Quick/Comparisons
46 | Gaming | 13 |
49 | Workstation | 22 |
39 | AI/ML |
Compare with something else?
Compare GeForce RTX 3070 with another →12 |
35 | Energy Efficiency | 32 |
46 | Quick Comparison Final Score | 16 |
| GA104 | Graphics Processor | Polaris 20 |
| 6144 | Cores | 2048 |
| 192 / 96 | TMUs / ROPs | 128 / 32 |
| 3746.83 | Blender GPU | 329.21 |
| 104 | Forza Horizon 5 | 36 |
| 143 | The Witcher 3 | 34 |
| 198 | Counter-Strike 2 | 43 |
| 125 | Far Cry 6 | 40 |
| 85 | Hogwarts Legacy | 22 |
| 120 | Call of Duty | 36 |
| 81 | Ghost of Tsushima | 21 |
| 102 | Cyberpunk 2077 | 26 |
| 161 | Tomb Raider | 42 |
| 124 | Average FPS | 33 |
| 150 | 1080p High | 45 |
| 124 | 1080p Ultra | 33 |
| 93 | 1440p Ultra | 23 |
| 54 | 4K Ultra | 12 |
| Low | Margin of Error | Low |
| 136096 | GB6 Compute Score | 37147 |
| 197 img/sec | Background Blur | 55.7 img/sec |
| 128.1 img/sec | Face Detection | 31.3 img/sec |
| 6.56 Gpixels/sec | Horizon Detection | 1.41 Gpixels/sec |
| 8.81 Gpixels/sec | Edge Detection | 2.82 Gpixels/sec |
| 6.47 Gpixels/sec | Gaussian Blur | 2.57 Gpixels/sec |
| 1.26 Gpixels/sec | Feature Matching | 0.31 Gpixels/sec |
| 609.3 Gpixels/sec | Stereo Matching | 143.2 Gpixels/sec |
| 18205 FPS | Particle Physics | 5080.1 FPS |
| OpenCL | API | OpenCL |
| 23260 | G3D Mark Score | 7922 |
| 1061 | G2D Mark | 724 |
| 192 FPS | DirectX 11 | 55 FPS |
| 91 FPS | DirectX 12 | 39 FPS |
| 11452 Ops/s | GPU Compute | 2890 Ops/s |
| 22516 | GB6 ML Single Precision | 6885 |
| 36377 | GB6 ML Half Precision | 7446 |
| 14215 | GB6 ML Quantized | 4931 |
| 8615 | Image Classification (SP) | 2676 |
| 24659 | Image Segmentation (HP) | 3757 |
| 20631 | Image Super Resolution (Q) | 6179 |
| 43834 | Face Detection (HP) | 10752 |
| 98347 | Pose Estimation (Q) | 35488 |
| 3162 | Text Classification (SP) | 1051 |
| 4911 | Machine Translation (HP) | 1933 |
| 11498 | Object Detection (SP) | 3048 |
| 34719 | Depth Estimation (Q) | 12040 |
| 196833 | Style Transfer (SP) | 71201 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 170 GPixel/s | Pixel Fill Rate | 40 GPixel/s |
| 340 GTexel/s | Texture Fill Rate | 159 GTexel/s |
| 21.7 TFLOPS | FLOPS (FP32) | 5.1 TFLOPS |
| GDDR6X | Memory Type | GDDR5 |
| 8 GB | Memory Size | 4 GB |
| 1188 MHz | Memory Clock | 1750 MHz |
| 19000 Mbps | Effective Memory Speed | 7000 Mbps |
| 256-bit | Bus | 256-bit |
| No | ECC | No |
| 608.3 GB/s | Memory Bandwidth | 224 GB/s |
| PCIe 4.0 x16 | Interface | PCIe 3.0 x16 |
| 290 W | TGP | 150 W |
| Samsung | Manufacturing | GlobalFoundries |
| 8 nm | Fabrication Process | 14 nm |
| 392 mm² | Die Size | 232 mm² |
| 17 billion | Transistor Count | 5 billion |
| 43.37 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 |
| June 10, 2021 | Released | April 18, 2017 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| Mid-range | Segment | Mid-range |
| Ampere | Architecture | GCN 4.0 |
| GA104 | GPU Codename | Polaris 20 |
| -Intel Core i7 12700Kor above | Recommended CPU | -Intel Core i5 8500or above |
| 15402 | Steel Nomad Lite Score | 3702 |
| 14880 | Time Spy | 3855 |
| 69900 | Solar Bay | — |
| 8884 | Port Royal | — |
| 37238 | Fire Strike | 13443 |
| 31911 | Wild Life Extreme | 7941 |
| 134012 | Night Raid | 50379 |
| 1575 MHz | Base Clock | 1168 MHz |
| 1770 MHz | Boost Clock | 1244 MHz |
| 6144 | Shading Units | 2048 |
| 192 | Texture Mapping Units (TMUs) | 128 |
| 96 | Render Output Units (ROPs) | 32 |
| 48 | Compute Units (Pipelines) | — |
| 192 | Tensor Cores | No |
| 48 | Ray-tracing Cores | No |
| 128KB per cluster | L1 Cache | 16KB per cluster |
| 4MB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |