Quick/Comparisons
63 | Gaming | 71 |
63 | Workstation | 73 |
55 | AI/ML |
Compare with something else?
Compare GeForce RTX 4090 with another →60 |
45 | Energy Efficiency | 50 |
59 | Quick Comparison Final Score | 67 |
| AD103 (GN21-X11) | Graphics Processor | GB203-400 |
| 9728 | Cores | 10496 |
| 304 / 112 | TMUs / ROPs | 328 / 112 |
| 7059.06 | Blender GPU | 8001.96 |
| 131 | Forza Horizon 5 | 154 |
| 182 | The Witcher 3 | 198 |
| 208 | Counter-Strike 2 | 213 |
| 147 | Far Cry 6 | 155 |
| 108 | Hogwarts Legacy | 123 |
| 164 | Call of Duty | 187 |
| 100 | Ghost of Tsushima | 115 |
| 137 | Cyberpunk 2077 | 155 |
| 231 | Tomb Raider | 260 |
| 156 | Average FPS | 173 |
| 187 | 1080p High | 202 |
| 156 | 1080p Ultra | 173 |
| 123 | 1440p Ultra | 137 |
| 72 | 4K Ultra | 80 |
| Low | Margin of Error | Low |
| 203263 | GB6 Compute Score | 226752 |
| 264 img/sec | Background Blur | 266.5 img/sec |
| 134.3 img/sec | Face Detection | 196.9 img/sec |
| 9.27 Gpixels/sec | Horizon Detection | 10.6 Gpixels/sec |
| 12.7 Gpixels/sec | Edge Detection | 15.2 Gpixels/sec |
| 13.2 Gpixels/sec | Gaussian Blur | 14.4 Gpixels/sec |
| 1.92 Gpixels/sec | Feature Matching | 1.89 Gpixels/sec |
| 1060 Gpixels/sec | Stereo Matching | 1150 Gpixels/sec |
| 29225.1 FPS | Particle Physics | 29783 FPS |
| OpenCL | API | OpenCL |
| 27060 | G3D Mark Score | 28298 |
| 988 | G2D Mark | 1094 |
| 258 FPS | DirectX 11 | 265 FPS |
| 106 FPS | DirectX 12 | 136 FPS |
| 12206 Ops/s | GPU Compute | 11559 Ops/s |
| 29362 | GB6 ML Single Precision | 30250 |
| 42611 | GB6 ML Half Precision | 52126 |
| 22212 | GB6 ML Quantized | 23496 |
| 11346 | Image Classification (SP) | 12121 |
| 24625 | Image Segmentation (HP) | 33054 |
| 36874 | Image Super Resolution (Q) | 33652 |
| 50088 | Face Detection (HP) | 67213 |
| 174844 | Pose Estimation (Q) | 151395 |
| 3226 | Text Classification (SP) | 3209 |
| 4905 | Machine Translation (HP) | 6292 |
| 14667 | Object Detection (SP) | 15618 |
| 53868 | Depth Estimation (Q) | 57894 |
| 340056 | Style Transfer (SP) | 272117 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 163-228 GPixel/s | Pixel Fill Rate | 170-242 GPixel/s |
| 442-620 GTexel/s | Texture Fill Rate | 497-708 GTexel/s |
| 28.3-39.7 TFLOPS | FLOPS (FP32) | 31.8-45.3 TFLOPS |
| GDDR6 | Memory Type | GDDR7 |
| 16 GB | Memory Size | 24 GB |
| 2000 MHz | Memory Clock | 1750 MHz |
| 16000 Mbps | Effective Memory Speed | 28000 Mbps |
| 256-bit | Bus | 256-bit |
| No | ECC | No |
| 512 GB/s | Memory Bandwidth | 896 GB/s |
| PCIe 4.0 x16 | Interface | PCIe 5.0 x16 |
| 80-175 W (configurable) | TGP | 95-175 W (configurable) |
| TSMC | Manufacturing | TSMC |
| 5 nm | Fabrication Process | 4 nm |
| 379 mm² | Die Size | 378 mm² |
| 45.9 billion | Transistor Count | 45.6 billion |
| 121.11 MTr/mm² | Transistor Density | 120.63 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.4 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 8.9 | CUDA | 10.1 |
| Yes | Ray Tracing | Yes |
| DLSS 3 | DLSS | DLSS 4 |
| 1.4a | DisplayPort | 2.1a |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| February 8, 2023 | Released | January 6, 2025 |
| Laptop | Case | Laptop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| Ada Lovelace | Architecture | Blackwell 2.0 |
| AD103 (GN21-X11) | GPU Codename | GB203-400 |
| -Intel Core Ultra 9 185Hor above | Recommended CPU | -Intel Core Ultra 9 275HXor above |
| 22501 | Steel Nomad Lite Score | 26140 |
| 21401 | Time Spy | 25013 |
| 96130 | Solar Bay | 115677 |
| 13733 | Port Royal | 16533 |
| 45218 | Fire Strike | 51096 |
| 42491 | Wild Life Extreme | 47099 |
| 147892 | Night Raid | 177139 |
| 930-1620 MHz (configurable) | Base Clock | 990-1515 MHz (configurable) |
| 1455-2040 MHz (configurable) | Boost Clock | 1515-2160 MHz (configurable) |
| 9728 | Shading Units | 10496 |
| 304 | Texture Mapping Units (TMUs) | 328 |
| 112 | Render Output Units (ROPs) | 112 |
| 76 | Compute Units (Pipelines) | 82 |
| 304 | Tensor Cores | 328 |
| 76 | Ray-tracing Cores | 82 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 64MB shared | L2 Cache | 64MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |