Quick/Comparisons
34 | Gaming | 17 |
41 | Workstation | 27 |
38 | AI/ML |
Compare with something else?
Compare Intel Arc B580 with another →17 |
47 | Energy Efficiency | 61 |
37 | Quick Comparison Final Score | 27 |
| BMG-G21 | Graphics Processor | GA107 |
| 2560 | Cores | 2048 |
| 160 / 80 | TMUs / ROPs | 64 / 32 |
| 2106.04 | Blender GPU | 1142.3 |
| 78 | Forza Horizon 5 | 43 |
| 128 | The Witcher 3 | 53 |
| 116 | Counter-Strike 2 | 79 |
| 98 | Far Cry 6 | 55 |
| 57 | Hogwarts Legacy | 31 |
| 85 | Call of Duty | 39 |
| 61 | Ghost of Tsushima | 26 |
| 82 | Cyberpunk 2077 | 36 |
| 107 | Tomb Raider | 61 |
| 90 | Average FPS | 47 |
| 109 | 1080p High | 59 |
| 90 | 1080p Ultra | 47 |
| 67 | 1440p Ultra | 32 |
| 38 | 4K Ultra | 17 |
| Low | Margin of Error | High |
| 111561 | GB6 Compute Score | 48591 |
| 284.6 img/sec | Background Blur | 80.6 img/sec |
| 150 img/sec | Face Detection | 47.7 img/sec |
| 2.67 Gpixels/sec | Horizon Detection | 2.14 Gpixels/sec |
| 5.22 Gpixels/sec | Edge Detection | 2.82 Gpixels/sec |
| 6.35 Gpixels/sec | Gaussian Blur | 2 Gpixels/sec |
| 1.18 Gpixels/sec | Feature Matching | 0.61 Gpixels/sec |
| 397 Gpixels/sec | Stereo Matching | 199.8 Gpixels/sec |
| 15177.3 FPS | Particle Physics | 6124.5 FPS |
| OpenCL | API | OpenCL |
| 15914 | G3D Mark Score | 9487 |
| 700 | G2D Mark | 524 |
| 126 FPS | DirectX 11 | 65 FPS |
| 77 FPS | DirectX 12 | 47 FPS |
| 7896 Ops/s | GPU Compute | 3856 Ops/s |
| 21661 | GB6 ML Single Precision | 9684 |
| 34463 | GB6 ML Half Precision | 18365 |
| 14366 | GB6 ML Quantized | 6700 |
| 10654 | Image Classification (SP) | 4109 |
| 18608 | Image Segmentation (HP) | 10426 |
| 16715 | Image Super Resolution (Q) | 8368 |
| 40348 | Face Detection (HP) | 18130 |
| 64494 | Pose Estimation (Q) | 32595 |
| 3091 | Text Classification (SP) | 2340 |
| 3530 | Machine Translation (HP) | 3225 |
| 13690 | Object Detection (SP) | 4684 |
| 36797 | Depth Estimation (Q) | 16324 |
| 197057 | Style Transfer (SP) | 70507 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 214 GPixel/s | Pixel Fill Rate | 34-56 GPixel/s |
| 427 GTexel/s | Texture Fill Rate | 68-111 GTexel/s |
| 13.7 TFLOPS | FLOPS (FP32) | 4.3-7.1 TFLOPS |
| GDDR6 | Memory Type | GDDR6 |
| 12 GB | Memory Size | 4 GB |
| 2375 MHz | Memory Clock | 1375 MHz |
| 19000 Mbps | Effective Memory Speed | 11000 Mbps |
| 192-bit | Bus | 128-bit |
| No | ECC | No |
| 456 GB/s | Memory Bandwidth | 176 GB/s |
| PCIe 4.0 x8 | Interface | PCIe 4.0 x8 |
| 190 W | TGP | 35-95 W (configurable) |
| TSMC | Manufacturing | Samsung |
| 5 nm | Fabrication Process | 8 nm |
| 272 mm² | Die Size | 200 mm² |
| 19.6 billion | Transistor Count | 8.7 billion |
| 72.06 MTr/mm² | Transistor Density | 43.5 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| No | CUDA | 8.6 |
| Yes | Ray Tracing | Yes |
| No | DLSS | DLSS 2 |
| 2.1 | DisplayPort | 1.4a |
| BMG-G21 | GPU Codename | GA107 |
| -Intel Core Ultra 5 245Kor above | Recommended CPU | -Intel Core i7 12800Hor above |
| Intel | Vendor | Nvidia |
| Discrete | Build | Discrete |
| December 3, 2024 | Released | March 30, 2022 |
| Desktop | Case | Laptop |
| Gaming | Purpose | Professional |
| Mid-range | Segment | Entry-level |
| Generation 12.7 | Architecture | Ampere |
| 11180 | Steel Nomad Lite Score | 5042 |
| 14665 | Time Spy | 4727 |
| 55557 | Solar Bay | 21233 |
| 7919 | Port Royal | 762 |
| 36685 | Fire Strike | 11952 |
| 28324 | Wild Life Extreme | — |
| 123864 | Night Raid | — |
| 1700 MHz | Base Clock | 712-1530 MHz (configurable) |
| 2670 MHz | Boost Clock | 1057-1740 MHz (configurable) |
| 2560 | Shading Units | 2048 |
| 160 | Texture Mapping Units (TMUs) | 64 |
| 80 | Render Output Units (ROPs) | 32 |
| 20 | Compute Units (Pipelines) | 16 |
| 160 | Tensor Cores | 64 |
| 20 | Ray-tracing Cores | 16 |
| 256KB per cluster | L1 Cache | 128KB per cluster |
| 12MB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |