Quick/Comparisons
9 | Gaming | 20 |
23 | Workstation | 26 |
12 | AI/ML |
Compare with something else?
Compare Apple M2 GPU with another →20 |
91 | Energy Efficiency | 59 |
27 | Quick Comparison Final Score | 28 |
| — | Graphics Processor | GA107 (GN20-P1) |
| 1024 | Cores | 2560 |
| 64 / 32 | TMUs / ROPs | 80 / 48 |
| 301.33 | Blender GPU | 1323.51 |
| — | Forza Horizon 5 | 50 |
| — | The Witcher 3 | 59 |
| — | Counter-Strike 2 | 89 |
| — | Far Cry 6 | 58 |
| — | Hogwarts Legacy | 35 |
| — | Call of Duty | 43 |
| — | Ghost of Tsushima | 32 |
| — | Cyberpunk 2077 | 39 |
| — | Tomb Raider | 69 |
| — | Average FPS | 53 |
| — | 1080p High | 67 |
| — | 1080p Ultra | 53 |
| — | 1440p Ultra | 37 |
| — | 4K Ultra | 19 |
| — | Margin of Error | Low |
| 29705 | GB6 Compute Score | 65121 |
| 58.7 img/sec | Background Blur | 107.9 img/sec |
| 35.2 img/sec | Face Detection | 62.6 img/sec |
| 1.1 Gpixels/sec | Horizon Detection | 2.75 Gpixels/sec |
| 1.82 Gpixels/sec | Edge Detection | 3.61 Gpixels/sec |
| 1.2 Gpixels/sec | Gaussian Blur | 2.89 Gpixels/sec |
| 0.37 Gpixels/sec | Feature Matching | 0.81 Gpixels/sec |
| 96 Gpixels/sec | Stereo Matching | 275.9 Gpixels/sec |
| 3825.9 FPS | Particle Physics | 8303.3 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 10086 |
| — | G2D Mark | 498 |
| — | DirectX 11 | 75 FPS |
| — | DirectX 12 | 49 FPS |
| — | GPU Compute | 4322 Ops/s |
| 6206 | GB6 ML Single Precision | 11209 |
| 6893 | GB6 ML Half Precision | 20383 |
| 6515 | GB6 ML Quantized | 7783 |
| 3242 | Image Classification (SP) | 4716 |
| 4890 | Image Segmentation (HP) | 12032 |
| 7149 | Image Super Resolution (Q) | 9965 |
| 10340 | Face Detection (HP) | 20723 |
| 21268 | Pose Estimation (Q) | 41186 |
| 2268 | Text Classification (SP) | 2133 |
| 1122 | Machine Translation (HP) | 3449 |
| 3007 | Object Detection (SP) | 5585 |
| 13941 | Depth Estimation (Q) | 19119 |
| 35943 | Style Transfer (SP) | 86331 |
| Core ML | Framework | ONNX |
| GPU | Backend | DirectML |
| 45 GPixel/s | Pixel Fill Rate | 50-81 GPixel/s |
| 89 GTexel/s | Texture Fill Rate | 83-136 GTexel/s |
| 2.9 TFLOPS | FLOPS (FP32) | 5.3-8.7 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 4 GB |
| 6400 MHz | Memory Clock | 1500 MHz |
| 12800 Mbps | Effective Memory Speed | 12000 Mbps |
| 128-bit | Bus | 128-bit |
| No | ECC | No |
| 102.4 GB/s | Memory Bandwidth | 192 GB/s |
| Custom | Interface | PCIe 4.0 x16 |
| 15 W | TGP | 35-95 W (configurable) |
| TSMC | Manufacturing | Samsung |
| 5 nm | Fabrication Process | 8 nm |
| — | Die Size | 276 mm² |
| 11.6 billion | Transistor Count | 12 billion |
| — | Transistor Density | 43.48 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.2 |
| — | OpenGL | 4.6 |
| — | OpenCL | 3 |
| — | CUDA | 8.6 |
| No | Ray Tracing | Yes |
| No | DLSS | DLSS 2 |
| — | DisplayPort | 1.4a |
| Apple | Vendor | Nvidia |
| Integrated | Build | Discrete |
| June 6, 2022 | Released | January 1, 2021 |
| Laptop | Case | Laptop |
| Professional | Purpose | Gaming |
| Mid-range | Segment | Mid-range |
| Apple M GPU | Architecture | Ampere |
| — | GPU Codename | GA107 (GN20-P1) |
| -Apple M2or above | Recommended CPU | -Intel Core i7 12700Hor above |
| 2312 | Steel Nomad Lite Score | 5684 |
| — | Time Spy | 5383 |
| 7310 | Solar Bay | 24147 |
| — | Port Royal | 570 |
| — | Fire Strike | 13499 |
| 5675 | Wild Life Extreme | 10255 |
| — | Night Raid | 56622 |
| 500 MHz | Base Clock | 735-1463 MHz (configurable) |
| 1398 MHz | Boost Clock | 1035-1695 MHz (configurable) |
| 1024 | Shading Units | 2560 |
| 64 | Texture Mapping Units (TMUs) | 80 |
| 32 | Render Output Units (ROPs) | 48 |
| 128 | Compute Units (Pipelines) | 20 |
| — | Tensor Cores | 80 |
| — | Ray-tracing Cores | 20 |
| — | L1 Cache | 128KB per cluster |
| — | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |