Use natural language to request, run, and release jobs on a supercomputer cluster via an AI agent
翻譯準備中
AI 摘要 · README
外掛Python · ShellGitHub @globus-labs自行安裝營運者未確認
試用過了嗎? 點一下就好,不需要登入。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
安裝型專案無需部署網域也可驗證。個人儲存庫可透過 GitHub 登入驗證,組織儲存庫可使用上方的儲存庫驗證檔案。
正在載入完整 README。 以下為已儲存的摘錄。
Drive a supercomputer from Claude Code. Ask for a compute node in plain language; the agent finds the
facility, logs you in to Globus once, starts a one-node scheduler job (a block, and it asks before it spends
anything), runs your commands on that node, and releases it when you are done.
Seven minutes on video: what hpc-bridge is, why an agent should sit outside the facility, and a live run on a cluster.
Two commands inside Claude Code; the repository is its own plugin marketplace.
Then ask "What HPC facilities can I use?" — that first answer needs no login and no configuration.檢視完整 README ↗目前無法載入完整 README。請透過原文連結閱讀。 以下為已儲存的摘錄。
Drive a supercomputer from Claude Code. Ask for a compute node in plain language; the agent finds the
facility, logs you in to Globus once, starts a one-node scheduler job (a block, and it asks before it spends
anything), runs your commands on that node, and releases it when you are done.
Seven minutes on video: what hpc-bridge is, why an agent should sit outside the facility, and a live run on a cluster.
Two commands inside Claude Code; the repository is its own plugin marketplace.
Then ask "What HPC facilities can I use?" — that first answer needs no login and no configuration.檢視完整 README ↗