Key Differences
Advantages of GeForce RTX 3090
Advantages of GeForce RTX 3070
Specifications
Performance
| TGP | 350 W | 220 W |
Storage Details
| Bus | 384-bit | 256-bit |
Ports
| DisplayPort | 1.4a | 1.4a |
Other
| Forza Horizon 5 | 18812711891 | 145958562 |
| The Witcher 3 | 20917512669 | 17014110356 |
| Counter-Strike 2 | 27421515685 | 25119213676 |
| Far Cry 6 | 15314512278 | 1271189557 |
| Hogwarts Legacy | 1271078747 | 97806233 |
| CoD: Modern Warfare III | 17116412482 | 1251138452 |
| Ghost of Tsushima | 120978248 | 87755630 |
| Cyberpunk 2077 | 1511328939 | 108925830 |
| Shadow of the Tomb Raider | 24221817290 | 16715211462 |
| 1080p High | 182 | 142 |
| 1080p Ultra | 153 | 118 |
| 1440p Ultra | 120 | 88 |
| 4K Ultra | 70 | 51 |
| Margin of Error | Low | Low |
| Time Spy | 19899 | 13642 |
| Solar Bay | 97411 | 64327 |
| Port Royal | 13633 | 8311 |
| Fire Strike | 47513 | 34046 |
| Wild Life Extreme | 43649 | 28473 |
| Night Raid | 139651 | 121837 |
| Background Blur | 320.8 img/sec | 161.6 img/sec |
| Face Detection | 207.2 img/sec | 117.2 img/sec |
| Horizon Detection | 10.1 Gpixels/sec | 5.89 Gpixels/sec |
| Edge Detection | 14.4 Gpixels/sec | 8.05 Gpixels/sec |
| Gaussian Blur | 11.5 Gpixels/sec | 6.65 Gpixels/sec |
| Feature Matching | 1.48 Gpixels/sec | 1.25 Gpixels/sec |
| Stereo Matching | 984.2 Gpixels/sec | 550.6 Gpixels/sec |
| Particle Physics | 27764.9 FPS | 17613.1 FPS |
| API | OpenGL | OpenGL |
| G2D Mark | 1066 | 1003 |
| DirectX 11 | 219 FPS | 182 FPS |
| DirectX 12 | 110 FPS | 85 FPS |
| GPU Compute | 15243 Ops/s | 11130 Ops/s |
| Aztec Ruins Normal Tier | 978.4 FPS | 856 FPS |
| Car Chase | 854.1 FPS | 668 FPS |
| Manhattan 3.1.1 (1440p) | 747.4 FPS | 592.9 FPS |
| T-Rex | 3732.1 FPS | 3078.4 FPS |
| Tessellation | 6622.8 FPS | 4563.3 FPS |
| ALU 2 | 6650.9 FPS | 4531.5 FPS |
| Driver Overhead 2 | 296 FPS | 283 FPS |
| OS | Windows | Windows |
| Image Classification (SP) | 10665 | 7624 |
| Image Segmentation (HP) | 28963 | 23420 |
| Image Super Resolution (Q) | 22697 | 17898 |
| Face Detection (HP) | 55232 | 40535 |
| Pose Estimation (Q) | 128155 | 83624 |
| Text Classification (SP) | 3152 | 2724 |
| Machine Translation (HP) | 5088 | 4481 |
| Object Detection (SP) | 13891 | 10304 |
| Depth Estimation (Q) | 37984 | 30418 |
| Style Transfer (SP) | 272000 | 180089 |
| Framework | ONNX | ONNX |
| Backend | DirectML | DirectML |
| Vendor | Nvidia | Nvidia |
| Build | Discrete | Discrete |
| Released | September 24, 2020 | October 22, 2020 |
| Launch price (MSRP) | $1499 | $499 |
| Case | Desktop | Desktop |
| Purpose | Gaming | Gaming |
| Segment | High-end | Mid-range |
| Architecture | Ampere | Ampere |
| GPU Codename | GA102 | GA104 |
| Rival Equivalent | Radeon RX 6900 XT | Radeon RX 7800 XT |
| Base Clock | 1395 MHz | 1500 MHz |
| Boost Clock | 1695 MHz | 1725 MHz |
| Shading Units | 10496 | 5888 |
| Texture Mapping Units (TMUs) | 328 | 184 |
| Render Output Units (ROPs) | 112 | 96 |
| Compute Units (Pipelines) | 82 | 46 |
| Tensor Cores | 328 | 184 |
| Ray-tracing Cores | 82 | 46 |
| L1 Cache | 128KB per cluster | 128KB per cluster |
| L2 Cache | 6MB shared | 4MB shared |
| Instructions Per Cycle | 2 IPC | 2 IPC |
| Pixel Fill Rate | 190 GPixel/s | 166 GPixel/s |
| Texture Fill Rate | 556 GTexel/s | 317 GTexel/s |
| Interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
| Manufacturing | Samsung | Samsung |
| Fabrication Process | 8 nm | 8 nm |
| Die Size | 628 mm² | 392 mm² |
| Transistor Count | 28 billion | 17 billion |
| Transistor Density | 44.59 MTr/mm² | 43.37 MTr/mm² |
| Memory Type | GDDR6X | GDDR6 |
| Memory Size | 24 GB | 8 GB |
| Memory Clock | 1219 MHz | 1750 MHz |
| Effective Memory Speed | 19500 Mbps | 14000 Mbps |
| ECC | No | No |
| DirectX | 12 | 12 |
| Vulkan | 1.3 | 1.3 |
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| CUDA | 8.6 | 8.6 |
| Ray Tracing | Yes | Yes |
| DLSS | DLSS 2 | DLSS 2 |
| Max. Temperature | - | - |