Quick/Comparisons
12 | Gaming | 22 |
21 | Workstation | 17 |
11 | AI/ML |
Compare with something else?
Compare GeForce GTX 970 with another →| — |
10 | Energy Efficiency | 66 |
15 | Quick Comparison Final Score | 28 |
| GM204 | Graphics Processor | Strix Halo |
| 1664 | Cores | 1024 |
| 104 / 56 | TMUs / ROPs | 64 / 32 |
| 286.89 | Blender GPU | 376.14 |
| 32 | Forza Horizon 5 | — |
| 35 | The Witcher 3 | — |
| 57 | Counter-Strike 2 | — |
| 39 | Far Cry 6 | — |
| 18 | Hogwarts Legacy | — |
| 31 | Call of Duty | — |
| 17 | Ghost of Tsushima | — |
| 24 | Cyberpunk 2077 | — |
| 44 | Tomb Raider | — |
| 33 | Average FPS | — |
| 44 | 1080p High | — |
| 33 | 1080p Ultra | — |
| 23 | 1440p Ultra | — |
| 12 | 4K Ultra | — |
| Low | Margin of Error | — |
| 28989 | GB6 Compute Score | 35976 |
| 77.1 img/sec | Background Blur | 106.8 img/sec |
| 38.1 img/sec | Face Detection | 43.8 img/sec |
| 1.28 Gpixels/sec | Horizon Detection | 1.12 Gpixels/sec |
| 2.08 Gpixels/sec | Edge Detection | 1.17 Gpixels/sec |
| 1.85 Gpixels/sec | Gaussian Blur | 1.07 Gpixels/sec |
| 0.17 Gpixels/sec | Feature Matching | 0.5 Gpixels/sec |
| 92.9 Gpixels/sec | Stereo Matching | 163 Gpixels/sec |
| 2443 FPS | Particle Physics | 5820.4 FPS |
| OpenCL | API | OpenCL |
| 9639 | G3D Mark Score | 10431 |
| 763 | G2D Mark | 1094 |
| 71 FPS | DirectX 11 | 79 FPS |
| 41 FPS | DirectX 12 | 46 FPS |
| 4080 Ops/s | GPU Compute | 5066 Ops/s |
| 6791 | GB6 ML Single Precision | — |
| 6738 | GB6 ML Half Precision | — |
| 5128 | GB6 ML Quantized | — |
| 1852 | Image Classification (SP) | — |
| 3622 | Image Segmentation (HP) | — |
| 8987 | Image Super Resolution (Q) | — |
| 7074 | Face Detection (HP) | — |
| 43726 | Pose Estimation (Q) | — |
| 1446 | Text Classification (SP) | — |
| 1576 | Machine Translation (HP) | — |
| 2425 | Object Detection (SP) | — |
| 12902 | Depth Estimation (Q) | — |
| 86714 | Style Transfer (SP) | — |
| ONNX | Framework | — |
| DirectML | Backend | — |
| 66 GPixel/s | Pixel Fill Rate | 90 GPixel/s |
| 123 GTexel/s | Texture Fill Rate | 179 GTexel/s |
| 3.9 TFLOPS | FLOPS (FP32) | 5.7 TFLOPS |
| GDDR5 | Memory Type | System Shared |
| 4 GB | Memory Size | — |
| 1753 MHz | Memory Clock | — |
| 7000 Mbps | Effective Memory Speed | — |
| 256-bit | Bus | — |
| No | ECC | — |
| 224.4 GB/s | Memory Bandwidth | — |
| PCIe 3.0 x16 | Interface | PCIe 5.0 x16 |
| 148 W | TGP | 45-120 W (configurable) |
| TSMC | Manufacturing | TSMC |
| 28 nm | Fabrication Process | 4 nm |
| 398 mm² | Die Size | 308 mm² |
| 5 billion | Transistor Count | — |
| 12.56 MTr/mm² | Transistor Density | — |
| 12 | DirectX | 12.2 |
| 1.3 | Vulkan | 1.4 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.1 |
| 5.2 | CUDA | No |
| No | Ray Tracing | Yes |
| No | DLSS | No |
| 1.2 | DisplayPort | 2.1a |
| Nvidia | Vendor | Amd |
| Discrete | Build | Integrated |
| September 19, 2014 | Released | January 7, 2025 |
| Desktop | Case | Laptop |
| Gaming | Purpose | Generic |
| Mid-range | Segment | Mid-range |
| Maxwell 2.0 | Architecture | RDNA 3.5 |
| GM204 | GPU Codename | Strix Halo |
| — | Recommended CPU | -AMD Ryzen AI Max Pro 380or above |
| 3577 | Steel Nomad Lite Score | — |
| 3666 | Time Spy | — |
| — | Solar Bay | — |
| 4353 | Port Royal | — |
| 12102 | Fire Strike | — |
| 6972 | Wild Life Extreme | — |
| 47591 | Night Raid | — |
| 1050 MHz | Base Clock | 1295 MHz |
| 1178 MHz | Boost Clock | 2800 MHz |
| 1664 | Shading Units | 1024 |
| 104 | Texture Mapping Units (TMUs) | 64 |
| 56 | Render Output Units (ROPs) | 32 |
| 13 | Compute Units (Pipelines) | 16 |
| No | Tensor Cores | No |
| No | Ray-tracing Cores | 16 |
| 48KB per cluster | L1 Cache | 512KB per cluster |
| 2MB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |