{"id":22878011,"url":"https://github.com/gmasse/gpu-specs","last_synced_at":"2025-04-14T19:33:23.707Z","repository":{"id":267370213,"uuid":"899725307","full_name":"gmasse/gpu-specs","owner":"gmasse","description":"This project aims to centralize detailed specifications for GPUs, particularly in the context of AI workloads.","archived":false,"fork":false,"pushed_at":"2025-04-04T21:46:27.000Z","size":35,"stargazers_count":7,"open_issues_count":0,"forks_count":3,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-04-04T22:29:48.142Z","etag":null,"topics":["ai","artificial-intelligence","gpu","hardware","json","nvidia","specs"],"latest_commit_sha":null,"homepage":"https://g.masse.me/gpu-specs/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/gmasse.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-12-06T21:55:43.000Z","updated_at":"2025-04-04T21:46:25.000Z","dependencies_parsed_at":"2025-01-08T18:46:11.538Z","dependency_job_id":"98d8e66d-f144-451e-ad7e-f61a2953caeb","html_url":"https://github.com/gmasse/gpu-specs","commit_stats":null,"previous_names":["gmasse/gpu-specs"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gmasse%2Fgpu-specs","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gmasse%2Fgpu-specs/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gmasse%2Fgpu-specs/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gmasse%2Fgpu-specs/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/gmasse","download_url":"https://codeload.github.com/gmasse/gpu-specs/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248945865,"owners_count":21187396,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai","artificial-intelligence","gpu","hardware","json","nvidia","specs"],"created_at":"2024-12-13T16:17:35.975Z","updated_at":"2025-04-14T19:33:23.701Z","avatar_url":"https://github.com/gmasse.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# GPU Specifications for AI Projects\n\nWelcome to the **GPU Specifications for AI Projects** repository!  \nThis project aims to centralize detailed specifications for GPUs, particularly in the context of AI workloads. \n\n## Why This Project?\n\nFinding comprehensive and accurate GPU specifications for AI can be challenging. Many sources are either incomplete, inconsistent, or unavailable. This repository consolidates information into a single, accessible resource to assist AI researchers, developers, and enthusiasts.\n\nIf you believe this project duplicates an existing effort, or if a similar database already exists, please let us know by opening an issue or contacting us directly. Collaboration or linking efforts can benefit the entire community!\n\n## What’s Included?\n\nThe core of this project is a [JSON database](data/specs.json) containing detailed specifications of GPUs:\n- Precision support (e.g., FP32, FP16, INT8, etc.)\n- Compute \u0026 Memory capabilities\n- Additional details relevant to AI workloads\n\nThe JSON data is rendered into:\n- [HTML](https://g.masse.me/gpu-specs) for easy browsing.\n- [Markdown](specs.md) for lightweight documentation.\n\n## Contributing\n\nIf you spot errors, have missing data, or can add reliable sources, we welcome your input!\nSee the [CONTRIBUTING.md](CONTRIBUTING.md) file for details.\n\n## Roadmap\n\n- Add missing specifications for GPUs.\n- Expand support to cover GPUs from various vendors (NVIDIA, AMD, etc.).\n\n## Sample GPU Specifications\n\nBelow is a sample of the GPU specifications data included in the repository:\n\nAttribute (Unit) | H100 | L40S | A100 PCIe 80GB\n--- | --- | --- | ---\nFP64 (TFLOPS) | 25.6 | 1.4 | 9.7\nFP64 Tensor Core (TFLOPS) | 51 | 1.4 | 19.5\nFP32 (TFLOPS) | 51.2 | 91.6 | 19.5\nTF32 Tensor Core (TFLOPS) | ? | 183 | 156\nTF32 Tensor Core with Sparsity (TFLOPS) | 756 | 366 | 312\nFP16 (TFLOPS) | 204.9 | 91.6 | 78\nFP16 Tensor Core (TFLOPS) | ? | 362 | 312\nFP16 Tensor Core with Sparsity (TFLOPS) | ? | 733 | 624\nBF16 Tensor Core (TFLOPS) | ? | 362 | 312\nBF16 Tensor Core with Sparsity (TFLOPS) | 1513 | 733 | 624\nFP8 Tensor Core (TFLOPS) | ? | 733 | N/A\nFP8 Tensor Core with Sparsity (TFLOPS) | 3026 | 1466 | N/A\nFP4 Tensor Core (TFLOPS) | N/A | N/A | N/A\nFP4 Tensor Core with Sparsity (TFLOPS) | N/A | N/A | N/A\nINT8 Tensor Core (TOPS) | ? | 733 | 624\nINT8 Tensor Core with Sparsity (TOPS) | 3026 | 1466 | 1248\nINT4 Tensor Core (TOPS) | ? | 733 | ?\nINT4 Tensor Core with Sparsity (TOPS) | ? | 1466 | ?\n**Architecture Details** |  |  |  | \nGPU Name | H100 | L40S | A100 PCIe 80GB\nManufacturer | NVIDIA | NVIDIA | NVIDIA\nArchitecture | Hopper | Ada Lovelace | Ampere\nManufacturing Process | ? | ? | ?\nNVIDIA RT Cores | ? | 142 (3rd gen) | ?\nNVIDIA Tensor Cores | 456 (4th gen) | 568 (4th gen) | 432 (3rd gen)\nNVIDIA CUDA Cores | 14592 | 18176 | 6912\nGPU Memory (GB) | 80 | 48 | 80\nMemory Type | HBM2e | GDDR6 | HBM2e\nMemory Bandwidth (GB/s) | 2048 | 864 | 1935\nInterconnect Type | PCIe Gen5 | PCIe Gen4 | PCIe Gen4\nEncoders and Decoders | 0, 7 | 3, 3 | 0, 5\nCUDA Compute Capability | 9 | 8.9 | 8\nPower Consumption (W) | 350 | 300 | 300\nDie Size (mm2) | ? | ? | ?\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgmasse%2Fgpu-specs","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fgmasse%2Fgpu-specs","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgmasse%2Fgpu-specs/lists"}