Quick/Comparisons
37 | Gaming | 87 |
41 | Workstation | 85 |
31 | AI/ML |
Compare with something else?
Compare GeForce RTX 3070 with another →70 |
47 | Energy Efficiency | 49 |
39 | Quick Comparison Final Score | 84 |
| GA104 (GN20-E6) | Graphics Processor | GB203-400 |
| 5632 | Cores | 10752 |
| 176 / 80 | TMUs / ROPs | 336 / 112 |
| 3120.06 | Blender GPU | 9138.54 |
| 85 | Forza Horizon 5 | 185 |
| 123 | The Witcher 3 | 237 |
| 167 | Counter-Strike 2 | 237 |
| 101 | Far Cry 6 | 176 |
| 67 | Hogwarts Legacy | 154 |
| 96 | Call of Duty | 241 |
| 63 | Ghost of Tsushima | 137 |
| 81 | Cyberpunk 2077 | 190 |
| 126 | Tomb Raider | 270 |
| 101 | Average FPS | 203 |
| 122 | 1080p High | 239 |
| 101 | 1080p Ultra | 203 |
| 74 | 1440p Ultra | 164 |
| 42 | 4K Ultra | 97 |
| Low | Margin of Error | Low |
| 116337 | GB6 Compute Score | 297747 |
| 173.8 img/sec | Background Blur | 319.6 img/sec |
| 96.9 img/sec | Face Detection | 225 img/sec |
| 5.69 Gpixels/sec | Horizon Detection | 13.6 Gpixels/sec |
| 7.75 Gpixels/sec | Edge Detection | 19.5 Gpixels/sec |
| 6.27 Gpixels/sec | Gaussian Blur | 20.4 Gpixels/sec |
| 1.14 Gpixels/sec | Feature Matching | 2.33 Gpixels/sec |
| 480.4 Gpixels/sec | Stereo Matching | 1750 Gpixels/sec |
| 14706.2 FPS | Particle Physics | 43676.6 FPS |
| OpenCL | API | OpenCL |
| 17472 | G3D Mark Score | 35695 |
| 773 | G2D Mark | 1407 |
| 152 FPS | DirectX 11 | 327 FPS |
| 72 FPS | DirectX 12 | 144 FPS |
| 7362 Ops/s | GPU Compute | 19397 Ops/s |
| 17250 | GB6 ML Single Precision | 38994 |
| 27936 | GB6 ML Half Precision | 66609 |
| 11249 | GB6 ML Quantized | 29206 |
| 6969 | Image Classification (SP) | 15441 |
| 17142 | Image Segmentation (HP) | 47173 |
| 16426 | Image Super Resolution (Q) | 43223 |
| 31628 | Face Detection (HP) | 92863 |
| 74413 | Pose Estimation (Q) | 202202 |
| 2522 | Text Classification (SP) | 3701 |
| 3639 | Machine Translation (HP) | 7905 |
| 9002 | Object Detection (SP) | 20955 |
| 27790 | Depth Estimation (Q) | 69068 |
| 160005 | Style Transfer (SP) | 349205 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 83-119 GPixel/s | Pixel Fill Rate | 293 GPixel/s |
| 182-261 GTexel/s | Texture Fill Rate | 879 GTexel/s |
| 11.7-16.7 TFLOPS | FLOPS (FP32) | 56.3 TFLOPS |
| GDDR6 | Memory Type | GDDR7 |
| 8 GB | Memory Size | 16 GB |
| 1750 MHz | Memory Clock | 1875 MHz |
| 14000 Mbps | Effective Memory Speed | 30000 Mbps |
| 256-bit | Bus | 256-bit |
| No | ECC | No |
| 448 GB/s | Memory Bandwidth | 960 GB/s |
| PCIe 4.0 x16 | Interface | PCIe 5.0 x16 |
| 80-150 W (configurable) | TGP | 360 W |
| Samsung | Manufacturing | TSMC |
| 8 nm | Fabrication Process | 4 nm |
| 392 mm² | Die Size | 378 mm² |
| 17.4 billion | Transistor Count | 45.6 billion |
| 44.39 MTr/mm² | Transistor Density | 120.63 MTr/mm² |
| 12 | DirectX | 12.2 |
| 1.2 | Vulkan | 1.4 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 8.6 | CUDA | 12 |
| Yes | Ray Tracing | Yes |
| DLSS 2 | DLSS | DLSS 4 |
| 1.4a | DisplayPort | 2.1b |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| — | Released | January 7, 2025 |
| Laptop | Case | Desktop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| Ampere | Architecture | Blackwell 2.0 |
| GA104 (GN20-E6) | GPU Codename | GB203-400 |
| -Intel Core i7 12700Hor above | Recommended CPU | -Intel Core Ultra 7 265Kor above |
| 11488 | Steel Nomad Lite Score | 36200 |
| 11331 | Time Spy | 32989 |
| 49158 | Solar Bay | 160901 |
| 6850 | Port Royal | 23257 |
| 28619 | Fire Strike | 75518 |
| 23071 | Wild Life Extreme | 69544 |
| 95266 | Night Raid | 200077 |
| 510-1035 MHz (configurable) | Base Clock | 2295 MHz |
| 1035-1485 MHz (configurable) | Boost Clock | 2617 MHz |
| 5632 | Shading Units | 10752 |
| 176 | Texture Mapping Units (TMUs) | 336 |
| 80 | Render Output Units (ROPs) | 112 |
| 44 | Compute Units (Pipelines) | 84 |
| 176 | Tensor Cores | 336 |
| 44 | Ray-tracing Cores | 84 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 4MB shared | L2 Cache | 64MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |