ASUS ROG Zephyrus G15 + RTX 3060 Performance Overview
The ASUS ROG Zephyrus G15, powered by the AMD Ryzen 9 6900HS and NVIDIA GeForce RTX 3060 Laptop GPU, offers strong AI performance across various workloads. With 16 GB of RAM, the system excels in tasks like Style Transfer, Pose Estimation, and Object Detection. The RTX 3060 delivers efficient GPU-accelerated performance, while the Ryzen 9 6900HS manages CPU-bound tasks with good efficiency.
Hardware Specs
- CPU: AMD Ryzen 9 6900HS, 8 cores (3.30 GHz)
- RAM: 16 GB
- GPU: NVIDIA GeForce RTX 3060 Laptop GPU
Estimated Total Price (2023): $1,099.99
Price-to-Performance: 1,168.5 per $100
The ASUS ROG Zephyrus G15 achieves a price-to-performance ratio of 1,168.5 per $100, based on its overall single precision GPU benchmark score of 12,858 with the NVIDIA GeForce RTX 3060.
Price-to-Performance Calculation:
The price-to-performance ratio is calculated by dividing the single precision GPU benchmark score by the system price, then multiplying by 100.
Calculation: (12858 / 1099.99) × 100 ≈ 1,168.5 per $100
DirectML GPU Benchmarks – NVIDIA GeForce RTX 3060
This benchmark focuses on the NVIDIA GeForce RTX 3060 Laptop GPU, highlighting its performance in tasks such as Style Transfer, Pose Estimation, and Image Segmentation. The RTX 3060 offers impressive AI capabilities, especially in single and half precision tasks, making it suitable for a variety of AI-driven workloads.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 12,858
- Half Precision: 21,751
- Quantized: 8,784
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 113,920 | 146.4 | 4. Depth Estimation | 31,569 | 243.2 |
| 2. Pose Estimation | 67,164 | 78.4 | 5. Face Detection | 13,772 | 163.6 |
| 3. Image Segmentation | 8,126 | 131.7 | 6. Object Detection | 6,502 | 515.8 |
DirectML GPU Benchmarks – AMD Radeon Graphics
This test measures the performance of the AMD Radeon™ Graphics found in the Ryzen 9 6900HS processor. While it performs reasonably well for integrated graphics, it is outclassed by the dedicated RTX 3060 when handling complex AI workloads.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 5,674
- Half Precision: 8,417
- Quantized: 3,961
ONNX CPU Benchmarks
This test measures the AI performance of the AMD Ryzen 9 6900HS CPU. The Ryzen 9 delivers solid results in single and quantized precision tasks, particularly in Style Transfer and Object Detection, but the GPU handles more intensive workloads with greater efficiency.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 2,902
- Half Precision: 794
- Quantized: 3,105
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 15,008 | 19.3 | 4. Depth Estimation | 7,083 | 54.6 |
| 2. Pose Estimation | 6,151 | 7.18 | 5. Face Detection | 2,315 | 27.5 |
| 3. Object Detection | 2,305 | 182.9 | 6. Image Segmentation | 593 | 9.61 |
OpenVINO CPU Benchmarks
In this benchmark using the OpenVINO Backend, the AMD Ryzen 9 6900HS shows balanced performance in quantized tasks, excelling particularly in Style Transfer, Pose Estimation, and Face Detection. However, the CPU lacks the throughput seen with the RTX 3060 GPU.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 4,218
- Half Precision: 4,219
- Quantized: 7,217
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 18,351 | 23.6 | 4. Depth Estimation | 8,186 | 63.1 |
| 2. Pose Estimation | 6,618 | 7.72 | 5. Face Detection | 5,995 | 71.2 |
| 3. Image Classification | 3,325 | 618.3 | 6. Object Detection | 3,067 | 243.3 |
Final Summary
The ASUS ROG Zephyrus G15 with AMD Ryzen 9 6900HS and NVIDIA GeForce RTX 3060 Laptop GPU provides excellent performance in AI tasks, particularly in Style Transfer, Pose Estimation, and Depth Estimation. The RTX 3060 offers efficient GPU-accelerated results, while the Ryzen 9 6900HS provides strong CPU support for less GPU-intensive tasks. This system is a great choice for users requiring a balance of both GPU and CPU AI performance.