From WikiChip
Ax - Apple
Apple Ax | |
Developer | Apple, ARM Holdings |
Manufacturer | Samsung, TSMC |
Type | System on Chips |
Architecture | High-performance ARM SoCs |
ISA | ARM |
Word size | 32 bit 4 octets , 64 bit8 nibbles 8 octets
16 nibbles |
Process | 45 nm 0.045 μm , 32 nm4.5e-5 mm 0.032 μm , 28 nm3.2e-5 mm 0.028 μm , 20 nm2.8e-5 mm 0.02 μm , 16 nm2.0e-5 mm 0.016 μm , 14 nm1.6e-5 mm 0.014 μm , 10 nm1.4e-5 mm 0.01 μm , 7 nm1.0e-5 mm 0.007 μm , 5 nm7.0e-6 mm 0.005 μm
5.0e-6 mm |
Technology | CMOS |
Clock | 800 MHz-3.1 GHz |
Ax is a family of high-performance 32-bit and 64-bit ARM system on chips designed by Apple exclusively for their consumer products.
Contents
Overview
Since the introduction of the original iPhone in 2007, Apple has been developing a series of chips designed exclusively for their own consumer products. Those system on chips are found in all iPhones, iPod Touches, iPads, Apple TVs, and Apple Watches. While initial models used IP cores licensed from ARM Holdings, since the introduction of the A6 in 2012, Apple has been developing their own microarchitecture for their chips which have become considerably more powerful than other ARM-based microprocessors.
Analysis
Die size comparison
Members
- A4 (Cortex-A8)
- A5 • A5X (Cortex-A9)
- A6 • A6X (Swift)
- A7 (Cyclone)
- A8 • A8X (Typhoon)
- A9 • A9X (Twister)
- A10 Fusion • A10X (Hurricane)
- A11 Bionic (Monsoon)
- A12 Bionic (Vortex)
- A13 Bionic (Lightning)
- A14 Bionic (Firestorm)
- A15 Bionic (Avalanche)
- A16 Bionic (Everest Gen 1)
- A17 Pro (Everest Gen 2)
- A18 • A18 Pro (Everest Gen 3)
Mobile processors
Model | CPU | GPU | NPU | RAM | Catalog number | Process | Release date |
---|---|---|---|---|---|---|---|
Apple A18 Pro | 2x Everest @4.05GHz 4x Sawtooth @2.42GHz |
Proprietary (hexa-core) (2227 GFLOPS) |
16x Neural Engine (35 TOPS) |
LPDDR5x-7500 single-channel 64-bit @3750MHz (60GB/s) |
APL1V07 | TSMC N3E 3 nm |
2024/09 |
Apple A18 | 2x Everest @4.05GHz 4x Sawtooth @2.42GHz |
Proprietary (penta-core) (1789 GFLOPS) |
16x Neural Engine (35 TOPS) |
LPDDR5x-7500 single-channel 64-bit @3750MHz (60GB/s) |
APL1V08 | "3 nm" TSMC N3E |
2024/09 |
Apple A17 | 2x Everest @3.78GHz 4x Sawtooth @2.11GHz |
(2147 GFLOPS) Proprietary (hexa-core) |
(35 TOPS) 16x Neural Engine |
LPDDR5-6400 single-channel 64-bit @3200MHz (51.2GB/s) |
APL1V02 | "3 nm" TSMC N3B |
2023/09 |
Apple A16 | 2x Everest @3.46GHz 4x Sawtooth @2.02GHz |
(1789 GFLOPS) Proprietary (penta-core) |
(17 TOPS) 16x Neural Engine |
LPDDR5-6400 single-channel 64-bit @3200MHz (51.2GB/s) |
APL1W10 | "4 nm" TSMC N4P |
2022/09 |
Apple A15 | 2x Avalanche @3.24GHz 4x Blizzard @2.0GHz |
(1712 GFLOPS) Proprietary (penta-core) |
(15.8 TOPS) 16x Neural Engine |
LPDDR4X-4266 single-channel 64-bit @2133MHz (34.1GB/s) |
APL1W07 | "5 nm" TSMC N5P |
2021/09 |
Apple A14 | 2x Firestorm @3.0GHz 4x Icestorm |
(748.8 GFLOPS) Proprietary (quad-core) |
(11 TOPS) 16x Neural Engine |
LPDDR4X-4266 single-channel 64-bit @2133MHz (34.1GB/s) |
APL1W87 | "5 nm" TSMC N5 |
2020/09 |
Apple A13 | 2x Lightning @2.65GHz 4x Thunder @1.8GHz |
(806 GFLOPS) Proprietary (quad-core) |
(6 TOPS) 8x Neural Engine |
LPDDR4X-4266 single-channel 64-bit @2133MHz (34.1GB/s) |
APL1W85 | "7 nm" TSMC N7P |
2019/09 |
Apple A12 | 2x Vortex @2.5GHz 4x Tempest @1.6GHz |
(576 GFLOPS) 4x Apple G11P |
(5 TOPS) 8x Neural Engine |
LPDDR4X-4266 single-channel 64-bit @2133 MHz (34.1GB/s) |
APL1W81 | "7 nm" TSMC N7 |
2018/09 |
Apple A11 | 2x Monsoon @2.4GHz 4x Mistral |
(408 GFLOPS) Proprietary (tri-core) |
(600 BOPS) 2x Neural Engine |
LPDDR4X-4266 single-channel 64-bit @2133MHz (34.1GB/s) |
APL1W72 | "10 nm" TSMC FinFET |
2017/09 |
Apple A10 | 2x Hurricane @2.34GHz 2x Zephyr @1.05GHz |
(345.6 GFLOPS) 6x PowerVR GT7600 Plus @900MHz |
LPDDR4-3200 single-channel 64-bit @1600MHz (25.6GB/s) |
APL1W24 | "16 nm" TSMC 16FFC |
2016/09 | |
Apple A9 | 2x Twister @1.85GHz | (249.6 GFLOPS) 6x PowerVR GT7600 @650MHz |
LPDDR4-3200 single-channel 64-bit @1600MHz (25.6GB/s) |
APL0898 APL1022 |
"14 nm" Samsung "16 nm" TSMC |
2015/09 | |
Apple A8 | 2x Apple Typhoon @1.5GHz | (136.5 GFLOPS) 4x PowerVR GXA6450 @533MHz |
LPDDR3-1600 single-channel 64-bit @800MHz (12.8GB/s) |
APL1011 | "20 nm" TSMC |
2014/09 | |
Apple A7 | 2x Cyclone @1.3GHz | (115.2 GFLOPS) 4x PowerVR G6430 @450MHz |
LPDDR3-1600 single-channel 64-bit @800MHz (12.8GB/s) |
APL0698 / APL5698 S5L8960X |
"28 nm" Samsung HKMG |
2013/09 | |
Apple A6 | 2x Swift @1.3GHz | (25 GFLOPS) 3x PowerVR SGX543 @266MHz |
LPDRR2-1066 dual-channel 32-bit @533MHz (8.5GB/s) |
APL0598 S5L8950X |
"32 nm" Samsung HKMG |
2012/09 | |
Apple A5 | 2x ARM Cortex-A9 @1.0GHz | (12.8 GFLOPS) 2x PowerVR SGX543 @200MHz |
LPDDR2-800 dual-channel 32-bit @400MHz (6.4GB/s) |
APL0498 / APL2498 S5L8940X |
"45 nm" Samsung "32 nm" Samsung HKMG |
2011/03 2012/03 | |
Apple A4 | 1x ARM Cortex-A8 @1.0GHz | PowerVR SGX535 | LPDDR-400 dual-channel 32-bit @200MHz (3.2GB/s) |
APL0398 S5L8930X |
"45 nm" Samsung |
2010/04 | |
Model | CPU | GPU | NPU | RAM | Catalog name | Lithography | Release date |
Tablet processors
Model | CPU | GPU | NPU | ISP | RAM | Modem | Catalog number | Lithography | Release date |
---|---|---|---|---|---|---|---|---|---|
Apple A12X |
|
(1200 GFLOPs) Proprietary (hexa/octa-core) |
(6 TOPS) 8x Neural Engine |
LPDDR4X-4266 dual-channel 64-bit @ 2133 MHz (68.2 GB/s) |
not available | APL1083 | "7 nm" TSMC N7 |
2018/10 | |
Apple A10X |
|
(768 GFLOPs) 12x PowerVR GT7500 Plus @ 900 MHz |
LPDDR4-3200 dual-channel 64-bit @ 1600 MHz (51.2 GB/s) |
not available | APL1071 | "10 nm" TSMC 10FF |
2017/06 | ||
Apple A9X |
|
(345.6 GFLOPs) 12x PowerVR GTA7850 @ 650 MHz |
LPDDR4-3200 dual-channel 64-bit @ 1600 MHz (51.2 GB/s) |
not available | APL1021 | "16 nm" TSMC 16FF |
2015/11 | ||
Apple A8X |
|
(230.4 GFLOPs) 8x PowerVR GXA6850 @ 450 MHz |
LPDDR3-1600 dual-channel 64-bit @ 800 MHz (25.6 GB/s) |
not available | APL1012 | "20 nm" TSMC 20 |
2014/10 | ||
Apple A6X |
|
(68.1 GFLOPs) 4x PowerVR SGX554 @ 266 MHz |
LPDDR2-1066 quad-channel 32-bit @ 533 MHz (17.1 GB/s) |
not available | APL5598 | "32 nm" Samsung 32 HKMG |
2012/10 | ||
Apple A5X |
|
(25 GFLOPs) 4x PowerVR SGX543 @ 200 MHz |
LPDDR2-800 quad-channel 32-bit @ 400 MHz (12.8 GB/s) |
not available | APL5498 | "45 nm" Samsung 45 |
2012/03 | ||
Model | CPU | GPU | NPU | ISP | RAM | Modem | Catalog name | Lithography | Release date |
PC processors
Model | CPU | GPU | NPU | ISP | RAM | Modem | Catalog number | Lithography | Release date |
---|---|---|---|---|---|---|---|---|---|
Apple M4 |
|
(3.7 TFLOPS) Proprietary (10 cores) |
(38 TOPS) 16x Neural Engine |
LPDDR5x-7500 128-bit @ 3750 MHz (120 GB/s) | not available | APL1206 | "3 nm" TSMC N3E |
2024/05 | |
Apple M3 Max |
|
(10.6 or 14.1 TFLOPS) Proprietary (30 or 40 cores) |
(18 TOPS) 16x Neural Engine |
LPDDR5-6400 384-bit or 512-bit @ 3200 MHz (307.2 or 409.6 GB/s) |
not available | APL1201 | "3 nm" TSMC N3B |
2023/10 | |
Apple M3 Pro |
|
(4.9 or 6.4 TFLOPS) Proprietary (14 or 18 cores) |
(18 TOPS) 16x Neural Engine |
LPDDR5-6400 192-bit @ 3200 MHz (153.6 GB/s) |
not available | APL1201 | "3 nm" TSMC N3B |
2023/10 | |
Apple M3 |
|
(2.8 TFLOPS) Proprietary (8 or 10 cores) |
(18 TOPS) 16x Neural Engine |
LPDDR5-6400 128-bit @ 3200 MHz (102.4 GB/s) |
not available | APL1201 | "3 nm" TSMC N3B |
2023/10 | |
Apple M2 Ultra |
|
(27.19 TFLOPS) Proprietary (60/76 cores) |
(31.6 TOPS) 32x Neural Engine |
LPDDR5-6400 1024-bit @ 3200 MHz (800 GB/s) |
not available | "5 nm" TSMC N5P |
2023/06 | ||
Apple M2 Max |
|
(13.59 TFLOPS) Proprietary (38 cores) |
(15.8 TOPS) 16x Neural Engine |
LPDDR5-6400 512-bit @ 3200 MHz (400 GB/s) |
not available | T6021 | "5 nm" TSMC N5P |
2023/01 | |
Apple M2 Pro |
|
(6.79 TFLOPS) Proprietary (19 cores) |
(15.8 TOPS) 16x Neural Engine |
LPDDR5-6400 256-bit @ 3200 MHz (200 GB/s) |
not available | T6020 | "5 nm" TSMC N5P |
2023/01 | |
Apple M2 |
|
(3.6 TFLOPs) Proprietary (10 cores) |
(15.8 TOPS) 16x Neural Engine |
LPDDR5-6400 128-bit @ 3200 MHz (100 GB/s) |
not available | APL1109 | "5 nm" TSMC N5P |
2022/06 | |
Apple M1 Ultra |
|
(21 TFLOPs) Proprietary (64 cores) |
(22 TOPS) 32x Neural Engine |
LPDDR5-6400 1024-bit @ 3200 MHz (800 GB/s) |
not available | APL1W06 | "5 nm" TSMC N5 |
2022/03 | |
Apple M1 Max |
|
(10.4 TFLOPs) Proprietary (32 cores) |
(11 TOPS) 16x Neural Engine |
LPDDR5-6400 512-bit @ 3200 MHz (400 GB/s) |
not available | APL1105 | "5 nm" TSMC N5 |
2021/10 | |
Apple M1 Pro |
|
(5.2 TFLOPs) Proprietary (16 cores) |
(11 TOPS) 16x Neural Engine |
LPDDR5-6400 256-bit @ 3200 MHz (200 GB/s) |
not available | APL1103 | "5 nm" TSMC N5 |
2021/10 | |
Apple M1 |
|
(2.6 TFLOPs) Proprietary (octa-core) |
(11 TOPS) 16x Neural Engine |
LPDDR4X-4266 dual-channel 64-bit @ 2133 MHz (68.2 GB/s) |
not available | APL1102 | "5 nm" TSMC N5 |
2020/11 | |
Model | CPU | GPU | NPU | ISP | RAM | Modem | Catalog name | Lithography | Release date |
See also
References
See also
- Qualcomm Snapdragon
- MediaTek Helio
- Intel Atom
Facts about "Ax - Apple"
designer | Apple + and ARM Holdings + |
full page name | apple/ax + |
instance of | system on a chip family + |
instruction set architecture | ARM + |
main designer | Apple + |
manufacturer | Samsung + and TSMC + |
name | Apple Ax + |
process | 45 nm (0.045 μm, 4.5e-5 mm) +, 32 nm (0.032 μm, 3.2e-5 mm) +, 28 nm (0.028 μm, 2.8e-5 mm) +, 20 nm (0.02 μm, 2.0e-5 mm) +, 16 nm (0.016 μm, 1.6e-5 mm) +, 14 nm (0.014 μm, 1.4e-5 mm) +, 10 nm (0.01 μm, 1.0e-5 mm) +, 7 nm (0.007 μm, 7.0e-6 mm) + and 5 nm (0.005 μm, 5.0e-6 mm) + |
technology | CMOS + |
word size | 32 bit (4 octets, 8 nibbles) + and 64 bit (8 octets, 16 nibbles) + |