Quick/Comparisons
The Radeon RX 7900 XT takes the lead in gaming performance, thanks to its higher benchmark score of 72 compared to the GeForce RTX 3090 Ti's 66.
The GeForce RTX 3090 Ti edges out the Radeon RX 7900 XT in workstation performance, with a higher score of 73 compared to the Radeon's 70.
The Radeon RX 7900 XT outperforms the GeForce RTX 3090 Ti in AI/ML tasks, boasting a higher score of 68 compared to the GeForce's 52.
66 | Gaming | 72 |
73 | Workstation | 70 |
52 | AI/ML | 68 |
37 | Energy Efficiency | 45 |
66 | Quick Comparison Final Score | 70 |
| GA102 | Graphics Processor | Navi 31 |
| 10752 | Cores | 5376 |
| 336 / 112 | TMUs / ROPs | 336 / 192 |
| 6057.83 | Blender GPU | 3548.59 |
| 141 | Forza Horizon 5 | 157 |
| 182 | The Witcher 3 | 228 |
| 210 | Counter-Strike 2 | 223 |
| 147 | Far Cry 6 | 173 |
| 112 | Hogwarts Legacy | 126 |
| 173 | Call of Duty | 222 |
| 108 | Ghost of Tsushima | 129 |
| 143 | Cyberpunk 2077 | 158 |
| 243 | Tomb Raider | 247 |
| 162 | Average FPS | 185 |
| 128 | 1440p Ultra | 144 |
| 75 | 4K Ultra | 85 |
| Low | Margin of Error | Low |
| 193 | 1080p High | 219 |
| 162 | 1080p Ultra | 185 |
| 232083 | GB6 Compute Score | 203754 |
| 362.7 img/sec | Background Blur | 419.8 img/sec |
| 227.9 img/sec | Face Detection | 240.5 img/sec |
| 11.3 Gpixels/sec | Horizon Detection | 8.72 Gpixels/sec |
| 15.9 Gpixels/sec | Edge Detection | 12 Gpixels/sec |
| 12.1 Gpixels/sec | Gaussian Blur | 9.88 Gpixels/sec |
| 1.56 Gpixels/sec | Feature Matching | 1.71 Gpixels/sec |
| 1080 Gpixels/sec | Stereo Matching | 701.7 Gpixels/sec |
| 31024.6 FPS | Particle Physics | 26631.4 FPS |
| OpenCL | API | OpenCL |
| 29288 | G3D Mark Score | 29015 |
| 1226 | G2D Mark | 1241 |
| 238 FPS | DirectX 11 | 340 FPS |
| 122 FPS | DirectX 12 | 111 FPS |
| 16569 Ops/s | GPU Compute | 16851 Ops/s |
| 26701 | GB6 ML Single Precision | 39099 |
| 38573 | GB6 ML Half Precision | 47298 |
| 15995 | GB6 ML Quantized | 30875 |
| 10246 | Image Classification (SP) | 16951 |
| 22830 | Image Segmentation (HP) | 23234 |
| 23532 | Image Super Resolution (Q) | 41128 |
| 48657 | Face Detection (HP) | 67615 |
| 147241 | Pose Estimation (Q) | 237026 |
| 2917 | Text Classification (SP) | 4961 |
| 4427 | Machine Translation (HP) | 7043 |
| 13323 | Object Detection (SP) | 18737 |
| 40055 | Depth Estimation (Q) | 67746 |
| 298683 | Style Transfer (SP) | 457919 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 208 GPixel/s | Pixel Fill Rate | 460 GPixel/s |
| 625 GTexel/s | Texture Fill Rate | 804 GTexel/s |
| 40 TFLOPS | FLOPS (FP32) | 51.5 TFLOPS |
| GDDR6X | Memory Type | GDDR6 |
| 24 GB | Memory Size | 20 GB |
| 1313 MHz | Memory Clock | 2500 MHz |
| 21000 Mbps | Effective Memory Speed | 20000 Mbps |
| 384-bit | Bus | 320-bit |
| No | ECC | No |
| 1008 GB/s | Memory Bandwidth | 800 GB/s |
| PCIe 4.0 x16 | Interface | PCIe 4.0 x16 |
| 450 W | TGP | 300 W |
| Samsung | Manufacturing | TSMC |
| 8 nm | Fabrication Process | 5 nm |
| 628 mm² | Die Size | 529 mm² |
| 28 billion | Transistor Count | 57 billion |
| 44.59 MTr/mm² | Transistor Density | 107.75 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.2 |
| 8.6 | CUDA | — |
| Yes | Ray Tracing | Yes |
| DLSS 2 | DLSS | No |
| 1.4a | DisplayPort | 2.1 |
| Nvidia | Vendor | Amd |
| Discrete | Build | Discrete |
| March 29, 2022 | Released | December 13, 2023 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| Ampere | Architecture | RDNA 3.0 |
| GA102 | GPU Codename | Navi 31 |
| -Intel Core i9 12900Kor above | Recommended CPU | -Intel Core i7 13700Kor above |
| 24378 | Steel Nomad Lite Score | 24675 |
| 21854 | Time Spy | 27067 |
| 109167 | Solar Bay | 104912 |
| 14888 | Port Royal | 14358 |
| 51943 | Fire Strike | 67256 |
| 48864 | Wild Life Extreme | 48319 |
| 156500 | Night Raid | 176267 |
| 1560 MHz | Base Clock | 1387 MHz |
| 1860 MHz | Boost Clock | 2394 MHz |
| 10752 | Shading Units | 5376 |
| 336 | Texture Mapping Units (TMUs) | 336 |
| 112 | Render Output Units (ROPs) | 192 |
| 84 | Compute Units (Pipelines) | — |
| 336 | Tensor Cores | — |
| 84 | Ray-tracing Cores | 84 |
| 128KB per cluster | L1 Cache | 256KB per cluster |
| 6MB shared | L2 Cache | 6MB shared |
| 2 IPC | Instructions Per Cycle | 4 IPC |
Compare with something else?
Compare GeForce RTX 3090 with another →