Key Differences
Advantages of GeForce RTX 3060
Advantages of GeForce RTX 4090
Specifications
Performance
| TGP | 170 W | 450 W |
Storage Details
| Bus | 192-bit | 384-bit |
Ports
| DisplayPort | 1.4a | 1.4a |
Other
| Forza Horizon 5 | 114675740 | 272199190149 |
| The Witcher 3 | 1241007240 | 287263184100 |
| Counter-Strike 2 | 1851329549 | 33524418289 |
| Far Cry 6 | 94846435 | 206185170103 |
| Hogwarts Legacy | 68544220 | 20516913378 |
| CoD: Modern Warfare III | 80744932 | 277269212149 |
| Ghost of Tsushima | 64513519 | 18115613990 |
| Cyberpunk 2077 | 68623819 | 22920514166 |
| Shadow of the Tomb Raider | 1101017037 | 306277215118 |
| 1080p High | 101 | 255 |
| 1080p Ultra | 81 | 219 |
| 1440p Ultra | 58 | 174 |
| 4K Ultra | 32 | 105 |
| Margin of Error | Low | Low |
| Time Spy | 8751 | 36318 |
| Solar Bay | 41040 | 187047 |
| Port Royal | 5144 | 26128 |
| Fire Strike | 22307 | 72296 |
| Wild Life Extreme | 17727 | 85170 |
| Night Raid | 98329 | 195691 |
| Background Blur | 149.6 img/sec | 318.9 img/sec |
| Face Detection | 91 img/sec | 222.8 img/sec |
| Horizon Detection | 4.07 Gpixels/sec | 14.2 Gpixels/sec |
| Edge Detection | 5.58 Gpixels/sec | 21 Gpixels/sec |
| Gaussian Blur | 3.92 Gpixels/sec | 23.9 Gpixels/sec |
| Feature Matching | 1.01 Gpixels/sec | 2.38 Gpixels/sec |
| Stereo Matching | 372.9 Gpixels/sec | 1950 Gpixels/sec |
| Particle Physics | 12226.3 FPS | 48090.4 FPS |
| API | OpenGL | OpenGL |
| G2D Mark | 965 | 1302 |
| DirectX 11 | 128 FPS | 325 FPS |
| DirectX 12 | 69 FPS | 150 FPS |
| GPU Compute | 8307 Ops/s | 26409 Ops/s |
| Aztec Ruins Normal Tier | 555.5 FPS | 1617.6 FPS |
| Car Chase | 492.3 FPS | 1228.6 FPS |
| Manhattan 3.1.1 (1440p) | 407.9 FPS | 1234.6 FPS |
| T-Rex | 1940.6 FPS | 6286 FPS |
| Tessellation | 2965.4 FPS | 12119.1 FPS |
| ALU 2 | 2862.2 FPS | 15073.4 FPS |
| Driver Overhead 2 | 295.2 FPS | 303.7 FPS |
| OS | Windows | Windows |
| Image Classification (SP) | 6645 | 15600 |
| Image Segmentation (HP) | 15338 | 36657 |
| Image Super Resolution (Q) | 14283 | 48114 |
| Face Detection (HP) | 28862 | 78111 |
| Pose Estimation (Q) | 61313 | 289639 |
| Text Classification (SP) | 2463 | 4074 |
| Machine Translation (HP) | 3793 | 6349 |
| Object Detection (SP) | 7712 | 21186 |
| Depth Estimation (Q) | 25461 | 75599 |
| Style Transfer (SP) | 123233 | 601428 |
| Framework | ONNX | ONNX |
| Backend | DirectML | DirectML |
| Vendor | Nvidia | Nvidia |
| Build | Discrete | Discrete |
| Released | January 12, 2021 | September 20, 2022 |
| Launch price (MSRP) | $329 | $1599 |
| Case | Desktop | Desktop |
| Purpose | Gaming | Gaming |
| Segment | Mid-range | High-end |
| Architecture | Ampere | Ada Lovelace |
| GPU Codename | GA106 | AD102 |
| Rival Equivalent | Radeon RX 6700 XT | - |
| Base Clock | 1320 MHz | 2235 MHz |
| Boost Clock | 1777 MHz | 2520 MHz |
| Shading Units | 3584 | 16384 |
| Texture Mapping Units (TMUs) | 112 | 512 |
| Render Output Units (ROPs) | 48 | 176 |
| Compute Units (Pipelines) | 28 | 128 |
| Tensor Cores | 112 | 512 |
| Ray-tracing Cores | 28 | 128 |
| L1 Cache | 128KB per cluster | 128KB per cluster |
| L2 Cache | 3MB shared | 72MB shared |
| Instructions Per Cycle | 2 IPC | 2 IPC |
| Pixel Fill Rate | 85 GPixel/s | 444 GPixel/s |
| Texture Fill Rate | 199 GTexel/s | 1290 GTexel/s |
| Interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
| Manufacturing | Samsung | TSMC |
| Fabrication Process | 8 nm | 5 nm |
| Die Size | 276 mm² | 609 mm² |
| Transistor Count | 12 billion | 76.3 billion |
| Transistor Density | 43.48 MTr/mm² | 125.29 MTr/mm² |
| Max. Temperature | - | - |
| Memory Type | GDDR6 | GDDR6X |
| Memory Size | 12 GB | 24 GB |
| Memory Clock | 1875 MHz | 2625 MHz |
| Effective Memory Speed | 15000 Mbps | 21000 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 | 8.9 |
| Ray Tracing | Yes | Yes |
| DLSS | DLSS 2 | DLSS 3 |