RISC is or isn't inherently "better", whether ARM can be a high performance architecture, etc. ad for forever and a day, the truth is, RISC architectures are less instruction dense than CISC architectures. What that comes down to is that you need more instructions to perform a task than CISC architectures do because the instructions are simpler.

I donā€™t know how much this is a factor. Code density of x86-64 and ARM64 are pretty much comparable. Optimized for size, x86-64 tends to do a bit better (since the CPU can utilize shorter instructions), optimized for performance, ARM64 tends to do better in my tests.

It is also not true that ARM64 instructions are simpler. They are just fixed length (compared to variable length x86) and they need separate instructions for load/store, where x86 sometimes can do memory operations as part of the base instruction. But at the same time ARM can do multiple memory operations in a single instruction and it has more registers (so complex code doesnā€™t need to do register juggling). The ā€œtrueā€ CISC instructions on x86 are all dead slow and deprecated, so nobody uses them anyway.
 
  • Like
Reactions: vigilant
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.