Quick/Comparisons
61 | Gaming | 15 |
69 | Workstation | 25 |
53 | AI/ML |
Compare with something else?
Compare GeForce RTX 3090 with another →17 |
35 | Energy Efficiency | 28 |
62 | Quick Comparison Final Score | 19 |
| GA102 | Graphics Processor | Polaris 30 |
| 10496 | Cores | 2304 |
| 328 / 112 | TMUs / ROPs | 144 / 32 |
| 5800.59 | Blender GPU | 487.92 |
| 125 | Forza Horizon 5 | 42 |
| 175 | The Witcher 3 | 47 |
| 212 | Counter-Strike 2 | 64 |
| 142 | Far Cry 6 | 50 |
| 106 | Hogwarts Legacy | 29 |
| 159 | Call of Duty | 42 |
| 97 | Ghost of Tsushima | 25 |
| 132 | Cyberpunk 2077 | 30 |
| 226 | Tomb Raider | 55 |
| 153 | Average FPS | 43 |
| 179 | 1080p High | 57 |
| 153 | 1080p Ultra | 43 |
| 119 | 1440p Ultra | 28 |
| 70 | 4K Ultra | 15 |
| Low | Margin of Error | Low |
| 211669 | GB6 Compute Score | 45783 |
| 320.8 img/sec | Background Blur | 69 img/sec |
| 207.2 img/sec | Face Detection | 38.2 img/sec |
| 10.1 Gpixels/sec | Horizon Detection | 1.75 Gpixels/sec |
| 14.4 Gpixels/sec | Edge Detection | 3.31 Gpixels/sec |
| 11.5 Gpixels/sec | Gaussian Blur | 3.31 Gpixels/sec |
| 1.48 Gpixels/sec | Feature Matching | 0.38 Gpixels/sec |
| 984.2 Gpixels/sec | Stereo Matching | 180.5 Gpixels/sec |
| 27764.9 FPS | Particle Physics | 6238 FPS |
| OpenCL | API | OpenCL |
| 26551 | G3D Mark Score | 9293 |
| 1069 | G2D Mark | 790 |
| 218 FPS | DirectX 11 | 65 FPS |
| 110 FPS | DirectX 12 | 45 FPS |
| 15117 Ops/s | GPU Compute | 3764 Ops/s |
| 27237 | GB6 ML Single Precision | 9064 |
| 41500 | GB6 ML Half Precision | 9971 |
| 15927 | GB6 ML Quantized | 6293 |
| 10665 | Image Classification (SP) | 3307 |
| 28963 | Image Segmentation (HP) | 4895 |
| 22697 | Image Super Resolution (Q) | 9072 |
| 55232 | Face Detection (HP) | 12467 |
| 128155 | Pose Estimation (Q) | 43468 |
| 3152 | Text Classification (SP) | 1547 |
| 5088 | Machine Translation (HP) | 2504 |
| 13891 | Object Detection (SP) | 3746 |
| 37984 | Depth Estimation (Q) | 15075 |
| 272000 | Style Transfer (SP) | 96813 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 190 GPixel/s | Pixel Fill Rate | 49 GPixel/s |
| 556 GTexel/s | Texture Fill Rate | 222 GTexel/s |
| 35.6 TFLOPS | FLOPS (FP32) | 7.1 TFLOPS |
| GDDR6X | Memory Type | GDDR5 |
| 24 GB | Memory Size | 8 GB |
| 1219 MHz | Memory Clock | 2000 MHz |
| 19500 Mbps | Effective Memory Speed | 8000 Mbps |
| 384-bit | Bus | 256-bit |
| No | ECC | No |
| 936.2 GB/s | Memory Bandwidth | 256 GB/s |
| PCIe 4.0 x16 | Interface | PCIe 3.0 x16 |
| 350 W | TGP | 175 W |
| Samsung | Manufacturing | GlobalFoundries |
| 8 nm | Fabrication Process | 12 nm |
| 628 mm² | Die Size | 232 mm² |
| 28 billion | Transistor Count | 5 billion |
| 44.59 MTr/mm² | Transistor Density | 21.55 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.1 |
| 8.6 | CUDA | — |
| Yes | Ray Tracing | No |
| DLSS 2 | DLSS | No |
| 1.4a | DisplayPort | 1.4a |
| Nvidia | Vendor | Amd |
| Discrete | Build | Discrete |
| September 24, 2020 | Released | November 15, 2018 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| Ampere | Architecture | GCN 4.0 |
| GA102 | GPU Codename | Polaris 30 |
| -Intel Core i9 12900Kor above | Recommended CPU | -Intel Core i5 8500or above |
| 21851 | Steel Nomad Lite Score | 4611 |
| 19896 | Time Spy | 4778 |
| 97375 | Solar Bay | — |
| 13631 | Port Royal | — |
| 47515 | Fire Strike | 16527 |
| 43673 | Wild Life Extreme | 9735 |
| 139755 | Night Raid | 62048 |
| 1395 MHz | Base Clock | 1469 MHz |
| 1695 MHz | Boost Clock | 1545 MHz |
| 10496 | Shading Units | 2304 |
| 328 | Texture Mapping Units (TMUs) | 144 |
| 112 | Render Output Units (ROPs) | 32 |
| 82 | Compute Units (Pipelines) | — |
| 328 | Tensor Cores | No |
| 82 | Ray-tracing Cores | No |
| 128KB per cluster | L1 Cache | 16KB per cluster |
| 6MB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |