Quick/Comparisons
The Apple M1 Max GPU (32-core) leads in gaming with a score of 22, thanks to its custom interface and ray-tracing cores.
The Apple M1 Max GPU (32-core) excels in content creation with a score of 37, leveraging its system shared memory and custom interface.
The Apple M1 Max GPU (32-core) takes the lead in AI/ML with a score of 24, despite lacking dedicated AI tensor cores.
22 | Gaming | 12 |
37 | Workstation | 23 |
24 | AI/ML |
Compare with something else?
Compare Apple M1 Max with another →14 |
83 | Energy Efficiency | 72 |
37 | Quick Comparison Final Score | 25 |
| — | Graphics Processor | GA107S |
| 4096 | Cores | 2048 |
| 256 / 128 | TMUs / ROPs | 64 / 32 |
| 933.73 | Blender GPU | 692.94 |
| — | Forza Horizon 5 | 24 |
| — | The Witcher 3 | 26 |
| — | Counter-Strike 2 | 42 |
| — | Far Cry 6 | 33 |
| — | Hogwarts Legacy | 14 |
| — | Call of Duty | 26 |
| — | Ghost of Tsushima | 13 |
| — | Cyberpunk 2077 | 18 |
| — | Tomb Raider | 36 |
| — | Average FPS | 26 |
| — | 1080p High | 35 |
| — | 1080p Ultra | 26 |
| — | 1440p Ultra | 18 |
| — | 4K Ultra | 9 |
| — | Margin of Error | Medium |
| 71977 | GB6 Compute Score | 40574 |
| 120.7 img/sec | Background Blur | 62.4 img/sec |
| 78.9 img/sec | Face Detection | 35.7 img/sec |
| 3.18 Gpixels/sec | Horizon Detection | 1.62 Gpixels/sec |
| 6.15 Gpixels/sec | Edge Detection | 2.11 Gpixels/sec |
| 3.44 Gpixels/sec | Gaussian Blur | 2.05 Gpixels/sec |
| 0.62 Gpixels/sec | Feature Matching | 0.52 Gpixels/sec |
| 229.8 Gpixels/sec | Stereo Matching | 185.1 Gpixels/sec |
| 8822.9 FPS | Particle Physics | 5502.5 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 6570 |
| — | G2D Mark | 416 |
| — | DirectX 11 | 40 FPS |
| — | DirectX 12 | 38 FPS |
| — | GPU Compute | 2911 Ops/s |
| 12213 | GB6 ML Single Precision | 7597 |
| 13810 | GB6 ML Half Precision | 15030 |
| 12739 | GB6 ML Quantized | 5337 |
| 5466 | Image Classification (SP) | 3298 |
| 12415 | Image Segmentation (HP) | 8358 |
| 15930 | Image Super Resolution (Q) | 6219 |
| 23179 | Face Detection (HP) | 14319 |
| 58520 | Pose Estimation (Q) | 22995 |
| 2299 | Text Classification (SP) | 2226 |
| 1713 | Machine Translation (HP) | 3163 |
| 5388 | Object Detection (SP) | 3768 |
| 26458 | Depth Estimation (Q) | 12079 |
| 98278 | Style Transfer (SP) | 53139 |
| Core ML | Framework | ONNX |
| GPU | Backend | DirectML |
| 166 GPixel/s | Pixel Fill Rate | 24-55 GPixel/s |
| 332 GTexel/s | Texture Fill Rate | 48-109 GTexel/s |
| 10.6 TFLOPS | FLOPS (FP32) | 3.1-7 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 4 GB |
| 6400 MHz | Memory Clock | 1500 MHz |
| 12800 Mbps | Effective Memory Speed | 12000 Mbps |
| 512-bit | Bus | 64-bit |
| No | ECC | No |
| 409.6 GB/s | Memory Bandwidth | 96 GB/s |
| Custom | Interface | PCIe 4.0 x8 |
| — | TGP | 20-60 W (configurable) |
| TSMC | Manufacturing | Samsung |
| 5 nm | Fabrication Process | 8 nm |
| — | Die Size | 200 mm² |
| 46.4 billion | Transistor Count | 8.7 billion |
| — | Transistor Density | 43.5 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.3 |
| — | 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 |
| October 18, 2021 | Released | March 22, 2022 |
| Laptop | Case | Laptop |
| Professional | Purpose | Professional |
| High-end | Segment | Entry-level |
| Apple M GPU | Architecture | Ampere |
| — | GPU Codename | GA107S |
| -Apple M1 Maxor above | Recommended CPU | -Intel Core i7 12700Hor above |
| 7412 | Steel Nomad Lite Score | 3021 |
| — | Time Spy | 3023 |
| 22583 | Solar Bay | 13710 |
| — | Port Royal | 625 |
| — | Fire Strike | 10265 |
| 19876 | Wild Life Extreme | 4053 |
| — | Night Raid | 38312 |
| — | Tensor Cores | 64 |
| No | Ray-tracing Cores | 16 |
| — | L1 Cache | 128KB per cluster |
| — | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |
| 450 MHz | Base Clock | 370-890 MHz (configurable) |
| 1296 MHz | Boost Clock | 745-1710 MHz (configurable) |
| 4096 | Shading Units | 2048 |
| 256 | Texture Mapping Units (TMUs) | 64 |
| 128 | Render Output Units (ROPs) | 32 |
| 512 | Compute Units (Pipelines) | 16 |