HiSilicon Kirin 960 vs HiSilicon Kirin 970
The HiSilicon Kirin 960 and Kirin 970 are two processors designed by Huawei's semiconductor subsidiary, HiSilicon. While they share some similarities in terms of their CPU cores and architecture, there are also some distinct differences in their specifications.
Both processors utilize a combination of 4 Cortex-A73 cores and 4 Cortex-A53 cores, providing a balanced blend of performance and power efficiency. This architectural similarity ensures that both processors can handle a variety of tasks with ease.
However, one notable difference between these processors lies in their lithography. The Kirin 960 features a 16 nm lithography, while the Kirin 970 boasts a more advanced 10 nm lithography. This smaller lithography allows for more transistors to be packed into the same physical space, resulting in improved performance and power efficiency.
Speaking of transistors, the Kirin 960 packs 4000 million transistors, while the Kirin 970 ups the ante with 5500 million transistors. This increase in transistor count on the Kirin 970 is likely a result of the smaller lithography, allowing for more complex processing and enhanced performance.
In terms of power consumption, the Kirin 960 has a thermal design power (TDP) rating of 5 Watts, while the Kirin 970 has a slightly higher TDP rating of 9 Watts. This increase can be attributed to the greater number of transistors and the additional power required for the integrated Neural Processing Unit (NPU), which is unique to the Kirin 970.
The inclusion of the NPU in the Kirin 970 is another significant differentiating factor. This dedicated hardware accelerator for artificial intelligence (AI) tasks enhances AI performance on the processor, making it particularly well-suited for tasks such as image recognition and voice processing.
In summary, while both the Kirin 960 and Kirin 970 share similar CPU cores and architecture, the latter offers better performance and power efficiency due to its smaller lithography and increased transistor count. Furthermore, the Kirin 970 sets itself apart with the inclusion of the NPU, enabling superior AI processing capabilities.
Both processors utilize a combination of 4 Cortex-A73 cores and 4 Cortex-A53 cores, providing a balanced blend of performance and power efficiency. This architectural similarity ensures that both processors can handle a variety of tasks with ease.
However, one notable difference between these processors lies in their lithography. The Kirin 960 features a 16 nm lithography, while the Kirin 970 boasts a more advanced 10 nm lithography. This smaller lithography allows for more transistors to be packed into the same physical space, resulting in improved performance and power efficiency.
Speaking of transistors, the Kirin 960 packs 4000 million transistors, while the Kirin 970 ups the ante with 5500 million transistors. This increase in transistor count on the Kirin 970 is likely a result of the smaller lithography, allowing for more complex processing and enhanced performance.
In terms of power consumption, the Kirin 960 has a thermal design power (TDP) rating of 5 Watts, while the Kirin 970 has a slightly higher TDP rating of 9 Watts. This increase can be attributed to the greater number of transistors and the additional power required for the integrated Neural Processing Unit (NPU), which is unique to the Kirin 970.
The inclusion of the NPU in the Kirin 970 is another significant differentiating factor. This dedicated hardware accelerator for artificial intelligence (AI) tasks enhances AI performance on the processor, making it particularly well-suited for tasks such as image recognition and voice processing.
In summary, while both the Kirin 960 and Kirin 970 share similar CPU cores and architecture, the latter offers better performance and power efficiency due to its smaller lithography and increased transistor count. Furthermore, the Kirin 970 sets itself apart with the inclusion of the NPU, enabling superior AI processing capabilities.
CPU cores and architecture
Architecture | 4x 2.4 GHz – Cortex-A73 4x 1.8 GHz – Cortex-A53 |
4x 2.4 GHz – Cortex-A73 4x 1.8 GHz – Cortex-A53 |
Number of cores | 8 | 8 |
Instruction Set | ARMv8-A | ARMv8-A |
Lithography | 16 nm | 10 nm |
Number of transistors | 4000 million | 5500 million |
TDP | 5 Watt | 9 Watt |
Neural Processing | HiSilicon NPU |
Memory (RAM)
Max amount | up to 6 GB | up to 8 GB |
Memory type | LPDDR4 | LPDDR4 |
Memory frequency | 1866 MHz | 1866 MHz |
Memory-bus | 2x32 bit | 4x16 bit |
Storage
Storage specification | UFS 2.1 | UFS 2.1 |
Graphics
GPU name | Mali-G71 MP8 | Mali-G72 MP12 |
GPU Architecture | Bifrost | Bifrost |
GPU frequency | 900 MHz | 750 MHz |
Execution units | 8 | 12 |
Shaders | 128 | 192 |
DirectX | 11.3 | 12 |
OpenCL API | 1.2 | 2.0 |
Vulkan API | 1.0 | 1.0 |
Camera, Video, Display
Max screen resolution | 2340x1080 | |
Max camera resolution | 1x 20MP, 2x 12MP | 1x 48MP, 2x 20MP |
Max Video Capture | 4K@30fps | 4K@30fps |
Video codec support | H.264 (AVC) H.265 (HEVC) VP8 VP9 |
H.264 (AVC) H.265 (HEVC) VP8 VP9 |
Wireless
4G network | Yes | Yes |
5G network | Yes | Yes |
Peak Download Speed | 0.6 Gbps | 1.2 Gbps |
Peak Upload Speed | 0.15 Gbps | 0.15 Gbps |
Wi-Fi | 5 (802.11ac) | 5 (802.11ac) |
Bluetooth | 4.2 | 4.2 |
Satellite navigation | BeiDou GPS Galileo GLONASS |
BeiDou GPS Galileo GLONASS |
Supplemental Information
Launch Date | 2016 October | 2017 September |
Partnumber | Hi3660 | Hi3670 |
Vertical Segment | Mobiles | Mobiles |
Positioning | Flagship | Flagship |
AnTuTu 10
Total Score
GeekBench 6 Single-Core
Score
GeekBench 6 Multi-Core
Score
Popular comparisons:
1
Samsung Exynos 850 vs Apple A10 Fusion
2
MediaTek Helio G37 vs Qualcomm Snapdragon 675
3
Unisoc SC9863A vs MediaTek Helio P60
4
Qualcomm Snapdragon 865 vs Unisoc Tiger T606
5
Samsung Exynos 9820 vs Samsung Exynos 980
6
Unisoc Tiger T710 vs Qualcomm Snapdragon 665
7
Apple A13 Bionic vs MediaTek Helio G35
8
HiSilicon Kirin 950 vs Unisoc Tiger T310
9
Qualcomm Snapdragon 685 vs Qualcomm Snapdragon 712
10
MediaTek Helio G90T vs Unisoc Tiger T610