Quick/Comparisons
14 | Gaming | 51 |
20 | Workstation | 62 |
14 | AI/ML |
Compare with something else?
Compare GeForce RTX 2050 with another →| — |
75 | Energy Efficiency | 49 |
26 | Quick Comparison Final Score | — |
| GA107 | Graphics Processor | BMG-G31 |
| 2048 | Cores | 4096 |
| 64 / 32 | TMUs / ROPs | 256 / 128 |
| 749.91 | Blender GPU | 3324.51 |
| 35 | Forza Horizon 5 | — |
| 41 | The Witcher 3 | — |
| 60 | Counter-Strike 2 | — |
| 44 | Far Cry 6 | — |
| 23 | Hogwarts Legacy | — |
| 32 | Call of Duty | — |
| 21 | Ghost of Tsushima | — |
| 26 | Cyberpunk 2077 | — |
| 50 | Tomb Raider | — |
| 37 | Average FPS | — |
| 47 | 1080p High | — |
| 37 | 1080p Ultra | — |
| 25 | 1440p Ultra | — |
| 13 | 4K Ultra | — |
| Low | Margin of Error | — |
| 42967 | GB6 Compute Score | 179586 |
| 73.3 img/sec | Background Blur | — |
| 39.7 img/sec | Face Detection | — |
| 1.65 Gpixels/sec | Horizon Detection | — |
| 2.13 Gpixels/sec | Edge Detection | — |
| 2.06 Gpixels/sec | Gaussian Blur | — |
| 0.55 Gpixels/sec | Feature Matching | — |
| 191.1 Gpixels/sec | Stereo Matching | — |
| 5921.5 FPS | Particle Physics | — |
| OpenCL | API | — |
| 7710 | G3D Mark Score | 17577 |
| 480 | G2D Mark | 601 |
| 47 FPS | DirectX 11 | 138 FPS |
| 42 FPS | DirectX 12 | 96 FPS |
| 3502 Ops/s | GPU Compute | 8261 Ops/s |
| 7807 | GB6 ML Single Precision | — |
| 14632 | GB6 ML Half Precision | — |
| 5491 | GB6 ML Quantized | — |
| 3489 | Image Classification (SP) | — |
| 7528 | Image Segmentation (HP) | — |
| 6509 | Image Super Resolution (Q) | — |
| 13217 | Face Detection (HP) | — |
| 24531 | Pose Estimation (Q) | — |
| 2148 | Text Classification (SP) | — |
| 2764 | Machine Translation (HP) | — |
| 3948 | Object Detection (SP) | — |
| 13355 | Depth Estimation (Q) | — |
| 60626 | Style Transfer (SP) | — |
| ONNX | Framework | — |
| DirectML | Backend | — |
| 40-47 GPixel/s | Pixel Fill Rate | 358 GPixel/s |
| 80-95 GTexel/s | Texture Fill Rate | 717 GTexel/s |
| 5.1-6 TFLOPS | FLOPS (FP32) | 22.9 TFLOPS |
| GDDR6 | Memory Type | GDDR6 |
| 4 GB | Memory Size | 32 GB |
| 1750 MHz | Memory Clock | 2375 MHz |
| 14000 Mbps | Effective Memory Speed | 19000 Mbps |
| 64-bit | Bus | 256-bit |
| No | ECC | No |
| 112 GB/s | Memory Bandwidth | 608 GB/s |
| PCIe 3.0 x8 | Interface | PCIe 5.0 x16 |
| 30-45 W (configurable) | TGP | 230 W |
| Samsung | Manufacturing | TSMC |
| 8 nm | Fabrication Process | 5 nm |
| 200 mm² | Die Size | — |
| 8 billion | Transistor Count | — |
| 40 MTr/mm² | Transistor Density | — |
| 12 | DirectX | 12.2 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 8.6 | CUDA | No |
| Yes | Ray Tracing | Yes |
| No | DLSS | No |
| 1.4a | DisplayPort | 2.1a |
| Nvidia | Vendor | Intel |
| Discrete | Build | Discrete |
| December 17, 2021 | Released | March 25, 2026 |
| Laptop | Case | Desktop |
| Gaming | Purpose | Professional |
| Mid-range | Segment | High-end |
| Ampere | Architecture | Battlemage Pro Series |
| GA107 | GPU Codename | BMG-G31 |
| -Intel Core i5 11500Hor above | Recommended CPU | — |
| 3959 | Steel Nomad Lite Score | — |
| 3924 | Time Spy | 19094 |
| 18938 | Solar Bay | — |
| 446 | Port Royal | — |
| 10490 | Fire Strike | 47615 |
| 7298 | Wild Life Extreme | — |
| 45541 | Night Raid | — |
| 735-1185 MHz (configurable) | Base Clock | 2280 MHz |
| 1245-1477 MHz (configurable) | Boost Clock | 2800 MHz |
| 2048 | Shading Units | 4096 |
| 64 | Texture Mapping Units (TMUs) | 256 |
| 32 | Render Output Units (ROPs) | 128 |
| 16 | Compute Units (Pipelines) | 32 |
| 64 | Tensor Cores | 256 |
| 16 | Ray-tracing Cores | 32 |
| 64KB per cluster | L1 Cache | 256KB per cluster |
| 2MB shared | L2 Cache | 16MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |