Quick/Comparisons
70 | Gaming | 22 |
72 | Workstation | 17 |
59 | AI/ML |
Compare with something else?
Compare GeForce RTX 4070 with another →| — |
46 | Energy Efficiency | 66 |
69 | Quick Comparison Final Score | 28 |
| AD103 | Graphics Processor | Strix Halo |
| 8448 | Cores | 1024 |
| 264 / 96 | TMUs / ROPs | 64 / 32 |
| 7089.35 | Blender GPU | 376.14 |
| 150 | Forza Horizon 5 | — |
| 198 | The Witcher 3 | — |
| 220 | Counter-Strike 2 | — |
| 152 | Far Cry 6 | — |
| 118 | Hogwarts Legacy | — |
| 188 | Call of Duty | — |
| 111 | Ghost of Tsushima | — |
| 155 | Cyberpunk 2077 | — |
| 249 | Tomb Raider | — |
| 171 | Average FPS | — |
| 200 | 1080p High | — |
| 171 | 1080p Ultra | — |
| 134 | 1440p Ultra | — |
| 79 | 4K Ultra | — |
| Low | Margin of Error | — |
| 237068 | GB6 Compute Score | 35976 |
| 331.7 img/sec | Background Blur | 106.8 img/sec |
| 211.6 img/sec | Face Detection | 43.8 img/sec |
| 9.63 Gpixels/sec | Horizon Detection | 1.12 Gpixels/sec |
| 13.7 Gpixels/sec | Edge Detection | 1.17 Gpixels/sec |
| 13.9 Gpixels/sec | Gaussian Blur | 1.07 Gpixels/sec |
| 2.15 Gpixels/sec | Feature Matching | 0.5 Gpixels/sec |
| 1210 Gpixels/sec | Stereo Matching | 163 Gpixels/sec |
| 33566 FPS | Particle Physics | 5820.4 FPS |
| OpenCL | API | OpenCL |
| 31798 | G3D Mark Score | 10431 |
| 1240 | G2D Mark | 1094 |
| 277 FPS | DirectX 11 | 79 FPS |
| 119 FPS | DirectX 12 | 46 FPS |
| 18264 Ops/s | GPU Compute | 5066 Ops/s |
| 32793 | GB6 ML Single Precision | — |
| 47412 | GB6 ML Half Precision | — |
| 24566 | GB6 ML Quantized | — |
| 12970 | Image Classification (SP) | — |
| 26364 | Image Segmentation (HP) | — |
| 37217 | Image Super Resolution (Q) | — |
| 58720 | Face Detection (HP) | — |
| 177604 | Pose Estimation (Q) | — |
| 3815 | Text Classification (SP) | — |
| 5829 | Machine Translation (HP) | — |
| 15675 | Object Detection (SP) | — |
| 62447 | Depth Estimation (Q) | — |
| 357918 | Style Transfer (SP) | — |
| ONNX | Framework | — |
| DirectML | Backend | — |
| 251 GPixel/s | Pixel Fill Rate | 90 GPixel/s |
| 689 GTexel/s | Texture Fill Rate | 179 GTexel/s |
| 44.1 TFLOPS | FLOPS (FP32) | 5.7 TFLOPS |
| GDDR6X | Memory Type | System Shared |
| 16 GB | Memory Size | — |
| 1313 MHz | Memory Clock | — |
| 21000 Mbps | Effective Memory Speed | — |
| 256-bit | Bus | — |
| No | ECC | — |
| 672.3 GB/s | Memory Bandwidth | — |
| PCIe 4.0 x16 | Interface | PCIe 5.0 x16 |
| 285 W | TGP | 45-120 W (configurable) |
| TSMC | Manufacturing | TSMC |
| 5 nm | Fabrication Process | 4 nm |
| 379 mm² | Die Size | 308 mm² |
| 45 billion | Transistor Count | — |
| 118.73 MTr/mm² | Transistor Density | — |
| 12 | DirectX | 12.2 |
| 1.3 | Vulkan | 1.4 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.1 |
| 8.9 | CUDA | No |
| Yes | Ray Tracing | Yes |
| DLSS 3 | DLSS | No |
| 1.4a | DisplayPort | 2.1a |
| Nvidia | Vendor | Amd |
| Discrete | Build | Integrated |
| January 24, 2024 | Released | January 7, 2025 |
| Desktop | Case | Laptop |
| Gaming | Purpose | Generic |
| High-end | Segment | Mid-range |
| Ada Lovelace | Architecture | RDNA 3.5 |
| AD103 | GPU Codename | Strix Halo |
| -Intel Core i7 14700Kor above | Recommended CPU | -AMD Ryzen AI Max Pro 380or above |
| 25741 | Steel Nomad Lite Score | — |
| 24222 | Time Spy | — |
| 116542 | Solar Bay | — |
| 15802 | Port Royal | — |
| 56340 | Fire Strike | — |
| 49901 | Wild Life Extreme | — |
| 176609 | Night Raid | — |
| 2340 MHz | Base Clock | 1295 MHz |
| 2610 MHz | Boost Clock | 2800 MHz |
| 8448 | Shading Units | 1024 |
| 264 | Texture Mapping Units (TMUs) | 64 |
| 96 | Render Output Units (ROPs) | 32 |
| 66 | Compute Units (Pipelines) | 16 |
| 264 | Tensor Cores | No |
| 66 | Ray-tracing Cores | 16 |
| 128KB per cluster | L1 Cache | 512KB per cluster |
| 48MB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |