Quick/Comparisons
12 | Gaming | 34 |
21 | Workstation | 41 |
11 | AI/ML |
Compare with something else?
Compare GeForce GTX 970 with another →31 |
10 | Energy Efficiency | 43 |
15 | Quick Comparison Final Score | 37 |
| GM204 | Graphics Processor | GA104 (GN20-E5) |
| 1664 | Cores | 5120 |
| 104 / 56 | TMUs / ROPs | 160 / 80 |
| 286.89 | Blender GPU | 2998.92 |
| 32 | Forza Horizon 5 | 79 |
| 35 | The Witcher 3 | 114 |
| 57 | Counter-Strike 2 | 150 |
| 39 | Far Cry 6 | 94 |
| 18 | Hogwarts Legacy | 61 |
| 31 | Call of Duty | 88 |
| 17 | Ghost of Tsushima | 56 |
| 24 | Cyberpunk 2077 | 73 |
| 44 | Tomb Raider | 116 |
| 33 | Average FPS | 92 |
| 44 | 1080p High | 114 |
| 33 | 1080p Ultra | 92 |
| 23 | 1440p Ultra | 68 |
| 12 | 4K Ultra | 38 |
| Low | Margin of Error | Low |
| 28989 | GB6 Compute Score | 118499 |
| 77.1 img/sec | Background Blur | 170.3 img/sec |
| 38.1 img/sec | Face Detection | 113.6 img/sec |
| 1.28 Gpixels/sec | Horizon Detection | 5.59 Gpixels/sec |
| 2.08 Gpixels/sec | Edge Detection | 7.94 Gpixels/sec |
| 1.85 Gpixels/sec | Gaussian Blur | 6.19 Gpixels/sec |
| 0.17 Gpixels/sec | Feature Matching | 1.16 Gpixels/sec |
| 92.9 Gpixels/sec | Stereo Matching | 467 Gpixels/sec |
| 2443 FPS | Particle Physics | 15131.3 FPS |
| OpenCL | API | OpenCL |
| 9639 | G3D Mark Score | 15105 |
| 763 | G2D Mark | 630 |
| 71 FPS | DirectX 11 | 137 FPS |
| 41 FPS | DirectX 12 | 63 FPS |
| 4080 Ops/s | GPU Compute | 6702 Ops/s |
| 6791 | GB6 ML Single Precision | 17250 |
| 6738 | GB6 ML Half Precision | 27936 |
| 5128 | GB6 ML Quantized | 11249 |
| 1852 | Image Classification (SP) | 6969 |
| 3622 | Image Segmentation (HP) | 17142 |
| 8987 | Image Super Resolution (Q) | 16426 |
| 7074 | Face Detection (HP) | 31628 |
| 43726 | Pose Estimation (Q) | 74413 |
| 1446 | Text Classification (SP) | 2522 |
| 1576 | Machine Translation (HP) | 3639 |
| 2425 | Object Detection (SP) | 9002 |
| 12902 | Depth Estimation (Q) | 27790 |
| 86714 | Style Transfer (SP) | 160005 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 66 GPixel/s | Pixel Fill Rate | 103-130 GPixel/s |
| 123 GTexel/s | Texture Fill Rate | 206-259 GTexel/s |
| 3.9 TFLOPS | FLOPS (FP32) | 13.2-16.6 TFLOPS |
| GDDR5 | Memory Type | GDDR6 |
| 4 GB | Memory Size | 8 GB |
| 1753 MHz | Memory Clock | 1500-1750 MHz |
| 7000 Mbps | Effective Memory Speed | 12000-14000 Mbps |
| 256-bit | Bus | 256-bit |
| No | ECC | No |
| 224.4 GB/s | Memory Bandwidth | 448 GB/s |
| PCIe 3.0 x16 | Interface | PCIe 4.0 x16 |
| 148 W | TGP | 80-150 W (configurable) |
| TSMC | Manufacturing | Samsung |
| 28 nm | Fabrication Process | 8 nm |
| 398 mm² | Die Size | 392 mm² |
| 5 billion | Transistor Count | 17.4 billion |
| 12.56 MTr/mm² | Transistor Density | 44.39 MTr/mm² |
| 12 | DirectX | 12 |
| 1.3 | Vulkan | 1.2 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2 |
| 5.2 | CUDA | 8.6 |
| No | Ray Tracing | Yes |
| No | DLSS | DLSS 2 |
| 1.2 | DisplayPort | 1.4a |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| September 19, 2014 | Released | January 12, 2021 |
| Desktop | Case | Laptop |
| Gaming | Purpose | Gaming |
| Mid-range | Segment | High-end |
| Maxwell 2.0 | Architecture | Ampere |
| GM204 | GPU Codename | GA104 (GN20-E5) |
| — | Recommended CPU | -Intel Core i7 12700Hor above |
| 3577 | Steel Nomad Lite Score | 10214 |
| 3666 | Time Spy | 10477 |
| — | Solar Bay | 45215 |
| 4353 | Port Royal | 6247 |
| 12102 | Fire Strike | 26729 |
| 6972 | Wild Life Extreme | 21137 |
| 47591 | Night Raid | 82533 |
| 1050 MHz | Base Clock | 780-1215 MHz (configurable) |
| 1178 MHz | Boost Clock | 1290-1620 MHz (configurable) |
| 1664 | Shading Units | 5120 |
| 104 | Texture Mapping Units (TMUs) | 160 |
| 56 | Render Output Units (ROPs) | 80 |
| 13 | Compute Units (Pipelines) | 40 |
| No | Tensor Cores | 160 |
| No | Ray-tracing Cores | 40 |
| 48KB per cluster | L1 Cache | 128KB per cluster |
| 2MB shared | L2 Cache | 4MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |