GeForce RTX 2050 vs GeForce GTX 660
Summary
Reasons to consider GeForce RTX 2050 |
| Higher theoretical gaming performance, based on specifications. |
| Supports PhysX |
| Supports G-Sync |
| Supports ShadowPlay (allows game streaming/recording with minimum performance penalty) |
| Supports Direct3D 12 Async Compute |
| Supports DirectX Raytracing (DXR) |
| Supports Deep Learning Super-Sampling (DLSS) |
Reasons to consider GeForce GTX 660 |
| Supports PhysX |
| Supports G-Sync |
| Supports ShadowPlay (allows game streaming/recording with minimum performance penalty) |
| Based on an outdated architecture (Nvidia Kepler), there are less performance optimizations for current games and applications |
HWBench recommends GeForce RTX 2050
Based on theoretical specifications.
Core Configuration
| GeForce RTX 2050 | | GeForce GTX 660 | |
|---|
| GPU Name | TU107 (TU107) | vs | GK106 (GK106-400-A1) |
| Fab Process | 12 nm | vs | 28 nm |
| Die Size | 0 mm² | vs | 221 mm² |
| Transistors | unknown | vs | 2,540 million |
| Shaders | 896 | vs | 960 |
| Compute Units | 14 | vs | 5 |
| Core clock | 1515 MHz | vs | 980 MHz |
| ROPs | 32 | vs | 24 |
| TMUs | 56 | vs | 80 |
Memory Configuration
| GeForce RTX 2050 | | GeForce GTX 660 | |
|---|
| Memory Type | GDDR6 | vs | GDDR5 |
| Bus Width | 128 bit | vs | 192 bit |
| Memory Speed | 1750 MHz
14000 MHz effective | vs | 1502 MHz
6008 MHz effective |
| Memory Size | 4096 Mb | vs | 2048 Mb |
Additional details
| GeForce RTX 2050 | | GeForce GTX 660 | |
|---|
| TDP | 120 watts | vs | 140 watts |
| Release Date | UNRELEASED | vs | 6 Sep 2012 |
GeForce RTX 2050
54.24 GP/s
GeForce GTX 660
19.60 GP/s
GigaPixels - higher is better
GeForce RTX 2050
94.92 GT/s
GeForce GTX 660
78.40 GT/s
GigaTexels - higher is better
GeForce RTX 2050
224.00 GB/s
GeForce GTX 660
144.00 GB/s
GB/s - higher is better
GeForce RTX 2050
3037.00 GFLOPs
GeForce GTX 660
1882.00 GFLOPs
GFLOPs - higher is better