Key Differences
Advantages of Apple M4 GPU (10-Core)
Advantages of Intel Arc B580
Specifications
Performance
| TGP | 18 W | 190 W |
Storage Details
| Bus | 128-bit | 192-bit |
Ports
| DisplayPort | - | 2.1 |
Other
| Wild Life Extreme | 9745 | 28283 |
| Background Blur | 72.2 img/sec | 284.6 img/sec |
| Face Detection | 48.3 img/sec | 150 img/sec |
| Horizon Detection | 1.4 Gpixels/sec | 2.67 Gpixels/sec |
| Edge Detection | 1.94 Gpixels/sec | 5.22 Gpixels/sec |
| Gaussian Blur | 1.5 Gpixels/sec | 6.35 Gpixels/sec |
| Feature Matching | 0.53 Gpixels/sec | 1.18 Gpixels/sec |
| Stereo Matching | 123.3 Gpixels/sec | 397 Gpixels/sec |
| Particle Physics | 4938.4 FPS | 15177.3 FPS |
| API | Metal | OpenGL |
| Aztec Ruins Normal Tier | 344.5 FPS | 847.7 FPS |
| Car Chase | 259.9 FPS | 779.3 FPS |
| Manhattan 3.1.1 (1440p) | 167.5 FPS | 608.2 FPS |
| T-Rex | 914.7 FPS | 2266.1 FPS |
| ALU 2 | 432.2 FPS | 3674.6 FPS |
| Driver Overhead 2 | 648.6 FPS | 184.2 FPS |
| OS | OS X | Windows |
| Image Classification (SP) | 4874 | 10654 |
| Image Segmentation (HP) | 8168 | 18608 |
| Image Super Resolution (Q) | 11983 | 16715 |
| Face Detection (HP) | 17036 | 40348 |
| Pose Estimation (Q) | 31171 | 64494 |
| Text Classification (SP) | 2964 | 3091 |
| Machine Translation (HP) | 3410 | 3530 |
| Object Detection (SP) | 5112 | 13690 |
| Depth Estimation (Q) | 22700 | 36797 |
| Style Transfer (SP) | 60872 | 197057 |
| Framework | Core ML | ONNX |
| Backend | GPU | DirectML |
| Vendor | Apple | Intel |
| Build | Integrated | Discrete |
| Released | May 7, 2024 | December 3, 2024 |
| Case | Laptop | Desktop |
| Purpose | Professional | Gaming |
| Segment | Mid-range | Mid-range |
| Architecture | Apple M GPU | Generation 12.7 |
| GPU Codename | Custom | BMG-G21 |
| Rival Equivalent | Adreno X1-85 | GeForce RTX 4060 |
| Base Clock | 500 MHz | 1700 MHz |
| Boost Clock | 1700 MHz | 2670 MHz |
| Shading Units | 1280 | 2560 |
| Texture Mapping Units (TMUs) | 80 | 160 |
| Render Output Units (ROPs) | 40 | 80 |
| Compute Units (Pipelines) | 160 | 20 |
| Instructions Per Cycle | 2 IPC | 2 IPC |
| Pixel Fill Rate | 68 GPixel/s | 214 GPixel/s |
| Texture Fill Rate | 136 GTexel/s | 427 GTexel/s |
| Interface | Custom | PCIe 4.0 x8 |
| Manufacturing | TSMC | TSMC |
| Fabrication Process | 3 nm | 5 nm |
| Transistor Count | 28 billion | 19.6 billion |
| Max. Temperature | 100°C | - |
| Memory Type | System Shared | GDDR6 |
| Memory Clock | 7500 MHz | 2375 MHz |
| Effective Memory Speed | 15000 Mbps | 19000 Mbps |
| ECC | No | No |
| Ray Tracing | Yes | Yes |
| DLSS | No | No |
| Forza Horizon 5 | - | 123786751 |
| The Witcher 3 | - | 1471289146 |
| Counter-Strike 2 | - | 1621168340 |
| Far Cry 6 | - | 108987643 |
| Hogwarts Legacy | - | 71574426 |
| CoD: Modern Warfare III | - | 82856138 |
| Ghost of Tsushima | - | 72614225 |
| Cyberpunk 2077 | - | 90825526 |
| Shadow of the Tomb Raider | - | 1241078343 |
| 1080p High | - | 109 |
| 1080p Ultra | - | 90 |
| 1440p Ultra | - | 67 |
| 4K Ultra | - | 38 |
| Margin of Error | - | Low |
| Time Spy | - | 14654 |
| Solar Bay | - | 55346 |
| Port Royal | - | 7910 |
| Fire Strike | - | 36715 |
| Night Raid | - | 123579 |
| G2D Mark | - | 703 |
| DirectX 11 | - | 129 FPS |
| DirectX 12 | - | 77 FPS |
| GPU Compute | - | 7865 Ops/s |
| Tessellation | - | 4148.9 FPS |
| Launch price (MSRP) | - | $249 |
| Tensor Cores | - | 160 |
| Ray-tracing Cores | - | 20 |
| L1 Cache | - | 256KB per cluster |
| L2 Cache | - | 12MB shared |
| Die Size | - | 272 mm² |
| Transistor Density | - | 72.06 MTr/mm² |
| Memory Size | - | 12 GB |
| DirectX | - | 12 |
| Vulkan | - | 1.3 |
| OpenGL | - | 4.6 |
| OpenCL | - | 3.0 |
| CUDA | - | No |