Apply software pipelining (double-buffering) to a tiled GPU kernel to overlap global memory loads with Tensor Core computation. Covers prologue/loop/epilogue restructuring, LDG-register vs cp.async (LDGSTS) variant selection based on compute/load ratio, shared memory budget verification against architecture-specific occupancy cliffs, and SASS-level verification of load/compute overlap.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
pjt222/agent-almanac☆ 422026년 10월 15일 업데이트
Generate Skills from HTTP MCP servers with async job patterns (submit/status/result). Use when converting MCP specifications (.mcp.json) into reusable Skills using mcp_tool_catalog.yaml. Supports --lazy mode for context-saving generation. Also use for calling async MCP tools via JSON-RPC 2.0 with session-based polling.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
el-el-san/KamuiCodeSkillsCreator☆ 42026년 1월 29일 업데이트