Key Differences
Advantages of Apple M4 Max GPU (40-core)
Advantages of GeForce RTX 4070 SUPER
Specifications
Performance
| TGP | 50 W | 220 W |
Storage Details
| Bus | 512-bit | 192-bit |
Ports
| DisplayPort | - | 1.4a |
Other
| Forza Horizon 5 | - | 18313111992 |
| The Witcher 3 | - | 20517812969 |
| Counter-Strike 2 | - | 28620815584 |
| Far Cry 6 | - | 15414212577 |
| Hogwarts Legacy | - | 1341088846 |
| CoD: Modern Warfare III | - | 17515912481 |
| Ghost of Tsushima | - | 1191018351 |
| Cyberpunk 2077 | - | 1481368942 |
| Shadow of the Tomb Raider | - | 24322217393 |
| 1080p High | - | 183 |
| 1080p Ultra | - | 154 |
| 1440p Ultra | - | 121 |
| 4K Ultra | - | 71 |
| Margin of Error | High | Low |
| Time Spy | - | 20886 |
| Solar Bay | 31561 | 98132 |
| Port Royal | - | 13122 |
| Fire Strike | - | 50125 |
| Wild Life Extreme | 20029 | 41653 |
| Night Raid | - | 164996 |
| Background Blur | 195 img/sec | 359.5 img/sec |
| Face Detection | 125.9 img/sec | 210 img/sec |
| Horizon Detection | 4.82 Gpixels/sec | 7.61 Gpixels/sec |
| Edge Detection | 7.47 Gpixels/sec | 10.4 Gpixels/sec |
| Gaussian Blur | 5.8 Gpixels/sec | 11.3 Gpixels/sec |
| Feature Matching | 1.06 Gpixels/sec | 1.97 Gpixels/sec |
| Stereo Matching | 417.4 Gpixels/sec | 969.3 Gpixels/sec |
| Particle Physics | 17079.9 FPS | 29061.4 FPS |
| API | Metal | OpenGL |
| G2D Mark | - | 1189 |
| DirectX 11 | - | 272 FPS |
| DirectX 12 | - | 110 FPS |
| GPU Compute | - | 17043 Ops/s |
| Aztec Ruins Normal Tier | 1316.3 FPS | 1132.4 FPS |
| Car Chase | 950.9 FPS | 981.8 FPS |
| Manhattan 3.1.1 (1440p) | 911 FPS | 840.6 FPS |
| T-Rex | 3917.4 FPS | 4325.3 FPS |
| Tessellation | - | 7090.8 FPS |
| ALU 2 | 1756 FPS | 6890.2 FPS |
| Driver Overhead 2 | 654.4 FPS | 290.1 FPS |
| OS | OS X | Windows |
| Image Classification (SP) | 10302 | 12146 |
| Image Segmentation (HP) | 23763 | 23932 |
| Image Super Resolution (Q) | 29314 | 36321 |
| Face Detection (HP) | 42016 | 52836 |
| Pose Estimation (Q) | 105972 | 179788 |
| Text Classification (SP) | 2963 | 3371 |
| Machine Translation (HP) | 6594 | 5017 |
| Object Detection (SP) | 8935 | 15164 |
| Depth Estimation (Q) | 43143 | 62500 |
| Style Transfer (SP) | 221563 | 357489 |
| Framework | Core ML | ONNX |
| Backend | GPU | DirectML |
| Vendor | Apple | Nvidia |
| Build | Integrated | Discrete |
| Released | October 30, 2024 | January 17, 2024 |
| Launch price (MSRP) | - | $599 |
| Case | Laptop | Desktop |
| Purpose | Professional | Gaming |
| Segment | Mid-range | Mid-range |
| Architecture | Apple M GPU | Ada Lovelace |
| GPU Codename | Custom | AD104 |
| Rival Equivalent | GeForce RTX 4070 Laptop | Radeon RX 7900 XT |
| Base Clock | 500 MHz | 1980 MHz |
| Boost Clock | 1800 MHz | 2475 MHz |
| Shading Units | 5120 | 7168 |
| Texture Mapping Units (TMUs) | 320 | 224 |
| Render Output Units (ROPs) | 160 | 80 |
| Compute Units (Pipelines) | 640 | 56 |
| Tensor Cores | - | 224 |
| Ray-tracing Cores | - | 56 |
| L1 Cache | - | 128KB per cluster |
| L2 Cache | - | 48MB shared |
| Instructions Per Cycle | 2 IPC | 2 IPC |
| Pixel Fill Rate | 288 GPixel/s | 198 GPixel/s |
| Texture Fill Rate | 576 GTexel/s | 554 GTexel/s |
| Interface | Custom | PCIe 4.0 x16 |
| Manufacturing | TSMC | TSMC |
| Fabrication Process | 3 nm | 5 nm |
| Die Size | - | 294 mm² |
| Transistor Count | - | 35 billion |
| Transistor Density | - | 119.05 MTr/mm² |
| Max. Temperature | 100°C | - |
| Memory Type | System Shared | GDDR6X |
| Memory Size | - | 12 GB |
| Memory Clock | 8533 MHz | 1313 MHz |
| Effective Memory Speed | - | 21000 Mbps |
| ECC | No | No |
| DirectX | - | 12 |
| Vulkan | - | 1.3 |
| OpenGL | - | 4.6 |
| OpenCL | - | 3.0 |
| CUDA | - | 8.9 |
| Ray Tracing | Yes | Yes |
| DLSS | No | DLSS 3 |