Quick/Comparisons
The GeForce GTX 1660 SUPER takes the lead in gaming performance, with a score of 19, thanks to its Turing architecture and 6GB GDDR6 memory.
The GeForce GTX 1660 SUPER outperforms the GeForce RTX 2050 Mobile in content creation, with a score of 29, due to its higher workstation score and more memory.
Both GPUs have dedicated AI tensor cores, but the GeForce GTX 1660 SUPER has a slightly higher AI/ML score of 17, making it a tie in this use case.
19 | Gaming | 14 |
29 | Workstation | 20 |
17 | AI/ML |
Compare with something else?
Compare GeForce GTX 1660 with another →14 |
43 | Energy Efficiency | 75 |
22 | Quick Comparison Final Score | 26 |
| TU116 | Graphics Processor | GA107 |
| 1408 | Cores | 2048 |
| 88 / 48 | TMUs / ROPs | 64 / 32 |
| 772.88 | Blender GPU | 749.91 |
| 49 | Forza Horizon 5 | 35 |
| 67 | The Witcher 3 | 41 |
| 91 | Counter-Strike 2 | 60 |
| 64 | Far Cry 6 | 44 |
| 37 | Hogwarts Legacy | 23 |
| 44 | Call of Duty | 32 |
| 33 | Ghost of Tsushima | 21 |
| 40 | Cyberpunk 2077 | 26 |
| 68 | Tomb Raider | 50 |
| 55 | Average FPS | 37 |
| 70 | 1080p High | 47 |
| 55 | 1080p Ultra | 37 |
| 39 | 1440p Ultra | 25 |
| 20 | 4K Ultra | 13 |
| Low | Margin of Error | Low |
| 67074 | GB6 Compute Score | 42967 |
| 122.5 img/sec | Background Blur | 73.3 img/sec |
| 68.6 img/sec | Face Detection | 39.7 img/sec |
| 2.56 Gpixels/sec | Horizon Detection | 1.65 Gpixels/sec |
| 3.33 Gpixels/sec | Edge Detection | 2.13 Gpixels/sec |
| 3.21 Gpixels/sec | Gaussian Blur | 2.06 Gpixels/sec |
| 0.86 Gpixels/sec | Feature Matching | 0.55 Gpixels/sec |
| 273.3 Gpixels/sec | Stereo Matching | 191.1 Gpixels/sec |
| 8476.3 FPS | Particle Physics | 5921.5 FPS |
| OpenCL | API | OpenCL |
| 12684 | G3D Mark Score | 7710 |
| 807 | G2D Mark | 480 |
| 104 FPS | DirectX 11 | 47 FPS |
| 50 FPS | DirectX 12 | 42 FPS |
| 5050 Ops/s | GPU Compute | 3502 Ops/s |
| 10238 | GB6 ML Single Precision | 7807 |
| 8153 | GB6 ML Half Precision | 14632 |
| 7392 | GB6 ML Quantized | 5491 |
| 4278 | Image Classification (SP) | 3489 |
| 5856 | Image Segmentation (HP) | 7528 |
| 9561 | Image Super Resolution (Q) | 6509 |
| 9830 | Face Detection (HP) | 13217 |
| 36109 | Pose Estimation (Q) | 24531 |
| 1858 | Text Classification (SP) | 2148 |
| 3286 | Machine Translation (HP) | 2764 |
| 5151 | Object Detection (SP) | 3948 |
| 18021 | Depth Estimation (Q) | 13355 |
| 73708 | Style Transfer (SP) | 60626 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 86 GPixel/s | Pixel Fill Rate | 40-47 GPixel/s |
| 157 GTexel/s | Texture Fill Rate | 80-95 GTexel/s |
| 5 TFLOPS | FLOPS (FP32) | 5.1-6 TFLOPS |
| GDDR6 | Memory Type | GDDR6 |
| 6 GB | Memory Size | 4 GB |
| 1750 MHz | Memory Clock | 1750 MHz |
| 14000 Mbps | Effective Memory Speed | 14000 Mbps |
| 192-bit | Bus | 64-bit |
| No | ECC | No |
| 336 GB/s | Memory Bandwidth | 112 GB/s |
| PCIe 3.0 x16 | Interface | PCIe 3.0 x8 |
| 125 W | TGP | 30-45 W (configurable) |
| TSMC | Manufacturing | Samsung |
| 12 nm | Fabrication Process | 8 nm |
| 284 mm² | Die Size | 200 mm² |
| 6 billion | Transistor Count | 8 billion |
| 21.13 MTr/mm² | Transistor Density | 40 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 7.5 | CUDA | 8.6 |
| No | Ray Tracing | Yes |
| No | DLSS | No |
| 1.4a | DisplayPort | 1.4a |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| October 29, 2019 | Released | December 17, 2021 |
| Desktop | Case | Laptop |
| Gaming | Purpose | Gaming |
| Entry-level | Segment | Mid-range |
| Turing | Architecture | Ampere |
| TU116 | GPU Codename | GA107 |
| -Intel Core i5 9500or above | Recommended CPU | -Intel Core i5 11500Hor above |
| 5672 | Steel Nomad Lite Score | 3959 |
| 6044 | Time Spy | 3924 |
| 19269 | Solar Bay | 18938 |
| 1548 | Port Royal | 446 |
| 16189 | Fire Strike | 10490 |
| 12328 | Wild Life Extreme | 7298 |
| 74668 | Night Raid | 45541 |
| 1530 MHz | Base Clock | 735-1185 MHz (configurable) |
| 1785 MHz | Boost Clock | 1245-1477 MHz (configurable) |
| 1408 | Shading Units | 2048 |
| 88 | Texture Mapping Units (TMUs) | 64 |
| 48 | Render Output Units (ROPs) | 32 |
| 22 | Compute Units (Pipelines) | 16 |
| No | Tensor Cores | 64 |
| No | Ray-tracing Cores | 16 |
| 64KB per cluster | L1 Cache | 64KB per cluster |
| 1536KB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |