Quick/Comparisons
The GeForce RTX 4090's superior gaming performance and higher frame rates make it the clear winner in this use case. Its 92 score is significantly higher than the GeForce RTX 2070 SUPER's 32 score.
The GeForce RTX 4090 and GeForce RTX 2070 SUPER are nearly tied in content creation, with the GeForce RTX 4090's strong workstation performance being matched by the GeForce RTX 2070 SUPER's 39 score.
The GeForce RTX 2070 SUPER's dedicated AI tensor cores and Turing architecture give it a slight edge in AI/ML compute, making it the winner in this use case despite the GeForce RTX 4090's 76 score.
32 | Gaming | 92 |
39 | Workstation | 91 |
33 | AI/ML | 76 |
27 | Energy Efficiency | 43 |
34 | Quick Comparison Final Score | 89 |
| TU104 | Graphics Processor | AD102 |
| 2560 | Cores | 16384 |
| 160 / 64 | TMUs / ROPs | 512 / 176 |
| 2361.66 | Blender GPU | 11688.48 |
| 77 | Forza Horizon 5 | 203 |
| 109 | The Witcher 3 | 270 |
| 150 | Counter-Strike 2 | 245 |
| 91 | Far Cry 6 | 183 |
| 61 | Hogwarts Legacy | 165 |
| 82 | Call of Duty | 269 |
| 58 | Ghost of Tsushima | 150 |
| 71 | Cyberpunk 2077 | 204 |
| 113 | Tomb Raider | 283 |
| 90 | Average FPS | 219 |
| 111 | 1080p High | 256 |
| 90 | 1080p Ultra | 219 |
| 66 | 1440p Ultra | 176 |
| 38 | 4K Ultra | 107 |
| Low | Margin of Error | Low |
| 105349 | GB6 Compute Score | 316301 |
| 166.7 img/sec | Background Blur | 318.9 img/sec |
| 103.3 img/sec | Face Detection | 222.8 img/sec |
| 4.21 Gpixels/sec | Horizon Detection | 14.2 Gpixels/sec |
| 5.56 Gpixels/sec | Edge Detection | 21 Gpixels/sec |
| 5.58 Gpixels/sec | Gaussian Blur | 23.9 Gpixels/sec |
| 1.17 Gpixels/sec | Feature Matching | 2.38 Gpixels/sec |
| 489.7 Gpixels/sec | Stereo Matching | 1950 Gpixels/sec |
| 13242.9 FPS | Particle Physics | 48090.4 FPS |
| OpenCL | API | OpenCL |
| 18141 | G3D Mark Score | 38071 |
| 878 | G2D Mark | 1303 |
| 153 FPS | DirectX 11 | 324 FPS |
| 67 FPS | DirectX 12 | 151 FPS |
| 7473 Ops/s | GPU Compute | 26193 Ops/s |
| 16276 | GB6 ML Single Precision | 42962 |
| 27406 | GB6 ML Half Precision | 59636 |
| 10983 | GB6 ML Quantized | 31670 |
| 6648 | Image Classification (SP) | 15600 |
| 16442 | Image Segmentation (HP) | 36657 |
| 15383 | Image Super Resolution (Q) | 48114 |
| 32861 | Face Detection (HP) | 78111 |
| 66348 | Pose Estimation (Q) | 289639 |
| 2166 | Text Classification (SP) | 4074 |
| 4239 | Machine Translation (HP) | 6349 |
| 8194 | Object Detection (SP) | 21186 |
| 26192 | Depth Estimation (Q) | 75599 |
| 137642 | Style Transfer (SP) | 601428 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 113 GPixel/s | Pixel Fill Rate | 444 GPixel/s |
| 283 GTexel/s | Texture Fill Rate | 1290 GTexel/s |
| 9.1 TFLOPS | FLOPS (FP32) | 82.6 TFLOPS |
| GDDR6 | Memory Type | GDDR6X |
| 8 GB | Memory Size | 24 GB |
| 1750 MHz | Memory Clock | 2625 MHz |
| 14000 Mbps | Effective Memory Speed | 21000 Mbps |
| 256-bit | Bus | 384-bit |
| No | ECC | No |
| 448 GB/s | Memory Bandwidth | 1010 GB/s |
| PCIe 3.0 x16 | Interface | PCIe 4.0 x16 |
| 215 W | TGP | 450 W |
| TSMC | Manufacturing | TSMC |
| 12 nm | Fabrication Process | 5 nm |
| 545 mm² | Die Size | 609 mm² |
| 13 billion | Transistor Count | 76.3 billion |
| 23.85 MTr/mm² | Transistor Density | 125.29 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 7.5 | CUDA | 8.9 |
| Yes | Ray Tracing | Yes |
| DLSS 2 | DLSS | DLSS 3 |
| 1.4a | DisplayPort | 1.4a |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| July 9, 2019 | Released | September 20, 2022 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| Mid-range | Segment | High-end |
| Turing | Architecture | Ada Lovelace |
| TU104 | GPU Codename | AD102 |
| -Intel Core i7 9700Kor above | Recommended CPU | -Intel Core i9 14900Kor above |
| 10053 | Steel Nomad Lite Score | 42169 |
| 10167 | Time Spy | 36311 |
| 45871 | Solar Bay | 187386 |
| 6032 | Port Royal | 26128 |
| 25675 | Fire Strike | 72387 |
| 20745 | Wild Life Extreme | 85230 |
| 102728 | Night Raid | 195880 |
| 1605 MHz | Base Clock | 2235 MHz |
| 1770 MHz | Boost Clock | 2520 MHz |
| 2560 | Shading Units | 16384 |
| 160 | Texture Mapping Units (TMUs) | 512 |
| 64 | Render Output Units (ROPs) | 176 |
| 40 | Compute Units (Pipelines) | 128 |
| 320 | Tensor Cores | 512 |
| 40 | Ray-tracing Cores | 128 |
| 64KB per cluster | L1 Cache | 128KB per cluster |
| 4MB shared | L2 Cache | 72MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |
Compare with something else?
Compare GeForce RTX 2070 with another →