Editing apple/ax/a12

{{apple title|A12 Bionic}}
{{chip
|name=A12 Bionic
|no image=No
|designer=Apple
|manufacturer=TSMC
|model number=A12 Bionic
|part number=APL1W81
|part number 2=T8020
|market=Mobile
|market 2=Embedded
|first announced=September 12, 2018
|first launched=September 21, 2018
|family=Ax
|frequency=2,490 MHz (Big)
|frequency 2=1,590 MHz (Little)
|isa=ARMv8.3
|isa family=ARM
|microarch=Vortex
|microarch 2=Tempest
|core name=Vortex
|core name 2=Tempest
|core name 3=Chinook
|process=7 nm
|transistors=6,900,000,000
|technology=CMOS
|die area=83.27 mm²
|die length=9.89 mm
|die width=8.42 mm
|word size=64 bit
|core count=6
|thread count=6
|max cpus=1
|predecessor=A11 Bionic
|predecessor link=apple/ax/a11
|successor=A13
|successor link=apple/ax/a13
|contemporary=A12X Bionic
|contemporary link=apple/ax/a12x
}}
'''A12 Bionic''', also known as Apple H11P (engineering name) is a {{arch|64}} [[hexa-core]] [[ARM]] performance mobile [[SoC]] introduced by [[Apple]] in late [[2018]] as a successor to the {{\\|A11}}.
It is codenamed Cyprus.

== Overview ==
[[File:apple a12 chip announcement.png|thumb|150px|right|A12 announcement]]
Fabricated on [[TSMC]] [[7 nm process]] and integrating 6.9 billion [[transistors]], the A12 features two [[big cores|big]] {{apple|Vortex|l=arch}} cores operating at up to 2.4 GHz along with four [[little cores|little]] {{apple|Tempest|l=arch}} high-efficiency cores. Apple claims the big cores are up to 15% faster than the one in the {{\\|A11}} while being up to 40% more efficient. For the four high-efficiency cores, Apple claims up to 50% higher efficiency. Note that all six cores can operate concurrently. 

The A12 features an Apple-designed graphics processor that is said to deliver up to 50% higher performance as well as a [[neural processor]] capable of peak performance of 5 TOPS (int8).


:[[File:apple a12.svg|500px]]

== ISP ==
The A12 incorporates Apple's [[image signal processor]] (ISP), codename Petra, for camera related applications. The ISP is said to deliver faster depth engine and an improved video processor.

== Integrated Graphics ==
The A12 incorporates an Apple-designed GPU. The GPU is a four-core design which is said to deliver 50% performance over the {{\\|A11}} in tessellation and multi-layer rendering as well as lossless memory compression.

== Neural Engine ==
The A12 chip features an improved [[neural engine]], codename Quin. Apple stated this is an 8-core design (the {{\\|A11}} featured what they called a dual-core design). The new engine is capable of performing real-time processing at 5 trillion 8-bit operations per second (up from 600 billion).

== Utilizing devices ==
* [[used by::iPhone XR]]
* [[used by::iPhone XS]]
* [[used by::iPhone XS Max]]
* [[used by::iPad Mini 2019]]
* [[used by::iPad Air 2019]]
* [[used by::iPad 2020]]
* [[used by::Apple TV 2021]]
* [[used by::iPad (8th Generation)]]

{{expand list}}

== Bibliography ==
* Apple Keynote. (September 12, 2018). Apple Special Event.
* Daniel Yang, Stacy Wegner. (September 17, 2018) "''[http://www.techinsights.com/about-techinsights/overview/blog/apple-iphone-xs-teardown/ Apple iPhone Xs Max Teardown]''". TechInsights.
core count	6 +
core name	Vortex +, Tempest + and Chinook +
designer	Apple +
die area	83.27 mm² (0.129 in², 0.833 cm², 83,270,000 µm²) +
die length	9.89 mm (0.989 cm, 0.389 in, 9,890 µm) +
die width	8.42 mm (0.842 cm, 0.331 in, 8,420 µm) +
family	Ax +
first announced	September 12, 2018 +
first launched	September 21, 2018 +
full page name	apple/ax/a12 +
instance of	microprocessor +
isa	ARMv8.3 +
isa family	ARM +
ldate	September 21, 2018 +
manufacturer	TSMC +
market segment	Mobile + and Embedded +
max cpu count	1 +
microarchitecture	Vortex + and Tempest +
model number	A12 Bionic +
name	A12 Bionic +
part number	APL1W81 + and T8020 +
process	7 nm (0.007 μm, 7.0e-6 mm) +
smp max ways	1 +
technology	CMOS +
thread count	6 +
transistor count	6,900,000,000 +
used by	iPhone XR +, iPhone XS +, iPhone XS Max +, iPad Mini 2019 +, iPad Air 2019 +, iPad 2020 +, Apple TV 2021 + and iPad (8th Generation) +
word size	64 bit (8 octets, 16 nibbles) +