Datasetpublic

Jr23xd23/gpu-database

GPU Database Comprehensive GPU specifications database with architecture, manufacturing, API support, performance details, and kernel development specs. 2,824 GPUs across NVIDIA, AMD, and Intel Part of RightNow — AI-powered code editor for GPU kernel development Data Vendor GPUs File NVIDIA 1,286 data/nvidia/all.json AMD 1,292 data/amd/all.json Intel 180 data/intel/all.json All 2,824 data/all-gpus.json Schema Each GPU contains… See the full description on the dataset page: https://huggingface.co/datasets/Jr23xd23/gpu-database.

sourceHugging Faceapache-2.0updated 9mo agoView on Hugging Face

1likes165downloads

rootREADME.md

README.md148 linesDownload Raw Back to root

1---2license: apache-2.03task_categories:4  - text-generation5  - feature-extraction6language:7  - en8tags:9  - gpu10  - nvidia11  - amd12  - intel13  - cuda14  - hardware15  - specifications16  - machine-learning17size_categories:18  - 1K<n<10K19---20 21# GPU Database22 23Comprehensive GPU specifications database with architecture, manufacturing, API support, performance details, and kernel development specs.24 25**2,824 GPUs** across NVIDIA, AMD, and Intel26 27Part of [RightNow](https://www.rightnowai.co) — AI-powered code editor for GPU kernel development28 29## Data30 31| Vendor | GPUs | File |32|--------|------|------|33| NVIDIA | 1,286 | `data/nvidia/all.json` |34| AMD | 1,292 | `data/amd/all.json` |35| Intel | 180 | `data/intel/all.json` |36| **All** | 2,824 | `data/all-gpus.json` |37 38## Schema39 40Each GPU contains up to 55 fields:41 42```json43{44  "name": "GeForce RTX 4090",45  "vendor": "nvidia",46  "manufacturer": "NVIDIA",47  "gpuName": "AD102",48  "architecture": "Ada Lovelace",49  "generation": "GeForce 40",50 51  "foundry": "TSMC",52  "processSize": 5,53  "transistors": 76.3,54  "transistorDensity": 125.3,55  "dieSize": 609.0,56  "chipPackage": "BGA-2150",57  "releaseDate": "2022-09-20",58 59  "baseClock": 2235.0,60  "boostClock": 2520.0,61  "memoryClock": 1313.0,62 63  "memorySize": 24.0,64  "memoryType": "GDDR6X",65  "memoryBus": 384,66  "memoryBandwidth": 1010.0,67 68  "shaders": 16384,69  "tmus": 512,70  "rops": 176,71  "sms": 128,72  "tensorCores": 512,73  "rtCores": 128,74  "coresPerSM": 128,75 76  "l1Cache": 128.0,77  "l2Cache": 72.0,78 79  "tdp": 450,80  "suggestedPSU": 850,81  "powerConnectors": "1x 16-pin",82 83  "length": 304.0,84  "width": 61.0,85  "slot": "Triple-slot",86  "displayOutputs": "1x HDMI 2.1, 3x DisplayPort 1.4a",87  "busInterface": "PCIe 4.0 x16",88 89  "pixelRate": 443.5,90  "textureRate": 1290.2,91  "fp16": 82.58,92  "fp32": 82.58,93  "fp64": 1.29,94 95  "directX": "12.2",96  "openGL": "4.6",97  "vulkan": "1.4",98  "openCL": "3.0",99  "cuda": "8.9",100  "shaderModel": "6.8",101 102  "warpSize": 32,103  "maxThreadsPerBlock": 1024,104  "maxThreadsPerSM": 1536,105  "maxBlocksPerSM": 24,106  "sharedMemPerSM": 102400,107  "registersPerSM": 65536,108 109  "url": "https://www.techpowerup.com/gpu-specs/geforce-rtx-4090.c3889"110}111```112 113Only populated fields are included — no nulls or zeros.114 115## Usage116 117**Python:**118```python119from datasets import load_dataset120 121ds = load_dataset("Jr23xd23/gpu-database")122```123 124**Direct JSON:**125```python126import requests127 128gpus = requests.get(129    'https://huggingface.co/datasets/Jr23xd23/gpu-database/resolve/main/data/nvidia/all.json'130).json()131 132rtx4090 = next(g for g in gpus if g['name'] == 'GeForce RTX 4090')133print(rtx4090['cuda'])            # 8.9134print(rtx4090['maxThreadsPerSM']) # 1536135```136 137## Source138 139Data sourced from [TechPowerUp GPU Database](https://www.techpowerup.com/gpu-specs/) via [dbgpu](https://github.com/painebenjamin/dbgpu).140 141## License142 143[Apache 2.0](LICENSE)144 145---146 147Built by [RightNow](https://www.rightnowai.co)148