Quick/Comparisons
The GeForce RTX 5090's superior gaming performance is due to its 100 score, significantly outpacing the Radeon RX 7900 XTX's 81.
The GeForce RTX 5090's strong workstation performance earns it the win, with a 93 score compared to the Radeon RX 7900 XTX's 75.
The GeForce RTX 5090's capable AI/ML compute capabilities are reflected in its 95 score, outperforming the Radeon RX 7900 XTX's 74.
100 | Gaming | 81 |
93 | Workstation | 75 |
95 | AI/ML |
Compare with something else?
Compare GeForce RTX 5090 with another →74 |
49 | Energy Efficiency | 43 |
96 | Quick Comparison Final Score | 77 |
| GB202-300 | Graphics Processor | Navi 31 |
| 21760 | Cores | 6144 |
| 680 / 176 | TMUs / ROPs | 384 / 192 |
| 15010.34 | Blender GPU | 4055.35 |
| 244 | Forza Horizon 5 | 174 |
| 323 | The Witcher 3 | 256 |
| 265 | Counter-Strike 2 | 232 |
| 202 | Far Cry 6 | 182 |
| 207 | Hogwarts Legacy | 137 |
| 327 | Call of Duty | 255 |
| 179 | Ghost of Tsushima | 138 |
| 229 | Cyberpunk 2077 | 181 |
| 299 | Tomb Raider | 277 |
| 253 | Average FPS | 204 |
| 294 | 1080p High | 241 |
| 253 | 1080p Ultra | 204 |
| 207 | 1440p Ultra | 159 |
| 124 | 4K Ultra | 94 |
| Low | Margin of Error | Low |
| 419309 | GB6 Compute Score | 224686 |
| 333 img/sec | Background Blur | 423.9 img/sec |
| 298.5 img/sec | Face Detection | 270.8 img/sec |
| 21.7 Gpixels/sec | Horizon Detection | 9.23 Gpixels/sec |
| 33.7 Gpixels/sec | Edge Detection | 13.7 Gpixels/sec |
| 34.5 Gpixels/sec | Gaussian Blur | 11.1 Gpixels/sec |
| 2.53 Gpixels/sec | Feature Matching | 1.74 Gpixels/sec |
| 2850 Gpixels/sec | Stereo Matching | 871 Gpixels/sec |
| 59146.9 FPS | Particle Physics | 29891.2 FPS |
| OpenCL | API | OpenCL |
| 38955 | G3D Mark Score | 31406 |
| 1413 | G2D Mark | 1285 |
| 341 FPS | DirectX 11 | 356 FPS |
| 178 FPS | DirectX 12 | 133 FPS |
| 24632 Ops/s | GPU Compute | 17768 Ops/s |
| 114685 | Face Detection (HP) | 71321 |
| 344729 | Pose Estimation (Q) | 257258 |
| 4464 | Text Classification (SP) | 5240 |
| 8184 | Machine Translation (HP) | 7425 |
| 27802 | Object Detection (SP) | 21335 |
| 85572 | Depth Estimation (Q) | 66140 |
| 687080 | Style Transfer (SP) | 483867 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 53861 | GB6 ML Single Precision | 42092 |
| 81080 | GB6 ML Half Precision | 50673 |
| 38938 | GB6 ML Quantized | 32874 |
| 18948 | Image Classification (SP) | 18608 |
| 55109 | Image Segmentation (HP) | 25385 |
| 58478 | Image Super Resolution (Q) | 45641 |
| 424 GPixel/s | Pixel Fill Rate | 480 GPixel/s |
| 1637 GTexel/s | Texture Fill Rate | 959 GTexel/s |
| 104.8 TFLOPS | FLOPS (FP32) | 61.4 TFLOPS |
| GDDR7 | Memory Type | GDDR6 |
| 32 GB | Memory Size | 24 GB |
| 1750 MHz | Memory Clock | 2500 MHz |
| 28000 Mbps | Effective Memory Speed | 20000 Mbps |
| 512-bit | Bus | 384-bit |
| No | ECC | No |
| 1792 GB/s | Memory Bandwidth | 960 GB/s |
| PCIe 5.0 x16 | Interface | PCIe 4.0 x16 |
| 575 W | TGP | 355 W |
| TSMC | Manufacturing | TSMC |
| 4 nm | Fabrication Process | 5 nm |
| 750 mm² | Die Size | 529 mm² |
| 92.2 billion | Transistor Count | 57 billion |
| 122.93 MTr/mm² | Transistor Density | 107.75 MTr/mm² |
| 12.2 | DirectX | 12 |
| 1.4 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.2 |
| 12 | CUDA | — |
| Yes | Ray Tracing | Yes |
| DLSS 4 | DLSS | No |
| 2.1b | DisplayPort | 2.1 |
| Nvidia | Vendor | Amd |
| Discrete | Build | Discrete |
| January 7, 2025 | Released | November 3, 2022 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| Blackwell 2.0 | Architecture | RDNA 3.0 |
| GB202-300 | GPU Codename | Navi 31 |
| -Intel Core Ultra 9 285Kor above | Recommended CPU | -Intel Core i7 13700Kor above |
| 52638 | Steel Nomad Lite Score | 29870 |
| 46941 | Time Spy | 30766 |
| 235224 | Solar Bay | 123758 |
| 38113 | Port Royal | 16749 |
| 87433 | Fire Strike | 73217 |
| 108736 | Wild Life Extreme | 58530 |
| 207058 | Night Raid | 184767 |
| 2017 MHz | Base Clock | 1929 MHz |
| 2407 MHz | Boost Clock | 2498 MHz |
| 21760 | Shading Units | 6144 |
| 680 | Texture Mapping Units (TMUs) | 384 |
| 176 | Render Output Units (ROPs) | 192 |
| 170 | Compute Units (Pipelines) | — |
| 680 | Tensor Cores | — |
| 170 | Ray-tracing Cores | 96 |
| 128KB per cluster | L1 Cache | 256KB per cluster |
| 96MB shared | L2 Cache | 6MB shared |
| 2 IPC | Instructions Per Cycle | 4 IPC |