| Name: | Google Tensor G3 |
| Family: | Google Tensor |
| Group: | Google Tensor G3 |
| Segment: | Mobile |
| Generation: | 3 |
Google Tensor G3 Benchmark, Test and specs
Last updated:
CPU lineage
The Google Tensor G3 features 8 processor cores and has the capability to manage 8 threads concurrently. It was released in Q3/2023 and belongs to the 3 generation of the Google Tensor series.
CPU Cores and Base Frequency
The Google Tensor G3 has 8 CPU cores and can calculate 8 threads in parallel. The clock frequency of the A-Core is 3.0 GHz. The number of CPU cores greatly affects the speed of the processor and is an important performance indicator.
| CPU Cores: | 8 |
| Threads: | 8 |
| Cores A: | 1x Cortex-X3 |
| Cores B: | 4x Cortex-A715 |
| Cores C: | 4x Cortex-A510 |
| Core architecture: | hybrid (Prime / big.LITTLE) |
| Hyperthreading: | No |
| Overclocking: | No |
| A-Core Frequency: | 3.0 GHz |
| B-Core Frequency: | 2.45 GHz |
| C-Core Frequency: | 2.15 GHz |
Internal Graphics
The Google Tensor G3 has integrated graphics, called iGPU for short. and 10 execution units. The iGPU uses the system's main memory as graphics memory and sits on the processor's die.
| GPU name: | ARM Immortalis-G715 MP10 |
| GPU frequency: | 0.89 GHz |
| GPU (Turbo): | -- |
| Execution units: | 10 |
| Shaders: | -- |
| Max. GPU Memory: | 0 bytes |
| Max. displays: | -- |
| Generation: | Vallhall |
| Direct X: | 12 |
| Technology: | 4 nm |
| Release date: | Q4/2022 |
Artificial Intelligence and Machine Learning
| AI hardware: | Google Tensor AI |
| AI specifications: | Google Edge TPU |
Hardware codec support
A photo or video codec that is accelerated in hardware can greatly accelerate the working speed of a processor and extend the battery life of notebooks or smartphones when playing videos.
| h265 / HEVC (8 bit): | Decode / Encode |
| h265 / HEVC (10 bit): | Decode / Encode |
| h264: | Decode / Encode |
| VP8: | Decode / Encode |
| VP9: | Decode / Encode |
| AV1: | Decode / Encode |
| AVC: | Decode / Encode |
| VC-1: | Decode / Encode |
| JPEG: | Decode / Encode |
Memory & PCIe
The processor supports a maximum memory capacity of 12.0 GB distributed across 2 memory channels. It offers a peak memory bandwidth of 53.0 GB/s. Both the type and quantity of memory can have a substantial impact on the overall system performance.
| Memory type: | LPDDR5-5500 |
| Max. Memory: | 12.0 GB |
| Memory channels: | 2 |
| Bandwidth: | 53.0 GB/s |
| ECC: | No |
| PCIe: | pci |
| AES-NI: | No |
Thermal Management
The processor has a thermal design power (TDP) of 10 W watts. TDP indicates the cooling solution needed to effectively manage the processor's heat. It generally provides an approximate indication of the actual power consumption of the CPU itself.
| TDP (PL1 / PBP): | 10 W |
| TDP (PL2): | None |
| TDP up: | None |
| TDP down: | None |
| Tjunction max: | -- |
Technical details
The Google Tensor G3 is manufactured using a 3 nm process. A smaller manufacturing process indicates a more contemporary and energy-efficient CPU. A substantial cache can significantly enhance the processor's performance, particularly in scenarios like gaming.
| Technology: | 3 nm |
| Chip design: | Chiplet |
| Socket: | |
| L2-Cache: | 0 bytes |
| L3-Cache: | 0 bytes |
| Architecture: | G3 |
| Operating systems: | Android |
| Virtualization: | None |
| Instruction set (ISA): | ARMv9-A64 (64 bit) |
| ISA extensions: | |
| Release date: | Q3/2023 |
| Release price: | -- |
| Part Number: | -- |
| Documents: | Technical data sheet |
ADS
Benchmarks
iGPU - FP32 Performance (Single-precision GFLOPS)
|
Samsung Exynos 3110
NoneC NoneT @ --
|
3
|
|
|
MediaTek MT6575M
NoneC NoneT @ --
|
2
|
|
|
Google Tensor G3
8C 8T @ 3.0 GHz
|
1
|
|
MediaTek Kompanio 828
8C 8T @ 2.6 GHz
|
0
|
|
|
MediaTek Kompanio 1200
8C 8T @ 2.6 GHz
|
0
|
