Quick/Comparisons
14 | Gaming | 14 |
24 | Workstation | 20 |
| — | AI/ML |
Compare with something else?
Compare Apple M4 GPU with another →97 | Energy Efficiency | 75 |
32 | Quick Comparison Final Score | 26 |
| Custom | Graphics Processor | GA107 |
| 1024 | Cores | 2048 |
| 64 / 32 | TMUs / ROPs | 64 / 32 |
| 1060.26 | Blender GPU | 749.91 |
| — | Forza Horizon 5 | 35 |
| — | The Witcher 3 | 41 |
| — | Counter-Strike 2 | 60 |
| — | Far Cry 6 | 44 |
| — | Hogwarts Legacy | 23 |
| — | Call of Duty | 32 |
| — | Ghost of Tsushima | 21 |
| — | Cyberpunk 2077 | 26 |
| — | Tomb Raider | 50 |
| — | Average FPS | 37 |
| — | 1080p High | 47 |
| — | 1080p Ultra | 37 |
| — | 1440p Ultra | 25 |
| — | 4K Ultra | 13 |
| — | Margin of Error | Low |
| 30791 | GB6 Compute Score | 42967 |
| 57.4 img/sec | Background Blur | 73.3 img/sec |
| 41.3 img/sec | Face Detection | 39.7 img/sec |
| 1.21 Gpixels/sec | Horizon Detection | 1.65 Gpixels/sec |
| 1.74 Gpixels/sec | Edge Detection | 2.13 Gpixels/sec |
| 1.12 Gpixels/sec | Gaussian Blur | 2.06 Gpixels/sec |
| 0.43 Gpixels/sec | Feature Matching | 0.55 Gpixels/sec |
| 94.9 Gpixels/sec | Stereo Matching | 191.1 Gpixels/sec |
| 3902.5 FPS | Particle Physics | 5921.5 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 7710 |
| — | G2D Mark | 480 |
| — | DirectX 11 | 47 FPS |
| — | DirectX 12 | 42 FPS |
| — | GPU Compute | 3502 Ops/s |
| — | GB6 ML Single Precision | 7807 |
| — | GB6 ML Half Precision | 14632 |
| — | GB6 ML Quantized | 5491 |
| — | Image Classification (SP) | 3489 |
| — | Image Segmentation (HP) | 7528 |
| — | Image Super Resolution (Q) | 6509 |
| — | Face Detection (HP) | 13217 |
| — | Pose Estimation (Q) | 24531 |
| — | Text Classification (SP) | 2148 |
| — | Machine Translation (HP) | 2764 |
| — | Object Detection (SP) | 3948 |
| — | Depth Estimation (Q) | 13355 |
| — | Style Transfer (SP) | 60626 |
| — | Framework | ONNX |
| — | Backend | DirectML |
| 47 GPixel/s | Pixel Fill Rate | 40-47 GPixel/s |
| 94 GTexel/s | Texture Fill Rate | 80-95 GTexel/s |
| 3 TFLOPS | FLOPS (FP32) | 5.1-6 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 4 GB |
| 7500 MHz | Memory Clock | 1750 MHz |
| 15000 Mbps | Effective Memory Speed | 14000 Mbps |
| 128-bit | Bus | 64-bit |
| No | ECC | No |
| 120 GB/s | Memory Bandwidth | 112 GB/s |
| Custom | Interface | PCIe 3.0 x8 |
| 15 W | TGP | 30-45 W (configurable) |
| TSMC | Manufacturing | Samsung |
| 3 nm | Fabrication Process | 8 nm |
| — | Die Size | 200 mm² |
| — | Transistor Count | 8 billion |
| — | Transistor Density | 40 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.3 |
| — | OpenGL | 4.6 |
| — | OpenCL | 3 |
| — | CUDA | 8.6 |
| Yes | Ray Tracing | Yes |
| No | DLSS | No |
| — | DisplayPort | 1.4a |
| Apple | Vendor | Nvidia |
| Integrated | Build | Discrete |
| October 29, 2024 | Released | December 17, 2021 |
| Laptop | Case | Laptop |
| Professional | Purpose | Gaming |
| Mid-range | Segment | Mid-range |
| Apple M GPU | Architecture | Ampere |
| Custom | GPU Codename | GA107 |
| -Apple M4 (8-Core)or above | Recommended CPU | -Intel Core i5 11500Hor above |
| 3450 | Steel Nomad Lite Score | 3959 |
| — | Time Spy | 3924 |
| 13908 | Solar Bay | 18938 |
| — | Port Royal | 446 |
| — | Fire Strike | 10490 |
| 7928 | Wild Life Extreme | 7298 |
| — | Night Raid | 45541 |
| 500 MHz | Base Clock | 735-1185 MHz (configurable) |
| 1470 MHz | Boost Clock | 1245-1477 MHz (configurable) |
| 1024 | Shading Units | 2048 |
| 64 | Texture Mapping Units (TMUs) | 64 |
| 32 | Render Output Units (ROPs) | 32 |
| 128 | Compute Units (Pipelines) | 16 |
| — | Tensor Cores | 64 |
| — | Ray-tracing Cores | 16 |
| — | L1 Cache | 64KB per cluster |
| — | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |