Quick/Comparisons
14 | Gaming | 21 |
23 | Workstation | 29 |
10 | AI/ML |
Compare with something else?
Compare Apple M3 GPU with another →22 |
95 | Energy Efficiency | 46 |
30 | Quick Comparison Final Score | 27 |
| — | Graphics Processor | TU106 (N18E-G1-B) |
| 1280 | Cores | 1920 |
| 80 / 40 | TMUs / ROPs | 120 / 48 |
| 155.27 | Blender GPU | 1387.02 |
| — | Forza Horizon 5 | 49 |
| — | The Witcher 3 | 66 |
| — | Counter-Strike 2 | 91 |
| — | Far Cry 6 | 61 |
| — | Hogwarts Legacy | 34 |
| — | Call of Duty | 46 |
| — | Ghost of Tsushima | 32 |
| — | Cyberpunk 2077 | 43 |
| — | Tomb Raider | 71 |
| — | Average FPS | 55 |
| — | 1080p High | 72 |
| — | 1080p Ultra | 55 |
| — | 1440p Ultra | 39 |
| — | 4K Ultra | 20 |
| — | Margin of Error | Low |
| 30027 | GB6 Compute Score | 73982 |
| 60.3 img/sec | Background Blur | 150.7 img/sec |
| 40.8 img/sec | Face Detection | 75.5 img/sec |
| 1.21 Gpixels/sec | Horizon Detection | 2.83 Gpixels/sec |
| 1.64 Gpixels/sec | Edge Detection | 3.68 Gpixels/sec |
| 1.2 Gpixels/sec | Gaussian Blur | 4.08 Gpixels/sec |
| 0.37 Gpixels/sec | Feature Matching | 0.89 Gpixels/sec |
| 102.6 Gpixels/sec | Stereo Matching | 257.7 Gpixels/sec |
| 3293.3 FPS | Particle Physics | 9069.6 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 11353 |
| — | G2D Mark | 540 |
| — | DirectX 11 | 96 FPS |
| — | DirectX 12 | 39 FPS |
| — | GPU Compute | 3992 Ops/s |
| 4120 | GB6 ML Single Precision | 10073 |
| 6992 | GB6 ML Half Precision | 17250 |
| 5638 | GB6 ML Quantized | 7165 |
| 1579 | Image Classification (SP) | 4450 |
| 2944 | Image Segmentation (HP) | 9120 |
| 4425 | Image Super Resolution (Q) | 9180 |
| 6661 | Face Detection (HP) | 18866 |
| 15020 | Pose Estimation (Q) | 41027 |
| 2984 | Text Classification (SP) | 1741 |
| 5876 | Machine Translation (HP) | 2269 |
| 1826 | Object Detection (SP) | 4980 |
| 9339 | Depth Estimation (Q) | 18044 |
| 22291 | Style Transfer (SP) | 91554 |
| Core ML | Framework | ONNX |
| CPU | Backend | DirectML |
| 55 GPixel/s | Pixel Fill Rate | 56-58 GPixel/s |
| 110 GTexel/s | Texture Fill Rate | 141-144 GTexel/s |
| 3.5 TFLOPS | FLOPS (FP32) | 4.5-4.6 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 6 GB |
| 6400 MHz | Memory Clock | 1375 MHz |
| 12800 Mbps | Effective Memory Speed | 11000 Mbps |
| 128-bit | Bus | 192-bit |
| No | ECC | No |
| 102.4 GB/s | Memory Bandwidth | 264 GB/s |
| Custom | Interface | PCIe 3.0 x16 |
| 15 W | TGP | 80-115 W (configurable) |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 12 nm |
| — | Die Size | 445 mm² |
| 13.8 billion | Transistor Count | 10.8 billion |
| — | Transistor Density | 24.27 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.2 |
| — | OpenGL | 4.6 |
| — | OpenCL | 1.2 |
| — | CUDA | 7.5 |
| Yes | Ray Tracing | Yes |
| No | DLSS | DLSS 2 |
| — | DisplayPort | 1.4 |
| Apple | Vendor | Nvidia |
| Integrated | Build | Discrete |
| October 31, 2023 | Released | January 29, 2019 |
| Laptop | Case | Laptop |
| Professional | Purpose | Gaming |
| Mid-range | Segment | High-end |
| Apple M GPU | Architecture | Turing |
| — | GPU Codename | TU106 (N18E-G1-B) |
| -Apple M3or above | Recommended CPU | -Intel Core i7 10750Hor above |
| 3478 | Steel Nomad Lite Score | 5764 |
| — | Time Spy | 5959 |
| 13529 | Solar Bay | 25775 |
| — | Port Royal | 3328 |
| — | Fire Strike | 15850 |
| 8229 | Wild Life Extreme | 11601 |
| — | Night Raid | 62117 |
| 500 MHz | Base Clock | 960 MHz |
| 1380 MHz | Boost Clock | 1175-1200 MHz (configurable) |
| 1280 | Shading Units | 1920 |
| 80 | Texture Mapping Units (TMUs) | 120 |
| 40 | Render Output Units (ROPs) | 48 |
| 160 | Compute Units (Pipelines) | 30 |
| — | Tensor Cores | 240 |
| — | Ray-tracing Cores | 30 |
| — | L1 Cache | 64KB per cluster |
| — | L2 Cache | 3MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |