Quick/Comparisons
14 | Gaming | 44 |
23 | Workstation | 48 |
10 | AI/ML |
Compare with something else?
Compare Apple M3 GPU with another →39 |
95 | Energy Efficiency | 27 |
30 | Quick Comparison Final Score | 44 |
| — | Graphics Processor | TU102 |
| 1280 | Cores | 4352 |
| 80 / 40 | TMUs / ROPs | 272 / 88 |
| 155.27 | Blender GPU | 3046.64 |
| — | Forza Horizon 5 | 101 |
| — | The Witcher 3 | 144 |
| — | Counter-Strike 2 | 199 |
| — | Far Cry 6 | 122 |
| — | Hogwarts Legacy | 82 |
| — | Call of Duty | 118 |
| — | Ghost of Tsushima | 78 |
| — | Cyberpunk 2077 | 98 |
| — | Tomb Raider | 155 |
| — | Average FPS | 122 |
| — | 1080p High | 145 |
| — | 1080p Ultra | 122 |
| — | 1440p Ultra | 92 |
| — | 4K Ultra | 52 |
| — | Margin of Error | Low |
| 30027 | GB6 Compute Score | 144745 |
| 60.3 img/sec | Background Blur | 211 img/sec |
| 40.8 img/sec | Face Detection | 141.3 img/sec |
| 1.21 Gpixels/sec | Horizon Detection | 6.24 Gpixels/sec |
| 1.64 Gpixels/sec | Edge Detection | 8.1 Gpixels/sec |
| 1.2 Gpixels/sec | Gaussian Blur | 9.22 Gpixels/sec |
| 0.37 Gpixels/sec | Feature Matching | 1.46 Gpixels/sec |
| 102.6 Gpixels/sec | Stereo Matching | 627.3 Gpixels/sec |
| 3293.3 FPS | Particle Physics | 16980.5 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 21477 |
| — | G2D Mark | 927 |
| — | DirectX 11 | 192 FPS |
| — | DirectX 12 | 83 FPS |
| — | GPU Compute | 9906 Ops/s |
| 4120 | GB6 ML Single Precision | 19878 |
| 6992 | GB6 ML Half Precision | 31237 |
| 5638 | GB6 ML Quantized | 12663 |
| 1579 | Image Classification (SP) | 8213 |
| 2944 | Image Segmentation (HP) | 19891 |
| 4425 | Image Super Resolution (Q) | 17598 |
| 6661 | Face Detection (HP) | 39785 |
| 15020 | Pose Estimation (Q) | 84003 |
| 2984 | Text Classification (SP) | 2062 |
| 5876 | Machine Translation (HP) | 4192 |
| 1826 | Object Detection (SP) | 10444 |
| 9339 | Depth Estimation (Q) | 31856 |
| 22291 | Style Transfer (SP) | 207773 |
| Core ML | Framework | ONNX |
| CPU | Backend | DirectML |
| 55 GPixel/s | Pixel Fill Rate | 136 GPixel/s |
| 110 GTexel/s | Texture Fill Rate | 420 GTexel/s |
| 3.5 TFLOPS | FLOPS (FP32) | 13.4 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 11 GB |
| 6400 MHz | Memory Clock | 1750 MHz |
| 12800 Mbps | Effective Memory Speed | 14000 Mbps |
| 128-bit | Bus | 352-bit |
| No | ECC | No |
| 102.4 GB/s | Memory Bandwidth | 616 GB/s |
| Custom | Interface | PCIe 3.0 x16 |
| 15 W | TGP | 250 W |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 12 nm |
| — | Die Size | 754 mm² |
| 13.8 billion | Transistor Count | 18 billion |
| — | Transistor Density | 23.87 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.3 |
| — | OpenGL | 4.6 |
| — | OpenCL | 3 |
| — | CUDA | 7.5 |
| Yes | Ray Tracing | Yes |
| No | DLSS | DLSS 2 |
| — | DisplayPort | 1.4a |
| Apple | Vendor | Nvidia |
| Integrated | Build | Discrete |
| October 31, 2023 | Released | September 27, 2018 |
| Laptop | Case | Desktop |
| Professional | Purpose | Gaming |
| Mid-range | Segment | High-end |
| Apple M GPU | Architecture | Turing |
| — | GPU Codename | TU102 |
| -Apple M3or above | Recommended CPU | -Intel Core i9 9900Kor above |
| 3478 | Steel Nomad Lite Score | 14284 |
| — | Time Spy | 14656 |
| 13529 | Solar Bay | 64076 |
| — | Port Royal | 9109 |
| — | Fire Strike | 35666 |
| 8229 | Wild Life Extreme | 29554 |
| — | Night Raid | 115500 |
| 500 MHz | Base Clock | 1350 MHz |
| 1380 MHz | Boost Clock | 1545 MHz |
| 1280 | Shading Units | 4352 |
| 80 | Texture Mapping Units (TMUs) | 272 |
| 40 | Render Output Units (ROPs) | 88 |
| 160 | Compute Units (Pipelines) | 68 |
| — | Tensor Cores | 544 |
| — | Ray-tracing Cores | 68 |
| — | L1 Cache | 64KB per cluster |
| — | L2 Cache | 5632KB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |