{"689583":{"#nid":"689583","#data":{"type":"event","title":"Building tools for tensor programming","body":[{"value":"\u003Cul\u003E\u003Cli\u003E\u003Ch3\u003ETitle:\u003C\/h3\u003E\u003Ch3\u003E\u003Cbr\u003E\u0026nbsp;Building Tools For Tensor Programming\u003C\/h3\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003C\/li\u003E\u003Cli\u003E\u003Ch3\u003EAbstract:\u003C\/h3\u003E\u003Cp\u003EMany of today\u0027s most important applications rely on high-performance manipulation of matrices, tensors, and general arrays of data. \u0026nbsp;The neural networks at the heart of AI language models are a prime example. \u0026nbsp;The AI software stack relies on kernels (e.g., for matrix-matrix product) that are specifically designed to deliver throughput close to the hardware\u0027s theoretical peak. \u0026nbsp;However, writing kernels that can deliver maximum performance is extremely challenging. \u0026nbsp;In this talk, I will discuss some of the work we have done to simplify the task of authoring such kernels. \u0026nbsp;This includes new abstractions to allow for algebraic manipulation of array layouts and new programming models for more easily expressing the kernels themselves.\u003C\/p\u003E\u003C\/li\u003E\u003Cli\u003E\u003Ch3\u003EBio:\u003C\/h3\u003E\u003Cp\u003EMichael Garland has been a researcher at NVIDIA since 2006. \u0026nbsp;He joined the company as one of the founding members of NVIDIA Research and is currently Senior Director of Programming Systems Research. \u0026nbsp;He leads a research group focused on developing technologies that will help programmers take advantage of modern high-performance machines. \u0026nbsp;Their work spans the software stack with a particular focus on parallel algorithms, programming languages, compilers and runtime systems, and low-level hardware\/software interfaces. \u0026nbsp;He and his team have both published numerous academic articles and contributed to many software packages in broad use by CUDA developers.\u003C\/p\u003E\u003C\/li\u003E\u003C\/ul\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EThis talk highlights the critical role of high-performance kernels in enabling efficient matrix and tensor computations for modern AI applications, including language models. It discusses the challenges of achieving near-peak hardware performance and introduces new abstractions and programming models designed to simplify kernel development, particularly through more flexible array layout manipulation and clearer ways to express complex computations.\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"A talk on new abstractions and programming models that make it easier to write high-performance kernels for matrix and tensor computations in modern AI systems."}],"uid":"36458","created_gmt":"2026-04-09 12:46:52","changed_gmt":"2026-04-09 12:57:54","author":"mellis74","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2026-04-24T11:00:00-04:00","event_time_end":"2026-04-24T12:00:00-04:00","event_time_end_last":"2026-04-24T12:00:00-04:00","gmt_time_start":"2026-04-24 15:00:00","gmt_time_end":"2026-04-24 16:00:00","gmt_time_end_last":"2026-04-24 16:00:00","rrule":null,"timezone":"America\/New_York"},"location":"Groseclose Executive Boardroom GC402","extras":["free_food"],"hg_media":{"679900":{"id":"679900","type":"image","title":"Michael Garland","body":null,"created":"1775739349","gmt_created":"2026-04-09 12:55:49","changed":"1775739349","gmt_changed":"2026-04-09 12:55:49","alt":"Michael Garland","file":{"fid":"264121","name":"Michael-Garland.jpg","image_path":"\/sites\/default\/files\/2026\/04\/09\/Michael-Garland.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/04\/09\/Michael-Garland.jpg","mime":"image\/jpeg","size":25226,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/04\/09\/Michael-Garland.jpg?itok=KSjtyW2s"}}},"media_ids":["679900"],"groups":[{"id":"1242","name":"School of Industrial and Systems Engineering (ISYE)"}],"categories":[],"keywords":[],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"194683","name":"Talk"},{"id":"1795","name":"Seminar\/Lecture\/Colloquium"}],"invited_audience":[{"id":"194945","name":"Alumni"},{"id":"78761","name":"Faculty\/Staff"},{"id":"177814","name":"Postdoc"},{"id":"174045","name":"Graduate students"},{"id":"78751","name":"Undergraduate students"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}