Thursday, October 29th 2020

Apple A14 SoC Put Under the Microscope; Die Size, and Transistor Density Calculated

Apple has established itself as a master of silicon integrated circuit design and has proven over the years that its processors deliver the best results, generation after generation. If we take a look at the performance numbers of the latest A14 Bionic, you can conclude that its performance is now rivaling some of the x86_64 chips. So you would wonder, what is inside this SoC that makes it so fast? That is exactly what ICmasters, a semiconductor reverse engineering and IP services company, has questioned and decided to find out. For starters, we know that Apple manufactures the new SoCs on TSMC's N5 5 nm node. The Taiwanese company promises to pack 171.3 million transistors per square millimeter, so how does it compare to an actual product?

ICmasters have used electron microscopy to see what the chip is made out of and to measure the transistor density. According to this source, Apple has a chip with a die size of 88 mm², which packs 11.8 billion N5 transistors. The density metric, however, doesn't correspond to that of TSMC. Instead of 171.3 million transistors per mm², the ICmasters measured 134.09 million transistors per mm². This is quite a difference, however, it is worth noting that each design will have it different due to different logic and cache layout.
Apple A14 SoC Die Apple A14 SoC
The SoC packs a 16-core neural engine, quad-cluster GPU, and six general-purpose cores, two of which are high-performance "big" FireStorm cores, while the remaining six are more efficient "LITTLE" IceStorm designs. If we break down the die usage of these cores, we can see that FireStorm cores with a big level two (L2) cache use about 9.1 mm² of the die, while the four IceStorm cores with small L2 cache use 6.44 mm². The GPU cluster is using about 11.65 mm² sizes of the portion. There is a unified system cache present on the die, however, it is a bit hard to find.
Source: SemiAnalysis
Add your own comment

9 Comments on Apple A14 SoC Put Under the Microscope; Die Size, and Transistor Density Calculated

#1
dj-electric
Not sure many people here understand how impossibly dense this chip is...
Posted on Reply
#2
Verpal
dj-electric
Not sure many people here understand how impossibly dense this chip is...
The future is now, old man.
Posted on Reply
#3
Turmania
Year 2020, AMD flipped both Intel and Nvidia, covid caused havoc. I would not be surprised to see Apple in x86 cpu sector. Nothing will surprise me anymore.
Posted on Reply
#4
roberto888
six general-purpose cores, two of which are high-performance "big" FireStorm cores, while the remaining six are more efficient "LITTLE" IceStorm designs.
Meant to be four if my maths don't fail me :p:p
Posted on Reply
#5
bonehead123
To put this in perspective, comptemplate this:

6,283,471.52 people living in a 2 BR efficiency apartment on the 82nd floor of your local slum neighborhood, hehehehe :roll:
Posted on Reply
#6
InVasMani
Turmania
Year 2020, AMD flipped both Intel and Nvidia, covid caused havoc. I would not be surprised to see Apple in x86 cpu sector. Nothing will surprise me anymore.
Apple in x86 ok with what license exactly!? Emulation I don't see unseating native x86 in a hurry that's a pipedream not at the high end of the spectrum anyway. Short of Apple licensing Ryzen tech from AMD. I'd include Intel in the licensing situation, but Apple and lax security seems like a no go. Now AMD licensing out Ryzen and RNDA2 IP to Apple to design that is certainly a real plausible scenario especially if Nvidia tries to stick it to them with ARM when they already seem to dislike them a fair bit.
Posted on Reply
#7
Turmania
InVasMani
Apple in x86 ok with what license exactly!? Emulation I don't see unseating native x86 in a hurry that's a pipedream not at the high end of the spectrum anyway. Short of Apple licensing Ryzen tech from AMD. I'd include Intel in the licensing situation, but Apple and lax security seems like a no go. Now AMD licensing out Ryzen and RNDA2 IP to Apple to design that is certainly a real plausible scenario especially if Nvidia tries to stick it to them with ARM when they already seem to dislike them a fair bit.
you can not be serious ! I was just having a light hearted fun !
Posted on Reply
#8
Fouquin
InVasMani
Emulation I don't see unseating native x86 in a hurry that's a pipedream not at the high end of the spectrum anyway.
You may be surprised to see how well custom silicon can translate x86 in real time with the right software. It's been done before.
InVasMani
Now AMD licensing out Ryzen and RNDA2 IP to Apple to design that is certainly a real plausible scenario especially if Nvidia tries to stick it to them with ARM when they already seem to dislike them a fair bit.
Nah. Imagine nVidia choosing to cut off their own leg by screwing with Apple's license in any way. nVidia doesn't want to build GPUs for Apple, that doesn't mean they want to actively hurt the bottom line by cutting their ARM license.
Posted on Reply
#9
InVasMani
Just do it leather jacket man and at a better price, TDP, and performance with equal or more capabilities...in a hurry ASAP.
Posted on Reply