Quick/Comparisons
The Radeon RX 6900 XT dominates in gaming with a 43-point lead, thanks to its significantly higher score of 57.
Although the Radeon RX 6900 XT has a respectable score, the Apple M4 GPU (8-core) takes the lead in content creation with a 25-point advantage.
The Radeon RX 6900 XT outperforms the Apple M4 GPU (8-core) by a wide margin in AI & Machine Learning, with a score of 56 compared to 0.
14 | Gaming | 57 |
24 | Workstation | 49 |
| — | AI/ML |
Compare with something else?
Compare Apple M4 GPU with another →97 | Energy Efficiency | 36 |
32 | Quick Comparison Final Score | 54 |
| Custom | Graphics Processor | Navi 21 |
| 1024 | Cores | 5120 |
| 64 / 32 | TMUs / ROPs | 320 / 128 |
| 1060.26 | Blender GPU | 2484.02 |
| — | Forza Horizon 5 | 139 |
| — | The Witcher 3 | 185 |
| — | Counter-Strike 2 | 208 |
| — | Far Cry 6 | 152 |
| — | Hogwarts Legacy | 103 |
| — | Call of Duty | 179 |
| — | Ghost of Tsushima | 112 |
| — | Cyberpunk 2077 | 127 |
| — | Tomb Raider | 208 |
| — | Average FPS | 157 |
| — | 1080p High | 188 |
| — | 1080p Ultra | 157 |
| — | 1440p Ultra | 121 |
| — | 4K Ultra | 71 |
| — | Margin of Error | Low |
| 30791 | GB6 Compute Score | 124121 |
| 57.4 img/sec | Background Blur | 231.6 img/sec |
| 41.3 img/sec | Face Detection | 152.5 img/sec |
| 1.21 Gpixels/sec | Horizon Detection | 4.26 Gpixels/sec |
| 1.74 Gpixels/sec | Edge Detection | 5.75 Gpixels/sec |
| 1.12 Gpixels/sec | Gaussian Blur | 5.87 Gpixels/sec |
| 0.43 Gpixels/sec | Feature Matching | 1.12 Gpixels/sec |
| 94.9 Gpixels/sec | Stereo Matching | 618.7 Gpixels/sec |
| 3902.5 FPS | Particle Physics | 17973.9 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 26671 |
| — | G2D Mark | 1062 |
| — | DirectX 11 | 279 FPS |
| — | DirectX 12 | 111 FPS |
| — | GPU Compute | 14356 Ops/s |
| — | GB6 ML Single Precision | 33651 |
| — | GB6 ML Half Precision | 36994 |
| — | GB6 ML Quantized | 24546 |
| — | Image Classification (SP) | 13868 |
| — | Image Segmentation (HP) | 17467 |
| — | Image Super Resolution (Q) | 32396 |
| — | Face Detection (HP) | 47808 |
| — | Pose Estimation (Q) | 178873 |
| — | Text Classification (SP) | 4358 |
| — | Machine Translation (HP) | 7910 |
| — | Object Detection (SP) | 15394 |
| — | Depth Estimation (Q) | 57351 |
| — | Style Transfer (SP) | 372610 |
| — | Framework | ONNX |
| — | Backend | DirectML |
| 47 GPixel/s | Pixel Fill Rate | 288 GPixel/s |
| 94 GTexel/s | Texture Fill Rate | 720 GTexel/s |
| 3 TFLOPS | FLOPS (FP32) | 23 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 16 GB |
| 7500 MHz | Memory Clock | 2000 MHz |
| 15000 Mbps | Effective Memory Speed | 16000 Mbps |
| 128-bit | Bus | 256-bit |
| No | ECC | No |
| 120 GB/s | Memory Bandwidth | 512 GB/s |
| Custom | Interface | PCIe 4.0 x16 |
| 15 W | TGP | 300 W |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 7 nm |
| — | Die Size | 520 mm² |
| — | Transistor Count | 26 billion |
| — | Transistor Density | 50 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.3 |
| — | OpenGL | 4.6 |
| — | OpenCL | 2.1 |
| — | CUDA | — |
| Yes | Ray Tracing | Yes |
| No | DLSS | No |
| — | DisplayPort | 1.4a |
| Apple | Vendor | Amd |
| Integrated | Build | Discrete |
| October 29, 2024 | Released | December 8, 2020 |
| Laptop | Case | Desktop |
| Professional | Purpose | Gaming |
| Mid-range | Segment | High-end |
| Apple M GPU | Architecture | RDNA 2.0 |
| Custom | GPU Codename | Navi 21 |
| -Apple M4 (8-Core)or above | Recommended CPU | -Intel Core i7 12700Kor above |
| 3450 | Steel Nomad Lite Score | 20125 |
| — | Time Spy | 20809 |
| 13908 | Solar Bay | 86980 |
| — | Port Royal | 10539 |
| — | Fire Strike | 57319 |
| 7928 | Wild Life Extreme | 39341 |
| — | Night Raid | 159586 |
| 500 MHz | Base Clock | 1825 MHz |
| 1470 MHz | Boost Clock | 2250 MHz |
| 1024 | Shading Units | 5120 |
| 64 | Texture Mapping Units (TMUs) | 320 |
| 32 | Render Output Units (ROPs) | 128 |
| 128 | Compute Units (Pipelines) | — |
| — | Tensor Cores | — |
| — | Ray-tracing Cores | 80 |
| — | L1 Cache | 128KB per cluster |
| — | L2 Cache | 4MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |