Quick/Comparisons
61 | Gaming | 7 |
69 | Workstation | 13 |
53 | AI/ML |
Compare with something else?
Compare GeForce RTX 3090 with another →8 |
35 | Energy Efficiency | 93 |
62 | Quick Comparison Final Score | 23 |
| GA102 | Graphics Processor | Phoenix |
| 10496 | Cores | 256 |
| 328 / 112 | TMUs / ROPs | 16 / 8 |
| 5800.59 | Blender GPU | 97.23 |
| 125 | Forza Horizon 5 | 14 |
| 175 | The Witcher 3 | 12 |
| 212 | Counter-Strike 2 | 20 |
| 142 | Far Cry 6 | 17 |
| 106 | Hogwarts Legacy | - |
| 159 | Call of Duty | 19 |
| 97 | Ghost of Tsushima | 10 |
| 132 | Cyberpunk 2077 | 10 |
| 226 | Tomb Raider | 18 |
| 153 | Average FPS | 14 |
| 179 | 1080p High | 17 |
| 153 | 1080p Ultra | 14 |
| 119 | 1440p Ultra | 10 |
| 70 | 4K Ultra | - |
| Low | Margin of Error | High |
| 211669 | GB6 Compute Score | 12774 |
| 320.8 img/sec | Background Blur | 31.6 img/sec |
| 207.2 img/sec | Face Detection | 15.3 img/sec |
| 10.1 Gpixels/sec | Horizon Detection | 0.44 Gpixels/sec |
| 14.4 Gpixels/sec | Edge Detection | 0.56 Gpixels/sec |
| 11.5 Gpixels/sec | Gaussian Blur | 0.47 Gpixels/sec |
| 1.48 Gpixels/sec | Feature Matching | 0.17 Gpixels/sec |
| 984.2 Gpixels/sec | Stereo Matching | 47 Gpixels/sec |
| 27764.9 FPS | Particle Physics | 1815.3 FPS |
| OpenCL | API | OpenCL |
| 26551 | G3D Mark Score | 2981 |
| 1069 | G2D Mark | 786 |
| 218 FPS | DirectX 11 | 25 FPS |
| 110 FPS | DirectX 12 | 16 FPS |
| 15117 Ops/s | GPU Compute | 1643 Ops/s |
| 27237 | GB6 ML Single Precision | 4462 |
| 41500 | GB6 ML Half Precision | 5298 |
| 15927 | GB6 ML Quantized | 3419 |
| 10665 | Image Classification (SP) | 1608 |
| 28963 | Image Segmentation (HP) | 2046 |
| 22697 | Image Super Resolution (Q) | 5074 |
| 55232 | Face Detection (HP) | 5435 |
| 128155 | Pose Estimation (Q) | 16992 |
| 3152 | Text Classification (SP) | 1469 |
| 5088 | Machine Translation (HP) | 1922 |
| 13891 | Object Detection (SP) | 1516 |
| 37984 | Depth Estimation (Q) | 7572 |
| 272000 | Style Transfer (SP) | 30182 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 190 GPixel/s | Pixel Fill Rate | 20-22 GPixel/s |
| 556 GTexel/s | Texture Fill Rate | 40-45 GTexel/s |
| 35.6 TFLOPS | FLOPS (FP32) | 1.3-1.4 TFLOPS |
| GDDR6X | Memory Type | System Shared |
| 24 GB | Memory Size | — |
| 1219 MHz | Memory Clock | — |
| 19500 Mbps | Effective Memory Speed | — |
| 384-bit | Bus | — |
| No | ECC | No |
| 936.2 GB/s | Memory Bandwidth | — |
| PCIe 4.0 x16 | Interface | PCIe 4.0 x8 |
| 350 W | TGP | 15 W |
| Samsung | Manufacturing | TSMC |
| 8 nm | Fabrication Process | 4 nm |
| 628 mm² | Die Size | 178 mm² |
| 28 billion | Transistor Count | 25.4 billion |
| 44.59 MTr/mm² | Transistor Density | 142.7 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.1 |
| 8.6 | CUDA | — |
| Yes | Ray Tracing | Yes |
| DLSS 2 | DLSS | No |
| 1.4a | DisplayPort | 2.1 |
| Nvidia | Vendor | Amd |
| Discrete | Build | Integrated |
| September 24, 2020 | Released | January 4, 2023 |
| Desktop | Case | Laptop |
| Gaming | Purpose | Generic |
| High-end | Segment | Entry-level |
| Ampere | Architecture | RDNA 3.0 |
| GA102 | GPU Codename | Phoenix |
| -Intel Core i9 12900Kor above | Recommended CPU | -AMD Ryzen 5 8540Uor above |
| 21851 | Steel Nomad Lite Score | 1405 |
| 19896 | Time Spy | 1613 |
| 97375 | Solar Bay | 4770 |
| 13631 | Port Royal | 722 |
| 47515 | Fire Strike | 5043 |
| 43673 | Wild Life Extreme | 3115 |
| 139755 | Night Raid | 21243 |
| 1395 MHz | Base Clock | 800 MHz |
| 1695 MHz | Boost Clock | 2500-2800 MHz (configurable) |
| 10496 | Shading Units | 256 |
| 328 | Texture Mapping Units (TMUs) | 16 |
| 112 | Render Output Units (ROPs) | 8 |
| 82 | Compute Units (Pipelines) | 4 |
| 328 | Tensor Cores | — |
| 82 | Ray-tracing Cores | 4 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 6MB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |