{"id":30101137,"url":"https://github.com/torotoki/simple-paged-attention","last_synced_at":"2026-05-18T02:33:17.045Z","repository":{"id":305787448,"uuid":"1011407413","full_name":"torotoki/simple-paged-attention","owner":"torotoki","description":"A simple implementation of PagedAttention purely written in CUDA and C++.","archived":false,"fork":false,"pushed_at":"2025-08-09T10:18:29.000Z","size":46,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-08-09T12:16:06.161Z","etag":null,"topics":["attention","cpp","cuda","llm","transformer"],"latest_commit_sha":null,"homepage":"","language":"Cuda","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/torotoki.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"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,"zenodo":null}},"created_at":"2025-06-30T19:10:46.000Z","updated_at":"2025-08-09T10:18:32.000Z","dependencies_parsed_at":"2025-07-22T03:26:22.266Z","dependency_job_id":"dde298cd-8c79-4aba-9597-00ff3bb67b74","html_url":"https://github.com/torotoki/simple-paged-attention","commit_stats":null,"previous_names":["torotoki/simple-paged-attention"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/torotoki/simple-paged-attention","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/torotoki%2Fsimple-paged-attention","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/torotoki%2Fsimple-paged-attention/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/torotoki%2Fsimple-paged-attention/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/torotoki%2Fsimple-paged-attention/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/torotoki","download_url":"https://codeload.github.com/torotoki/simple-paged-attention/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/torotoki%2Fsimple-paged-attention/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33162626,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-17T22:39:12.733Z","status":"online","status_checked_at":"2026-05-18T02:00:06.436Z","response_time":71,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"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":["attention","cpp","cuda","llm","transformer"],"created_at":"2025-08-09T17:26:19.274Z","updated_at":"2026-05-18T02:33:17.040Z","avatar_url":"https://github.com/torotoki.png","language":"Cuda","funding_links":[],"categories":[],"sub_categories":[],"readme":"# simple-paged-attention\n\nThis is a CUDA and C++ implementation of PagedAttention.\n\nThis repo contains five types of attention implementations with and without the Key-Value caching mechanism (KV cache) as follows:\n\n| Method                      | Non KV cache | KV cache |\n|----------------------------------|:----------:|:----------:|\n| Standard causal attention on CPU | ✅        | -         |\n| Standard causal attention on GPU | ✅        | -             |\n| Attention with autoregressive output (common in inference) on CPU  | ✅        | ✅             |\n| Attention with autoregressive output (common in inference) on GPU  | ✅        | ✅            |\n| PagedAttention on GPU | - | 🚧 |\n\n## 📊 Benchmark Results:\n\n```\nCommand: attention_cpu\nAveraged Time (msec): 3.42877\n\nCommand: attention_gpu\nAveraged Time (msec): 1.26602\n\nCommand: attention_cpu_autoregressive\nEnable KV cache: 0\nAveraged Time (msec): 18.6311\n\nCommand: attention_cpu_autoregressive\nEnable KV cache: 1\nAveraged Time (msec): 3.65721\n\nCommand: attention_gpu_autoregressive\nEnable KV cache: 0\nAveraged Time (msec): 3.11079\n\nCommand: attention_gpu_autoregressive\nEnable KV cache: 1\nAveraged Time (msec): 2.88444\n```\n\n## 📥 Get Started\n\nComing soon: installation, usage examples, and code walkthroughs.\n\nStay tuned and ⭐️ the repo to keep updated!\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftorotoki%2Fsimple-paged-attention","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ftorotoki%2Fsimple-paged-attention","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftorotoki%2Fsimple-paged-attention/lists"}