Quick/Comparisons
The GeForce RTX 4090 offers significantly better gaming performance, thanks to its higher benchmark scores and excellent gaming performance. Its 92 score far surpasses the GeForce RTX 3090's 61.
The GeForce RTX 4090 takes the lead in content creation, boasting a 91 score compared to the GeForce RTX 3090's 69, thanks to its strong workstation performance and capable AI/ML compute.
The GeForce RTX 4090 outperforms the GeForce RTX 3090 in AI/ML tasks, with a 76 score indicating its capable AI/ML compute capabilities, while the GeForce RTX 3090's 53 score falls short.
61 | Gaming | 92 |
69 | Workstation | 91 |
53 | AI/ML | 76 |
35 | Energy Efficiency | 43 |
62 | Quick Comparison Final Score | 89 |
| GA102 | Graphics Processor | AD102 |
| 10496 | Cores | 16384 |
| 328 / 112 | TMUs / ROPs | 512 / 176 |
| 5800.59 | Blender GPU | 11688.48 |
| 125 | Forza Horizon 5 | 203 |
| 175 | The Witcher 3 | 270 |
| 212 | Counter-Strike 2 | 245 |
| 142 | Far Cry 6 | 183 |
| 106 | Hogwarts Legacy | 165 |
| 159 | Call of Duty | 269 |
| 97 | Ghost of Tsushima | 150 |
| 132 | Cyberpunk 2077 | 204 |
| 226 | Tomb Raider | 283 |
| 153 | Average FPS | 219 |
| 179 | 1080p High | 256 |
| 153 | 1080p Ultra | 219 |
| 119 | 1440p Ultra | 176 |
| 70 | 4K Ultra | 107 |
| Low | Margin of Error | Low |
| 211669 | GB6 Compute Score | 316301 |
| 320.8 img/sec | Background Blur | 318.9 img/sec |
| 207.2 img/sec | Face Detection | 222.8 img/sec |
| 10.1 Gpixels/sec | Horizon Detection | 14.2 Gpixels/sec |
| 14.4 Gpixels/sec | Edge Detection | 21 Gpixels/sec |
| 11.5 Gpixels/sec | Gaussian Blur | 23.9 Gpixels/sec |
| 1.48 Gpixels/sec | Feature Matching | 2.38 Gpixels/sec |
| 984.2 Gpixels/sec | Stereo Matching | 1950 Gpixels/sec |
| 27764.9 FPS | Particle Physics | 48090.4 FPS |
| OpenCL | API | OpenCL |
| 26551 | G3D Mark Score | 38071 |
| 1069 | G2D Mark | 1303 |
| 218 FPS | DirectX 11 | 324 FPS |
| 110 FPS | DirectX 12 | 151 FPS |
| 15117 Ops/s | GPU Compute | 26193 Ops/s |
| 27237 | GB6 ML Single Precision | 42962 |
| 41500 | GB6 ML Half Precision | 59636 |
| 15927 | GB6 ML Quantized | 31670 |
| 10665 | Image Classification (SP) | 15600 |
| 28963 | Image Segmentation (HP) | 36657 |
| 22697 | Image Super Resolution (Q) | 48114 |
| 55232 | Face Detection (HP) | 78111 |
| 128155 | Pose Estimation (Q) | 289639 |
| 3152 | Text Classification (SP) | 4074 |
| 5088 | Machine Translation (HP) | 6349 |
| 13891 | Object Detection (SP) | 21186 |
| 37984 | Depth Estimation (Q) | 75599 |
| 272000 | Style Transfer (SP) | 601428 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 190 GPixel/s | Pixel Fill Rate | 444 GPixel/s |
| 556 GTexel/s | Texture Fill Rate | 1290 GTexel/s |
| 35.6 TFLOPS | FLOPS (FP32) | 82.6 TFLOPS |
| GDDR6X | Memory Type | GDDR6X |
| 24 GB | Memory Size | 24 GB |
| 1219 MHz | Memory Clock | 2625 MHz |
| 19500 Mbps | Effective Memory Speed | 21000 Mbps |
| 384-bit | Bus | 384-bit |
| No | ECC | No |
| 936.2 GB/s | Memory Bandwidth | 1010 GB/s |
| PCIe 4.0 x16 | Interface | PCIe 4.0 x16 |
| 350 W | TGP | 450 W |
| Samsung | Manufacturing | TSMC |
| 8 nm | Fabrication Process | 5 nm |
| 628 mm² | Die Size | 609 mm² |
| 28 billion | Transistor Count | 76.3 billion |
| 44.59 MTr/mm² | Transistor Density | 125.29 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 8.6 | CUDA | 8.9 |
| Yes | Ray Tracing | Yes |
| DLSS 2 | DLSS | DLSS 3 |
| 1.4a | DisplayPort | 1.4a |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| September 24, 2020 | Released | September 20, 2022 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| Ampere | Architecture | Ada Lovelace |
| GA102 | GPU Codename | AD102 |
| -Intel Core i9 12900Kor above | Recommended CPU | -Intel Core i9 14900Kor above |
| 21851 | Steel Nomad Lite Score | 42169 |
| 19896 | Time Spy | 36311 |
| 97375 | Solar Bay | 187386 |
| 13631 | Port Royal | 26128 |
| 47515 | Fire Strike | 72387 |
| 43673 | Wild Life Extreme | 85230 |
| 139755 | Night Raid | 195880 |
| 1395 MHz | Base Clock | 2235 MHz |
| 1695 MHz | Boost Clock | 2520 MHz |
| 10496 | Shading Units | 16384 |
| 328 | Texture Mapping Units (TMUs) | 512 |
| 112 | Render Output Units (ROPs) | 176 |
| 82 | Compute Units (Pipelines) | 128 |
| 328 | Tensor Cores | 512 |
| 82 | Ray-tracing Cores | 128 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 6MB shared | L2 Cache | 72MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |
Compare with something else?
Compare GeForce RTX 3090 with another →