Quick/Comparisons
The Radeon RX 6800's superior gaming performance is due to its higher score of 45, significantly outpacing the GeForce GTX 1650 SUPER's score of 16.
Both the GeForce GTX 1650 SUPER and Radeon RX 6800 have similar scores of 25 and 47, respectively, making them nearly tied in content creation performance.
The Radeon RX 6800's score of 44 in AI/ML outperforms the GeForce GTX 1650 SUPER's score of 14, making it the clear winner in this use case.
16 | Gaming | 45 |
25 | Workstation | 47 |
14 | AI/ML |
Compare with something else?
Compare GeForce GTX 1650 with another →44 |
50 | Energy Efficiency | 36 |
19 | Quick Comparison Final Score | 45 |
| TU116 | Graphics Processor | Navi 21 |
| 1280 | Cores | 3840 |
| 80 / 32 | TMUs / ROPs | 240 / 96 |
| 555.23 | Blender GPU | 1894.91 |
| 42 | Forza Horizon 5 | 116 |
| 48 | The Witcher 3 | 152 |
| 72 | Counter-Strike 2 | 186 |
| 52 | Far Cry 6 | 125 |
| 27 | Hogwarts Legacy | 88 |
| 34 | Call of Duty | 134 |
| 24 | Ghost of Tsushima | 90 |
| 29 | Cyberpunk 2077 | 98 |
| 56 | Tomb Raider | 168 |
| 43 | Average FPS | 129 |
| 55 | 1080p High | 157 |
| 43 | 1080p Ultra | 129 |
| 31 | 1440p Ultra | 95 |
| 15 | 4K Ultra | 55 |
| Low | Margin of Error | Low |
| 57455 | GB6 Compute Score | 135060 |
| 106.6 img/sec | Background Blur | 298.6 img/sec |
| 58.3 img/sec | Face Detection | 186.2 img/sec |
| 1.98 Gpixels/sec | Horizon Detection | 4.75 Gpixels/sec |
| 2.67 Gpixels/sec | Edge Detection | 6.68 Gpixels/sec |
| 2.79 Gpixels/sec | Gaussian Blur | 5.58 Gpixels/sec |
| 0.75 Gpixels/sec | Feature Matching | 1.16 Gpixels/sec |
| 248.2 Gpixels/sec | Stereo Matching | 579 Gpixels/sec |
| 7783.2 FPS | Particle Physics | 18755.1 FPS |
| OpenCL | API | OpenCL |
| 10186 | G3D Mark Score | 22023 |
| 749 | G2D Mark | 998 |
| 73 FPS | DirectX 11 | 212 FPS |
| 45 FPS | DirectX 12 | 89 FPS |
| 4460 Ops/s | GPU Compute | 10764 Ops/s |
| 8531 | GB6 ML Single Precision | 25097 |
| 7319 | GB6 ML Half Precision | 27988 |
| 6132 | GB6 ML Quantized | 19116 |
| 3533 | Image Classification (SP) | 11245 |
| 5144 | Image Segmentation (HP) | 13617 |
| 7927 | Image Super Resolution (Q) | 27747 |
| 8860 | Face Detection (HP) | 38124 |
| 30433 | Pose Estimation (Q) | 119493 |
| 1622 | Text Classification (SP) | 2425 |
| 2754 | Machine Translation (HP) | 5228 |
| 4278 | Object Detection (SP) | 11986 |
| 15240 | Depth Estimation (Q) | 46083 |
| 66527 | Style Transfer (SP) | 281860 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 55 GPixel/s | Pixel Fill Rate | 202 GPixel/s |
| 138 GTexel/s | Texture Fill Rate | 505 GTexel/s |
| 4.4 TFLOPS | FLOPS (FP32) | 16.2 TFLOPS |
| GDDR6 | Memory Type | GDDR6 |
| 4 GB | Memory Size | 16 GB |
| 2000 MHz | Memory Clock | 2000 MHz |
| 12000 Mbps | Effective Memory Speed | 16000 Mbps |
| 128-bit | Bus | 256-bit |
| No | ECC | No |
| 192 GB/s | Memory Bandwidth | 512 GB/s |
| PCIe 3.0 x16 | Interface | PCIe 4.0 x16 |
| 100 W | TGP | 250 W |
| TSMC | Manufacturing | TSMC |
| 12 nm | Fabrication Process | 7 nm |
| 284 mm² | Die Size | 520 mm² |
| 6 billion | Transistor Count | 26 billion |
| 21.13 MTr/mm² | Transistor Density | 50 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.1 |
| 7.5 | CUDA | — |
| No | Ray Tracing | Yes |
| No | DLSS | No |
| 1.4a | DisplayPort | 1.4a |
| Nvidia | Vendor | Amd |
| Discrete | Build | Discrete |
| April 23, 2019 | Released | October 28, 2020 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| Entry-level | Segment | High-end |
| Turing | Architecture | RDNA 2.0 |
| TU116 | GPU Codename | Navi 21 |
| -Intel Core i5 9400or above | Recommended CPU | -Intel Core i5 12600or above |
| 4388 | Steel Nomad Lite Score | 14548 |
| 4691 | Time Spy | 16220 |
| — | Solar Bay | 64755 |
| — | Port Royal | 8137 |
| 12265 | Fire Strike | 44377 |
| 9633 | Wild Life Extreme | 29764 |
| 60029 | Night Raid | 142721 |
| 1530 MHz | Base Clock | 1700 MHz |
| 1725 MHz | Boost Clock | 2105 MHz |
| 1280 | Shading Units | 3840 |
| 80 | Texture Mapping Units (TMUs) | 240 |
| 32 | Render Output Units (ROPs) | 96 |
| 20 | Compute Units (Pipelines) | — |
| No | Tensor Cores | No |
| No | Ray-tracing Cores | 60 |
| 64KB per cluster | L1 Cache | 128KB per cluster |
| 1024KB shared | L2 Cache | 4MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |