Quick/Comparisons
Both GPUs have identical gaming scores, indicating a tie in this use case.
The GeForce MX450 leads in workstation performance, thanks to its higher score of 14.
The GeForce MX450 has a clear advantage in AI/ML performance, with a score of 7 compared to the Radeon RX Vega 8's score of 4.
6 | Gaming | 6 |
14 | Workstation | 12 |
7 | AI/ML |
Compare with something else?
Compare GeForce MX450 with another →4 |
71 | Energy Efficiency | 84 |
19 | Quick Comparison Final Score | 20 |
| TU117 | Graphics Processor | Raven |
| 896 | Cores | 512 |
| 56 / 32 | TMUs / ROPs | 32 / 8 |
| 201.01 | Blender GPU | — |
| 10 | Forza Horizon 5 | 13 |
| 12 | The Witcher 3 | 10 |
| 22 | Counter-Strike 2 | 15 |
| 21 | Far Cry 6 | 16 |
| 10 | Hogwarts Legacy | - |
| 23 | Call of Duty | 16 |
| - | Ghost of Tsushima | 9 |
| - | Cyberpunk 2077 | 9 |
| 17 | Tomb Raider | 16 |
| 14 | Average FPS | 12 |
| 20 | 1080p High | 16 |
| 14 | 1080p Ultra | 12 |
| 10 | 1440p Ultra | 9 |
| - | 4K Ultra | - |
| Medium | Margin of Error | Low |
| 34221 | GB6 Compute Score | 10472 |
| 65.7 img/sec | Background Blur | 21.2 img/sec |
| 33.3 img/sec | Face Detection | 10.9 img/sec |
| 1.01 Gpixels/sec | Horizon Detection | 0.34 Gpixels/sec |
| 1.39 Gpixels/sec | Edge Detection | 0.65 Gpixels/sec |
| 1.61 Gpixels/sec | Gaussian Blur | 0.51 Gpixels/sec |
| 0.48 Gpixels/sec | Feature Matching | 0.13 Gpixels/sec |
| 176 Gpixels/sec | Stereo Matching | 34.9 Gpixels/sec |
| 5012.4 FPS | Particle Physics | 1322.9 FPS |
| OpenCL | API | OpenCL |
| 3718 | G3D Mark Score | 1566 |
| 305 | G2D Mark | 400 |
| 26 FPS | DirectX 11 | 10 FPS |
| 23 FPS | DirectX 12 | 8 FPS |
| 1867 Ops/s | GPU Compute | 863 Ops/s |
| 4394 | GB6 ML Single Precision | 2257 |
| 4387 | GB6 ML Half Precision | 2817 |
| 3246 | GB6 ML Quantized | 1643 |
| 1911 | Image Classification (SP) | 924 |
| 2536 | Image Segmentation (HP) | 1148 |
| 4111 | Image Super Resolution (Q) | 2211 |
| 5175 | Face Detection (HP) | 2819 |
| 13374 | Pose Estimation (Q) | 7106 |
| 1514 | Text Classification (SP) | 832 |
| 1481 | Machine Translation (HP) | 1021 |
| 2070 | Object Detection (SP) | 1040 |
| 7813 | Depth Estimation (Q) | 3885 |
| 36156 | Style Transfer (SP) | 14495 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 30-50 GPixel/s | Pixel Fill Rate | 9-17 GPixel/s |
| 52-88 GTexel/s | Texture Fill Rate | 35-67 GTexel/s |
| 1.7-2.8 TFLOPS | FLOPS (FP32) | 1.1-2.2 TFLOPS |
| 10000 Mbps | Effective Memory Speed | System RAM Dependent |
| 64-bit | Bus | — |
| — | ECC | No |
| 80 GB/s | Memory Bandwidth | — |
| GDDR6 | Memory Type | System Shared |
| 2 GB | Memory Size | System RAM Dependent |
| 1250 MHz | Memory Clock | System RAM Dependent |
| PCIe 4.0 x4 | Interface | IGP |
| 10-30 W (configurable) | TGP | 15 W |
| TSMC | Manufacturing | TSMC |
| 12 nm | Fabrication Process | 7 nm |
| 200 mm² | Die Size | 210 mm² |
| 4.7 billion | Transistor Count | 4.94 billion |
| 23.5 MTr/mm² | Transistor Density | 23.52 MTr/mm² |
| 12 | DirectX | 12 |
| 1.2 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 1.2 | OpenCL | 2 |
| 7.5 | CUDA | — |
| No | Ray Tracing | No |
| No | DLSS | No |
| 1.4 | DisplayPort | — |
| Nvidia | Vendor | Amd |
| Discrete | Build | Integrated |
| August 25, 2020 | Released | February 13, 2018 |
| Laptop | Case | Laptop |
| Gaming | Purpose | Generic |
| Mid-range | Segment | Entry-level |
| Turing | Architecture | GCN 5.0 |
| TU117 | GPU Codename | Raven |
| -Intel Core i5 11300Hor above | Recommended CPU | -AMD Ryzen 7 4800Uor above |
| 1174 | Steel Nomad Lite Score | 1260 |
| 2118 | Time Spy | 1410 |
| — | Solar Bay | — |
| — | Port Royal | — |
| 5267 | Fire Strike | 4499 |
| 4259 | Wild Life Extreme | 2532 |
| 26243 | Night Raid | 17993 |
| 720-1395 MHz (configurable) | Base Clock | 300 MHz |
| 930-1575 MHz (configurable) | Boost Clock | 1100-2100 MHz (configurable) |
| 896 | Shading Units | 512 |
| 56 | Texture Mapping Units (TMUs) | 32 |
| 32 | Render Output Units (ROPs) | 8 |
| 14 | Compute Units (Pipelines) | 8 |
| No | Tensor Cores | No |
| No | Ray-tracing Cores | No |
| 64KB per cluster | L1 Cache | — |
| 512KB shared | L2 Cache | — |
| 2 IPC | Instructions Per Cycle | 2 IPC |