| Market (main) |
mobile phone |
mobile phone |
| ISA |
ARMv8.2-A(64-bit) |
ARMv8.5-A(64-bit) |
| Micro-architecture |
Cortex-A77、Cortex-A55、DynamIQ |
Avalanche， Blizzard |
| Part number, s specification |
Kirin 9000, Kirin 90005G |
APL1W07 |
| Release date |
The fourth quarter of 2020 |
The third quarter of 2021 |
| craft |
5nm FinFET EUV |
5nm FinFET |
| transistor |
– |
15000000000 |
| core |
eight |
six |
| thread |
eight |
six |
| fundamental frequency |
2.0GHz |
1.8GHz |
| Turbine frequency |
3.1GHz |
3.2GHz |
| core |
4x ARM Cortex-A55@2，05GHz |
4x Apple Blizzard @1，82GHz |
| High performance kernel |
1x ARM Cortex-A77@3，13GHz
3x ARM Cortex-A77@2，54GHz
|
2x Apple Avalanche @3，23GHz |
| Cache memory |
2MB |
12MB |
| Maximum memory capacity |
8GB |
6GB |
| Memory type |
LPDDR4X-2133、LPDDR5-2750 |
LPDDR4X memory |
| Maximum number of memory channels |
four |
four |
| Maximum memory bandwidth |
44GB/ s |
42.7GB/ s |
| TDP |
6 watts |
6 watts |
| GPU integrated graphics card |
ARM Mali-G78MP24 |
Apple A15 bionic GPU |
| GPU core |
– |
four |
| GPU execution unit |
24 |
80 |
| GPU coloring unit |
384 |
640 |
| Graphics card core |
759 MHz |
1200 MHz |
| GPU FP32 floating-point number |
2331，6GFLOPS |
1500GFLOPS |
| Artificial intelligence accelerator |
– |
Neuroengine 16 nucleus |
| AI calculation operations per second |
– |
15, 8 tops |
| Antu rabbit |
689081 |
839664 |
| （Android64-bit | iOS64-bit）
Geekbench4 single core
|
4693 |
7106 |
| （Android64-bit | iOS64-bit）
Geekbench4 multi-core
|
13767 |
18595 |
| （Android | iOS）
Geekbench5 single core
|
963 |
1698 |
| （Android | iOS）
Geekbench5 multi-core
|
3196 |
4425 |