[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"site-home-live-entry-setting":3,"site-header-contests":8,"site-header-camps":49,"study-material-detail-study-19":104},{"enabled":4,"title":5,"statusText":5,"buttonText":5,"campSlug":5,"stage":6,"classroomId":7,"hasClassroom":4},false,"",0,null,{"total":9,"list":10,"pageNum":44,"pageSize":45,"size":9,"startRow":44,"endRow":9,"pages":44,"prePage":6,"nextPage":6,"isFirstPage":46,"isLastPage":46,"hasPreviousPage":4,"hasNextPage":4,"navigatePages":47,"navigatepageNums":48,"navigateFirstPage":44,"navigateLastPage":44},3,[11,25,35],{"slug":12,"title":13,"logo":7,"liveUrl":14,"prize":15,"sponsor":16,"description":17,"organizer":18,"startTime":19,"endTime":20,"registrationDeadline":21,"status":22,"dashboardVisibility":23,"createdAt":24},"spring2026","2026春季启元人工智能大赛","https:\u002F\u002Fwww.bilibili.com\u002Fblackboard\u002Flive\u002Flive-activity-player.html?cid=1822703816&quality=0","200,000","麒麟软件有限公司","“2026春季启元人工智能大赛”是由启元实验室主办的面向全国人工智能领域开发者、科研人员及高校学生的权威赛事。大赛旨在促进与培养具备工程实践能力和创新能力的优秀人工智能领域开发者、推广国产软硬件技术与生态、助力国家人工智能产业高质量发展。\n各参赛选手需在规定的比赛周期内，依据赛事规则选定赛题并完成相应挑战，优胜者将共同角逐总额20万元的赛事奖金池，并且还将有机会获得实习和知名企业优质岗位内推资格。","启元实验室","2026-04-19 16:00:00","2026-07-13 16:00:00","2026-06-16 16:00:00","ongoing","private","2026-04-16 08:00:00",{"slug":26,"title":27,"logo":7,"liveUrl":14,"prize":28,"sponsor":7,"description":29,"organizer":18,"startTime":30,"endTime":31,"registrationDeadline":32,"status":33,"dashboardVisibility":23,"createdAt":34},"autumn2025","2025秋季启元人工智能大赛","150,000","“2025秋季启元人工智能大赛”是由启元实验室主办的面向全国人工智能领域开发者、科研人员及高校学生的权威赛事。大赛旨在促进与培养具备工程实践能力和创新能力的优秀人工智能领域开发者、推广国产软硬件技术与生态、助力国家人工智能产业高质量发展。","2025-11-08 08:00:00","2026-01-13 08:00:00","2025-12-30 08:00:00","ended","2025-11-03 08:00:00",{"slug":36,"title":37,"logo":7,"liveUrl":14,"prize":38,"sponsor":7,"description":39,"organizer":18,"startTime":40,"endTime":41,"registrationDeadline":42,"status":33,"dashboardVisibility":23,"createdAt":43},"summer2025","启元人工智能大赛(系统方向)","80,000","启元人工智能大赛由启元实验室主办，其中系统赛道基于启元实验室九源软件栈的相关开源项目，通过拉动开源社区力量，一方面促进国产智能计算生态的发展，另一方面也为人工智能系统方向的学生及从业者提供一个实践平台。","2025-07-08 08:00:00","2025-08-25 08:00:00","2025-07-22 08:00:00","2025-06-22 08:00:00",1,5,true,8,[44],{"total":45,"list":50,"pageNum":44,"pageSize":45,"size":45,"startRow":44,"endRow":45,"pages":44,"prePage":6,"nextPage":6,"isFirstPage":46,"isLastPage":46,"hasPreviousPage":4,"hasNextPage":4,"navigatePages":47,"navigatepageNums":103,"navigateFirstPage":44,"navigateLastPage":44},[51,62,73,84,94],{"id":52,"slug":53,"title":54,"logo":55,"publicLiveEmbedUrl":14,"description":56,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":16,"startTime":58,"endTime":59,"status":22,"dashboardVisibility":60,"createdAt":58,"updatedAt":61},"019eeeff-6b00-7d55-a629-efafded3e232","summer2026","2026年夏季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fuploads\u002F2026-06\u002F019eef7f-7816-7828-9309-79c2c61dd568.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=o29Dy5KMbkQVCWTVaFUX1%2F9v%2FI8%3D","由启元实验室发起的InfiniTensor大模型与人工智能系统训练营，面向开源社区、高校及产业界伙伴开放。活动以「普及AI全栈知识技术、培养工程实践与创新型人才、推广国产软硬件技术生态」为核心目标，助力国家人工智能产业高质量发展。","...","2026-06-22 08:00:00","2026-10-19 08:00:00","public","2026-10-12 08:00:00",{"id":63,"slug":64,"title":65,"logo":66,"publicLiveEmbedUrl":7,"description":67,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":68,"startTime":69,"endTime":70,"status":33,"dashboardVisibility":23,"createdAt":71,"updatedAt":72},"019af3f3-e06a-7245-aa8d-37dc99a69c73","winter2025","2025年冬季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fuploads\u002F2025-12\u002F019afc69-b7ca-7805-91fb-605429671c01.png?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=RF6et0tKpeNsR%2FwxsCjy7zBTZbU%3D","由启元实验室发起的InfiniTensor大模型与人工智能系统训练营，面向开源社区、高校及产业界伙伴开放。活动以「普及AI全栈知识技术、培养工程实践与创新型人才、推广国产软硬件技术生态」为核心目标，助力国家人工智能产业高质量发展。 \n","新华三技术有限公司","2025-12-25 08:00:00","2026-03-27 08:00:00","2025-12-06 08:00:00","2026-10-22 20:24:30",{"id":74,"slug":36,"title":75,"logo":76,"publicLiveEmbedUrl":7,"description":77,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":78,"startTime":79,"endTime":80,"status":81,"dashboardVisibility":23,"createdAt":82,"updatedAt":83},"01980342-208c-7803-b3f3-19e1d31960ba","2025年夏季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fuploads\u002F2025-07\u002F0198045e-28fe-78f0-95fe-23835ce58395.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=mArtxwAhCa0dE5Cm17SDDzrBtQ8%3D","InfiniTensor大模型与人工智能系统训练营由启元实验室及清华大学 AI 系统方向团队发起，面向开源社区、高校和产业界，普及人工智能各方面知识和技术，培养相关人才。训练营联合开源项目开发者、学术界大牛、工业界资深工程师参与授课与项目指导。课程采用「线上直播授课 + 在线答疑 + 项目实践辅导 」的方式进行，确保每位同学都能获得充分成长。训练营全程免费，通过视频、直播课、习题和项目实践，在人工智能飞速发展的背景下，为学员提供一窥深层技术的机会。","北京积算科技有限公司","2025-07-14 08:00:00","2025-09-30 08:00:00","upcoming","2025-07-13 08:00:00","2026-10-22 20:24:58",{"id":85,"slug":86,"title":87,"logo":88,"publicLiveEmbedUrl":7,"description":89,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":5,"startTime":90,"endTime":91,"status":81,"dashboardVisibility":23,"createdAt":92,"updatedAt":93},"01980cf9-a7d8-7796-ae86-0bbe05334add","winter2024","2024年冬季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Ftesting-cdn.infinitensor.com\u002Fuploads\u002F2025-07\u002F01981269-ce23-7624-a287-9f07dd0d85f6.jpg","InfiniTensor大模型与人工智能系统训练营由清华大学发起，面向开源社区、高校和产业界，普及人工智能各方面知识和技术，培养相关人才。训练营全程免费，通过视频、直播课、习题和项目实践，在人工智能飞速发展的背景下，为学员提供一窥深层技术的机会。报名学员请阅读“导学”课程页的训练营介绍。","2025-01-05 08:00:00","2025-03-22 08:00:00","2025-07-15 08:00:00","2026-10-22 20:24:39",{"id":95,"slug":96,"title":97,"logo":98,"publicLiveEmbedUrl":7,"description":99,"introduction":57,"organizer":18,"coOrganizer":5,"sponsor":5,"startTime":100,"endTime":101,"status":81,"dashboardVisibility":23,"createdAt":92,"updatedAt":102},"01980c20-363a-7dc3-8f5a-58a68e08dcc4","summer2024","2024年夏季InfiniTensor大模型与人工智能系统训练营","https:\u002F\u002Ftesting-cdn.infinitensor.com\u002Fuploads\u002F2025-07\u002F01981269-fbfa-788a-afcd-17fa9be74ff9.jpg","InfiniTensor大模型与人工智能系统训练营由清华大学发起，面向开源社区和高校，普及人工智能各方面知识和技术，培养相关人才。训练营全程免费，通过课程、习题和项目实践，在人工智能飞速发展的背景下，为学员提供一窥深层技术的机会。","2024-07-20 08:00:00","2024-09-29 08:00:00","2026-10-22 20:24:49",[44],{"currentItem":105,"processedData":125},{"img":106,"keywords":107,"treeType":44,"description":108,"showStatus":44,"updateTime":109,"type":110,"title":111,"content":112,"nextArticle":113,"createTime":109,"sortOrder":117,"canonicalPath":118,"id":119,"contentType":120,"prevArticle":121},"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_fm.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=Qaxr54F4nRa0VDDy6m2AdiUeKws%3D","InfiniTensor,冬季训练营,Triton-Ascend,算子开发,性能优化","讲解 Triton-Ascend 编程实践，覆盖算子编写流程、适配方法与性能优化要点。","2026-10-22 20:38:11",2,"InfiniTensor 冬季训练营Triton & 九齿第二课《Triton-Ascend 编程》","![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw01.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=oA%2B%2F%2BE5xy%2FZixnTTt4BBKshUf%2FE%3D)\r\n\r\n2026 年 1 月 16 日，2025 冬季 InfiniTensor 大模型与人工智能系统训练营 Triton & 九齿系列第二课《Triton-Ascend 编程》开讲。\r\n\r\n本节课将详解 Triton 在昇腾 NPU 上的适配与优化。前置课程为《[Triton 练气术](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FwZTZJJUHlZuu3vgNHQEzzg)》，如果有需要的同学可以点击复习。\r\n\r\n## Triton-Ascend 架构\r\n\r\n### Triton 简介\r\nTriton 是一种面向 GPU 的高效算子编程语言，其核心特点包括：\r\n- 高效的并行计算模型\r\n- 显式但高效的内存访问模型\r\n- 简洁的 API 设计\r\n\r\n### 昇腾使能：\u003Cbr>Triton-Ascend + AscendNPU IR\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw02.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=%2FbZPCpy89Lxgm62lc5FE0%2BuvTpw%3D)\r\n\r\n关键区别：\r\n昇腾的 CMD 架构要求数据访问尽可能连续，而 GPU 架构对数据连续性要求相对宽松。\r\n\r\n## Triton 算子开发范式\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw03.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=pPxTElgYRJtFlnfAeRlq9iL7QEI%3D)\r\n\r\n### Triton kernel 开发\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw04.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=H7LUxudgQk7tjRTxbaaUXhjljTo%3D)\r\n\r\n### Triton kernel 调用\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw05.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=aCDH5O95SiV8dUUv3Qt9MeFt2TA%3D)\r\n\r\n\r\n## 面向昇腾算子开发\r\n\r\n### Grid 设置与物理核匹配\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw06.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=c4yPd6MCDgbnzex6GD8Ohin1wXA%3D)\r\n\r\n\r\n昇腾的并行基于物理核，与 GPU 不同：\r\n- 推荐设置：Grid 大小应等于物理核数\r\n- 原因：避免逻辑核数量超过物理核，减少调度开销\r\n- 通过代码获取物理核数\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw07.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=qPX1M%2Fw9fdCPnMys3eDlc6mmb9E%3D)\r\n\r\n### 昇腾硬件架构\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw08.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=fFitWkf4EaHzqzFF7jtJUg1%2Bq9k%3D)\r\n\r\n\r\n## 性能优化\r\n### 算子性能测试\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw09.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=QPxtaeFXNnWn0%2BBigNPa%2BjIBj5g%3D)\r\n\r\n### 数据类型调整\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw10.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=Pf5D143t84wxlB2FEZe6zJ%2FNjqc%3D)\r\n\r\n### 离散访存优化\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw11.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=dK1n%2FUraOcGL4tNSCPkfovztiak%3D)\r\n\r\n### 访存调度优化\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw12.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=wM5eVqaikdjgCItgdaP59bsZDP8%3D)\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw13.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=%2B%2BQUHQuQb4LK7bEMV9shlFyizyk%3D)\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw14.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=RwZJIjO%2FUJxZ3oUiNghofx4bdgs%3D)\r\n\r\n## 高性能样例：matmul\r\n\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw15.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=6O%2FrkBJOYKXbHi3c7oGglzwddfk%3D)\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw16.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=eFaMQZCfCnfQCxmDAm3rgsvmiCk%3D)\r\n\r\n\r\n\r\n\r\n\r\n### 总结与展望\r\n随着 Triton-Ascend 的不断成熟，更多算子被 Triton-Ascend 适配，昇腾 NPU 的算子开发将更加高效，昇腾平台的生态将更加丰富，为 AI 应用在昇腾平台上的部署提供强大支持。\r\n\r\n\r\n\r\n![2025冬季训练营课程表](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw6.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=nRtncQE8d9WeqAJkebtpCuy2cYI%3D)\r\n\r\n### 💡 学习贴士\r\n更多细节及完整课程内容，请观看直播或查看课程回放。\r\n\r\n链接：https:\u002F\u002Fwww.infinitensor.com\u002Fcamp\u002Fwinter2025\r\n\r\n### 📚 训练营报名持续进行中\r\n扫码报名与加入群聊，或点击 **「阅读原文」** 直达官网报名\r\n\r\n\r\n![](https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw7.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=cXo0Vj%2B8TkiETVdZDjCIcCGN6sw%3D)",{"publishTime":7,"createTime":109,"canonicalPath":114,"id":115,"title":116},"\u002Fstudy-material\u002Fstudy-18","study-18","InfiniTensor 冬季训练营 AI 编译器第二课《AI 编译器中的前端优化》",90,"\u002Fstudy-material\u002Fstudy-19","study-19","markdown",{"publishTime":7,"createTime":109,"canonicalPath":122,"id":123,"title":124},"\u002Fstudy-material\u002Fstudy-20","study-20","InfiniTensor 冬季训练营 OpenCL 编程第一课《OpenCL 概述与运行时》",{"htmlContent":126,"headings":127},"\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw01.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=oA%2B%2F%2BE5xy%2FZixnTTt4BBKshUf%2FE%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>2026 年 1 月 16 日，2025 冬季 InfiniTensor 大模型与人工智能系统训练营 Triton & 九齿系列第二课《Triton-Ascend 编程》开讲。\u003C\u002Fp>\n\u003Cp>本节课将详解 Triton 在昇腾 NPU 上的适配与优化。前置课程为《\u003Ca href=\"https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FwZTZJJUHlZuu3vgNHQEzzg\">Triton 练气术\u003C\u002Fa>》，如果有需要的同学可以点击复习。\u003C\u002Fp>\n\u003Ch2 id=\"triton-ascend-架构\">\u003Ca class=\"anchor\" href=\"#triton-ascend-架构\" aria-label=\"跳转到标题：Triton-Ascend 架构\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>Triton-Ascend 架构\u003C\u002Fh2>\u003Ch3 id=\"triton-简介\">\u003Ca class=\"anchor\" href=\"#triton-简介\" aria-label=\"跳转到标题：Triton 简介\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>Triton 简介\u003C\u002Fh3>\u003Cp>Triton 是一种面向 GPU 的高效算子编程语言，其核心特点包括：\u003C\u002Fp>\n\u003Cul>\n\u003Cli>高效的并行计算模型\u003C\u002Fli>\n\u003Cli>显式但高效的内存访问模型\u003C\u002Fli>\n\u003Cli>简洁的 API 设计\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3 id=\"昇腾使能-triton-ascend-ascendnpu-ir\">\u003Ca class=\"anchor\" href=\"#昇腾使能-triton-ascend-ascendnpu-ir\" aria-label=\"跳转到标题：昇腾使能：Triton-Ascend + AscendNPU IR\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>昇腾使能：\u003Cbr>Triton-Ascend + AscendNPU IR\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw02.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=%2FbZPCpy89Lxgm62lc5FE0%2BuvTpw%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>关键区别：\u003Cbr>昇腾的 CMD 架构要求数据访问尽可能连续，而 GPU 架构对数据连续性要求相对宽松。\u003C\u002Fp>\n\u003Ch2 id=\"triton-算子开发范式\">\u003Ca class=\"anchor\" href=\"#triton-算子开发范式\" aria-label=\"跳转到标题：Triton 算子开发范式\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>Triton 算子开发范式\u003C\u002Fh2>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw03.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=pPxTElgYRJtFlnfAeRlq9iL7QEI%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"triton-kernel-开发\">\u003Ca class=\"anchor\" href=\"#triton-kernel-开发\" aria-label=\"跳转到标题：Triton kernel 开发\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>Triton kernel 开发\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw04.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=H7LUxudgQk7tjRTxbaaUXhjljTo%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"triton-kernel-调用\">\u003Ca class=\"anchor\" href=\"#triton-kernel-调用\" aria-label=\"跳转到标题：Triton kernel 调用\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>Triton kernel 调用\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw05.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=aCDH5O95SiV8dUUv3Qt9MeFt2TA%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch2 id=\"面向昇腾算子开发\">\u003Ca class=\"anchor\" href=\"#面向昇腾算子开发\" aria-label=\"跳转到标题：面向昇腾算子开发\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>面向昇腾算子开发\u003C\u002Fh2>\u003Ch3 id=\"grid-设置与物理核匹配\">\u003Ca class=\"anchor\" href=\"#grid-设置与物理核匹配\" aria-label=\"跳转到标题：Grid 设置与物理核匹配\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>Grid 设置与物理核匹配\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw06.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=c4yPd6MCDgbnzex6GD8Ohin1wXA%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>昇腾的并行基于物理核，与 GPU 不同：\u003C\u002Fp>\n\u003Cul>\n\u003Cli>推荐设置：Grid 大小应等于物理核数\u003C\u002Fli>\n\u003Cli>原因：避免逻辑核数量超过物理核，减少调度开销\u003C\u002Fli>\n\u003Cli>通过代码获取物理核数\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw07.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=qPX1M%2Fw9fdCPnMys3eDlc6mmb9E%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"昇腾硬件架构\">\u003Ca class=\"anchor\" href=\"#昇腾硬件架构\" aria-label=\"跳转到标题：昇腾硬件架构\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>昇腾硬件架构\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw08.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=fFitWkf4EaHzqzFF7jtJUg1%2Bq9k%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch2 id=\"性能优化\">\u003Ca class=\"anchor\" href=\"#性能优化\" aria-label=\"跳转到标题：性能优化\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>性能优化\u003C\u002Fh2>\u003Ch3 id=\"算子性能测试\">\u003Ca class=\"anchor\" href=\"#算子性能测试\" aria-label=\"跳转到标题：算子性能测试\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>算子性能测试\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw09.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=QPxtaeFXNnWn0%2BBigNPa%2BjIBj5g%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"数据类型调整\">\u003Ca class=\"anchor\" href=\"#数据类型调整\" aria-label=\"跳转到标题：数据类型调整\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>数据类型调整\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw10.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=Pf5D143t84wxlB2FEZe6zJ%2FNjqc%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"离散访存优化\">\u003Ca class=\"anchor\" href=\"#离散访存优化\" aria-label=\"跳转到标题：离散访存优化\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>离散访存优化\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw11.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=dK1n%2FUraOcGL4tNSCPkfovztiak%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"访存调度优化\">\u003Ca class=\"anchor\" href=\"#访存调度优化\" aria-label=\"跳转到标题：访存调度优化\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>访存调度优化\u003C\u002Fh3>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw12.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=wM5eVqaikdjgCItgdaP59bsZDP8%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw13.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=%2B%2BQUHQuQb4LK7bEMV9shlFyizyk%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw14.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=RwZJIjO%2FUJxZ3oUiNghofx4bdgs%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch2 id=\"高性能样例-matmul\">\u003Ca class=\"anchor\" href=\"#高性能样例-matmul\" aria-label=\"跳转到标题：高性能样例：matmul\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>高性能样例：matmul\u003C\u002Fh2>\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw15.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=6O%2FrkBJOYKXbHi3c7oGglzwddfk%3D\" alt=\"\">\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_19_zw16.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=eFaMQZCfCnfQCxmDAm3rgsvmiCk%3D\" alt=\"\">\u003C\u002Fp>\n\u003Ch3 id=\"总结与展望\">\u003Ca class=\"anchor\" href=\"#总结与展望\" aria-label=\"跳转到标题：总结与展望\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>总结与展望\u003C\u002Fh3>\u003Cp>随着 Triton-Ascend 的不断成熟，更多算子被 Triton-Ascend 适配，昇腾 NPU 的算子开发将更加高效，昇腾平台的生态将更加丰富，为 AI 应用在昇腾平台上的部署提供强大支持。\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw6.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=nRtncQE8d9WeqAJkebtpCuy2cYI%3D\" alt=\"2025冬季训练营课程表\">\u003C\u002Fp>\n\u003Ch3 id=\"学习贴士\">\u003Ca class=\"anchor\" href=\"#学习贴士\" aria-label=\"跳转到标题：💡 学习贴士\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>💡 学习贴士\u003C\u002Fh3>\u003Cp>更多细节及完整课程内容，请观看直播或查看课程回放。\u003C\u002Fp>\n\u003Cp>链接：\u003Ca href=\"https:\u002F\u002Fwww.infinitensor.com\u002Fcamp\u002Fwinter2025\">https:\u002F\u002Fwww.infinitensor.com\u002Fcamp\u002Fwinter2025\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch3 id=\"训练营报名持续进行中\">\u003Ca class=\"anchor\" href=\"#训练营报名持续进行中\" aria-label=\"跳转到标题：📚 训练营报名持续进行中\">\u003Cspan class=\"octicon octicon-link\">\u003C\u002Fspan>\u003C\u002Fa>📚 训练营报名持续进行中\u003C\u002Fh3>\u003Cp>扫码报名与加入群聊，或点击 \u003Cstrong>「阅读原文」\u003C\u002Fstrong> 直达官网报名\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Foss.infinitensor.com\u002Fprod\u002Fimages\u002FKnowledge_Base\u002FKB_01_zw7.jpg?Expires=1792752820&OSSAccessKeyId=LTAI5t9R6uzvTa69CT92cFvV&Signature=cXo0Vj%2B8TkiETVdZDjCIcCGN6sw%3D\" alt=\"\">\u003C\u002Fp>\n",[128,131,134,137,140,143,146,148,151,153,155,157,159,161,163,166,168,171],{"id":129,"text":130,"level":110},"triton-ascend-架构","Triton-Ascend 架构",{"id":132,"text":133,"level":9},"triton-简介","Triton 简介",{"id":135,"text":136,"level":9},"昇腾使能-triton-ascend-ascendnpu-ir","昇腾使能：Triton-Ascend + AscendNPU IR",{"id":138,"text":139,"level":110},"triton-算子开发范式","Triton 算子开发范式",{"id":141,"text":142,"level":9},"triton-kernel-开发","Triton kernel 开发",{"id":144,"text":145,"level":9},"triton-kernel-调用","Triton kernel 调用",{"id":147,"text":147,"level":110},"面向昇腾算子开发",{"id":149,"text":150,"level":9},"grid-设置与物理核匹配","Grid 设置与物理核匹配",{"id":152,"text":152,"level":9},"昇腾硬件架构",{"id":154,"text":154,"level":110},"性能优化",{"id":156,"text":156,"level":9},"算子性能测试",{"id":158,"text":158,"level":9},"数据类型调整",{"id":160,"text":160,"level":9},"离散访存优化",{"id":162,"text":162,"level":9},"访存调度优化",{"id":164,"text":165,"level":110},"高性能样例-matmul","高性能样例：matmul",{"id":167,"text":167,"level":9},"总结与展望",{"id":169,"text":170,"level":9},"学习贴士","💡 学习贴士",{"id":172,"text":173,"level":9},"训练营报名持续进行中","📚 训练营报名持续进行中"]