Quick/Comparisons
The GeForce RTX 5090's exceptional gaming performance is unmatched, thanks to its Blackwell 2.0 architecture and 32GB GDDR7 memory.
The GeForce RTX 5090's strong workstation performance and PCIe 5.0 x16 interface give it a slight edge over the Nvidia RTX 5000 Ada.
The GeForce RTX 5090's capable AI/ML compute and 32GB GDDR7 memory make it a clear winner in this use case, outperforming the Nvidia RTX 5000 Ada by a significant margin.
100 | Gaming | 73 |
93 | Workstation | 82 |
95 | AI/ML | 54 |
49 | Energy Efficiency | 45 |
96 | Quick Comparison Final Score | 73 |
| GB202-300 | Graphics Processor | AD102 |
| 21760 | Cores | 12800 |
| 680 / 176 | TMUs / ROPs | 400 / 176 |
| 15010.34 | Blender GPU | 7348.98 |
| 244 | Forza Horizon 5 | 146 |
| 323 | The Witcher 3 | 186 |
| 265 | Counter-Strike 2 | 213 |
| 202 | Far Cry 6 | 151 |
| 207 | Hogwarts Legacy | 117 |
| 327 | Call of Duty | 179 |
| 179 | Ghost of Tsushima | 108 |
| 229 | Cyberpunk 2077 | 148 |
| 299 | Tomb Raider | 242 |
| 253 | Average FPS | 166 |
| 294 | 1080p High | 196 |
| 253 | 1080p Ultra | 166 |
| 207 | 1440p Ultra | 132 |
| 124 | 4K Ultra | 78 |
| Low | Margin of Error | High |
| 419309 | GB6 Compute Score | 260099 |
| 333 img/sec | Background Blur | 286.6 img/sec |
| 298.5 img/sec | Face Detection | 220.4 img/sec |
| 21.7 Gpixels/sec | Horizon Detection | 9.93 Gpixels/sec |
| 33.7 Gpixels/sec | Edge Detection | 13.7 Gpixels/sec |
| 34.5 Gpixels/sec | Gaussian Blur | 18.9 Gpixels/sec |
| 2.53 Gpixels/sec | Feature Matching | 2.2 Gpixels/sec |
| 2850 Gpixels/sec | Stereo Matching | 1580 Gpixels/sec |
| 59146.9 FPS | Particle Physics | 41235.6 FPS |
| OpenCL | API | OpenCL |
| 38955 | G3D Mark Score | 30196 |
| 1413 | G2D Mark | 998 |
| 341 FPS | DirectX 11 | 273 FPS |
| 178 FPS | DirectX 12 | 106 FPS |
| 24632 Ops/s | GPU Compute | 20736 Ops/s |
| 114685 | Face Detection (HP) | 43098 |
| 344729 | Pose Estimation (Q) | 138689 |
| 4464 | Text Classification (SP) | 2893 |
| 8184 | Machine Translation (HP) | 4064 |
| 27802 | Object Detection (SP) | 12467 |
| 85572 | Depth Estimation (Q) | 50352 |
| 687080 | Style Transfer (SP) | 265482 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 53861 | GB6 ML Single Precision | 24715 |
| 81080 | GB6 ML Half Precision | 36753 |
| 38938 | GB6 ML Quantized | 19649 |
| 18948 | Image Classification (SP) | 9943 |
| 55109 | Image Segmentation (HP) | 19696 |
| 58478 | Image Super Resolution (Q) | 32252 |
| 424 GPixel/s | Pixel Fill Rate | 449 GPixel/s |
| 1637 GTexel/s | Texture Fill Rate | 1020 GTexel/s |
| 104.8 TFLOPS | FLOPS (FP32) | 65.3 TFLOPS |
| GDDR7 | Memory Type | GDDR6 |
| 32 GB | Memory Size | 32 GB |
| 1750 MHz | Memory Clock | 2250 MHz |
| 28000 Mbps | Effective Memory Speed | 18000 Mbps |
| 512-bit | Bus | 256-bit |
| No | ECC | No |
| 1792 GB/s | Memory Bandwidth | 576 GB/s |
| PCIe 5.0 x16 | Interface | PCIe 4.0 x16 |
| 575 W | TGP | 250 W |
| TSMC | Manufacturing | TSMC |
| 4 nm | Fabrication Process | 5 nm |
| 750 mm² | Die Size | 609 mm² |
| 92.2 billion | Transistor Count | 76 billion |
| 122.93 MTr/mm² | Transistor Density | 124.79 MTr/mm² |
| 12.2 | DirectX | 12 |
| 1.4 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 12 | CUDA | 8.9 |
| Yes | Ray Tracing | Yes |
| DLSS 4 | DLSS | DLSS 3 |
| 2.1b | DisplayPort | 1.4a |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| January 7, 2025 | Released | August 9, 2023 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Professional |
| High-end | Segment | High-end |
| Blackwell 2.0 | Architecture | Ada Lovelace |
| GB202-300 | GPU Codename | AD102 |
| -Intel Core Ultra 9 285Kor above | Recommended CPU | -Intel Core i9 14900Kor above |
| 52638 | Steel Nomad Lite Score | 29190 |
| 46941 | Time Spy | 18675 |
| 235224 | Solar Bay | 128428 |
| 38113 | Port Royal | 17754 |
| 87433 | Fire Strike | 62027 |
| 108736 | Wild Life Extreme | 52866 |
| 207058 | Night Raid | 210443 |
| 2017 MHz | Base Clock | 1155 MHz |
| 2407 MHz | Boost Clock | 2550 MHz |
| 21760 | Shading Units | 12800 |
| 680 | Texture Mapping Units (TMUs) | 400 |
| 176 | Render Output Units (ROPs) | 176 |
| 170 | Compute Units (Pipelines) | 100 |
| 680 | Tensor Cores | 400 |
| 170 | Ray-tracing Cores | 100 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 96MB shared | L2 Cache | 72MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |
Compare with something else?
Compare GeForce RTX 5090 with another →