Quick/Comparisons
The Radeon RX 7900 XTX dominates in gaming with a score of 81, thanks to its superior performance in demanding titles.
The Radeon RX 6650 XT and Radeon RX 7900 XTX are closely matched in content creation, with the former's efficiency and the latter's workstation capabilities making them nearly equal.
The Radeon RX 7900 XTX's capable AI/ML compute and large memory pool are matched by the Radeon RX 6650 XT's dedicated AI tensor cores, resulting in a tie in this use case.
30 | Gaming | 81 |
34 | Workstation | 75 |
31 | AI/ML | 74 |
39 | Energy Efficiency | 43 |
31 | Quick Comparison Final Score | 77 |
| Navi 23 | Graphics Processor | Navi 31 |
| 2048 | Cores | 6144 |
| 128 / 64 | TMUs / ROPs | 384 / 192 |
| 1138.24 | Blender GPU | 4055.35 |
| 76 | Forza Horizon 5 | 174 |
| 101 | The Witcher 3 | 256 |
| 130 | Counter-Strike 2 | 232 |
| 93 | Far Cry 6 | 182 |
| 59 | Hogwarts Legacy | 137 |
| 87 | Call of Duty | 255 |
| 63 | Ghost of Tsushima | 138 |
| 65 | Cyberpunk 2077 | 181 |
| 111 | Tomb Raider | 277 |
| 87 | Average FPS | 204 |
| 112 | 1080p High | 241 |
| 87 | 1080p Ultra | 204 |
| 61 | 1440p Ultra | 159 |
| 32 | 4K Ultra | 94 |
| Low | Margin of Error | Low |
| 90831 | GB6 Compute Score | 224686 |
| 220.1 img/sec | Background Blur | 423.9 img/sec |
| 128.3 img/sec | Face Detection | 270.8 img/sec |
| 2.94 Gpixels/sec | Horizon Detection | 9.23 Gpixels/sec |
| 4.13 Gpixels/sec | Edge Detection | 13.7 Gpixels/sec |
| 3.68 Gpixels/sec | Gaussian Blur | 11.1 Gpixels/sec |
| 0.94 Gpixels/sec | Feature Matching | 1.74 Gpixels/sec |
| 408.2 Gpixels/sec | Stereo Matching | 871 Gpixels/sec |
| 10694.7 FPS | Particle Physics | 29891.2 FPS |
| OpenCL | API | OpenCL |
| 17117 | G3D Mark Score | 31406 |
| 924 | G2D Mark | 1285 |
| 168 FPS | DirectX 11 | 356 FPS |
| 62 FPS | DirectX 12 | 133 FPS |
| 7461 Ops/s | GPU Compute | 17768 Ops/s |
| 17945 | GB6 ML Single Precision | 42092 |
| 21242 | GB6 ML Half Precision | 50673 |
| 13940 | GB6 ML Quantized | 32874 |
| 8220 | Image Classification (SP) | 18608 |
| 9263 | Image Segmentation (HP) | 25385 |
| 20723 | Image Super Resolution (Q) | 45641 |
| 27074 | Face Detection (HP) | 71321 |
| 76541 | Pose Estimation (Q) | 257258 |
| 2796 | Text Classification (SP) | 5240 |
| 4281 | Machine Translation (HP) | 7425 |
| 6880 | Object Detection (SP) | 21335 |
| 33354 | Depth Estimation (Q) | 66140 |
| 185358 | Style Transfer (SP) | 483867 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 169 GPixel/s | Pixel Fill Rate | 480 GPixel/s |
| 337 GTexel/s | Texture Fill Rate | 959 GTexel/s |
| 10.8 TFLOPS | FLOPS (FP32) | 61.4 TFLOPS |
| GDDR6 | Memory Type | GDDR6 |
| 8 GB | Memory Size | 24 GB |
| 2190 MHz | Memory Clock | 2500 MHz |
| 17500 Mbps | Effective Memory Speed | 20000 Mbps |
| 128-bit | Bus | 384-bit |
| No | ECC | No |
| 280.3 GB/s | Memory Bandwidth | 960 GB/s |
| PCIe 4.0 x8 | Interface | PCIe 4.0 x16 |
| 176 W | TGP | 355 W |
| TSMC | Manufacturing | TSMC |
| 7 nm | Fabrication Process | 5 nm |
| 237 mm² | Die Size | 529 mm² |
| 11 billion | Transistor Count | 57 billion |
| 46.41 MTr/mm² | Transistor Density | 107.75 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 2.1 | OpenCL | 2.2 |
| — | CUDA | — |
| Yes | Ray Tracing | Yes |
| No | DLSS | No |
| 1.4a | DisplayPort | 2.1 |
| Amd | Vendor | Amd |
| Discrete | Build | Discrete |
| May 10, 2022 | Released | November 3, 2022 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| RDNA 2.0 | Architecture | RDNA 3.0 |
| Navi 23 | GPU Codename | Navi 31 |
| -Intel Core i5 12600Kor above | Recommended CPU | -Intel Core i7 13700Kor above |
| 9265 | Steel Nomad Lite Score | 29870 |
| 9897 | Time Spy | 30766 |
| 38892 | Solar Bay | 123758 |
| 4897 | Port Royal | 16749 |
| 29357 | Fire Strike | 73217 |
| 18034 | Wild Life Extreme | 58530 |
| 116256 | Night Raid | 184767 |
| 2055 MHz | Base Clock | 1929 MHz |
| 2635 MHz | Boost Clock | 2498 MHz |
| 2048 | Shading Units | 6144 |
| 128 | Texture Mapping Units (TMUs) | 384 |
| 64 | Render Output Units (ROPs) | 192 |
| — | Compute Units (Pipelines) | — |
| No | Tensor Cores | — |
| 32 | Ray-tracing Cores | 96 |
| 128KB per cluster | L1 Cache | 256KB per cluster |
| 2MB shared | L2 Cache | 6MB shared |
| 2 IPC | Instructions Per Cycle | 4 IPC |
Compare with something else?
Compare Radeon RX 6650 with another →