Quick/Comparisons
A budget Nvidia GeForce GTX 1050 Ti graphics card with limited performance overall. Suited for casual gaming and basic graphics tasks.
| Gaming | 8 /100 |
| Workstation | 17 /100 |
| AI/ML | 8 /100 |
| Energy Efficiency | 54 |
Looking for more comparisons?
Compare GeForce GTX →| Quick Comparison Final Score | 12 /100 |
| Graphics Processor | GP107 |
| Cores | 768 |
| TMUs / ROPs | 48 / 32 |
| Blender GPU | 214.48 |
| Forza Horizon 5 | 15 |
| The Witcher 3 | 20 |
| Counter-Strike 2 | 29 |
| Far Cry 6 | 28 |
| Hogwarts Legacy | 13 |
| Call of Duty | 26 |
| Ghost of Tsushima | 9 |
| Cyberpunk 2077 | 12 |
| Tomb Raider | 27 |
| Average FPS | 20 |
| 1080p High | 26 |
| 1080p Ultra | 20 |
| 1440p Ultra | 14 |
| 4K Ultra | - |
| Margin of Error | Low |
| GB6 Compute Score | 21180 |
| Background Blur | 56.6 img/sec |
| Face Detection | 27.4 img/sec |
| Horizon Detection | 0.82 Gpixels/sec |
| Edge Detection | 1.2 Gpixels/sec |
| Gaussian Blur | 1.27 Gpixels/sec |
| Feature Matching | 0.18 Gpixels/sec |
| Stereo Matching | 66.3 Gpixels/sec |
| Particle Physics | 1916.5 FPS |
| API | OpenCL |
| G3D Mark Score | 6362 |
| G2D Mark | 655 |
| DirectX 11 | 44 FPS |
| DirectX 12 | 26 FPS |
| GPU Compute | 2665 Ops/s |
| GB6 ML Single Precision | 4117 |
| GB6 ML Half Precision | 3560 |
| GB6 ML Quantized | 3084 |
| Image Classification (SP) | 1518 |
| Image Segmentation (HP) | 2085 |
| Image Super Resolution (Q) | 4847 |
| Face Detection (HP) | 3275 |
| Pose Estimation (Q) | 16118 |
| Text Classification (SP) | 1440 |
| Machine Translation (HP) | 1875 |
| Object Detection (SP) | 1743 |
| Depth Estimation (Q) | 6258 |
| Style Transfer (SP) | 26368 |
| Framework | ONNX |
| Backend | DirectML |
| Pixel Fill Rate | 45 GPixel/s |
| Texture Fill Rate | 67 GTexel/s |
| FLOPS (FP32) | 2.1 TFLOPS |
| Memory Type | GDDR5 |
| Memory Size | 4 GB |
| Memory Clock | 1752 MHz |
| Effective Memory Speed | 7000 Mbps |
| Bus | 128-bit |
| ECC | No |
| Memory Bandwidth | 112.1 GB/s |
| Interface | PCIe 3.0 x16 |
| TGP | 75 W |
| Manufacturing | Samsung |
| Fabrication Process | 14 nm |
| Die Size | 132 mm² |
| Transistor Count | 3 billion |
| Transistor Density | 22.73 MTr/mm² |
| DirectX | 12 |
| Vulkan | 1.3 |
| OpenGL | 4.6 |
| OpenCL | 3 |
| CUDA | 6.1 |
| Ray Tracing | No |
| DLSS | No |
| DisplayPort | 1.4a |
| Vendor | Nvidia |
| Build | Discrete |
| Released | October 25, 2016 |
| Case | Desktop |
| Purpose | Gaming |
| Segment | Entry-level |
| Architecture | Pascal |
| GPU Codename | GP107 |
| Recommended CPU | — |
| Steel Nomad Lite Score | 2312 |
| Time Spy | 2352 |
| Solar Bay | — |
| Port Royal | — |
| Fire Strike | 7702 |
| Wild Life Extreme | 4908 |
| Night Raid | 33485 |
| Base Clock | 1291 MHz |
| Boost Clock | 1392 MHz |
| Shading Units | 768 |
| Texture Mapping Units (TMUs) | 48 |
| Render Output Units (ROPs) | 32 |
| Compute Units (Pipelines) | 6 |
| Tensor Cores | No |
| Ray-tracing Cores | No |
| L1 Cache | 48KB per cluster |
| L2 Cache | 1024KB shared |
| Instructions Per Cycle | 2 IPC |
84/100