BEGIN:VCALENDAR VERSION:2.0 PRODID:-//CERN//INDICO//EN BEGIN:VEVENT SUMMARY:[Online] Choosing GPU Programming Approaches DTSTART:20261109T080000Z DTEND:20261110T120000Z DTSTAMP:20261031T104300Z UID:indico-event-209@indico.ecap.work DESCRIPTION:Choosing GPU Programming Approaches\n\nSchedule & Format\n\nDa te: 2026\, November 9-10\nTimes:\n\nNov 9: 9:00 - 13:00 CE(S)T\nNov 10: 9: 00 - 13:00 CE(S)T\n\n\nFormat: Two half-days\nLocation: Online\nLanguage: English\n\nRegistered participants will receive the video conferencing lin k via email on the day before the course.\nInstructor\n\nDr. Sebastian Kuc kuk\, NHR@FAU\, certified NVIDIA DLI Ambassador\n\nThis course is organize d by Erlangen National High Performance Computing Center (NHR@FAU).\nCours e Description\nThe GPU programming landscape has grown from a single domin ant framework (CUDA) into a diverse ecosystem of vendor-neutral and perfor mance-portable alternatives. Choosing the right approach for a given appli cation - considering hardware targets\, portability requirements\, team ex pertise\, and performance goals - is a non-trivial decision. This course s urveys the most widely used GPU programming models: CUDA/HIP\, SYCL\, mode rn C++ parallel algorithms\, Thrust\, OpenACC\, OpenMP offloading\, and Ko kkos. For each approach\, participants see representative code\, learn the key abstractions\, and assess the trade-offs in portability\, expressiven ess\, and performance.\nPrerequisites\nKnowledge\n\nFamiliarity with moder n C++ programming (templates\, lambdas\, and the STL)\nPrior experience wi th at least one GPU programming approach is recommended but not required\n \nTechnical\n\nA modern web browser (exercises run on NHR@FAU's HPC cluste rs via JupyterHub - no local installation required)\n\nCourse Structure\n\ nGPU programming landscape: hardware diversity\, portability challenges\, and framework taxonomy\nLow-level\, vendor-specific approaches: CUDA and H IP\nOpen standard directives: OpenACC and OpenMP target offloading\nPerfor mance portability libraries: Kokkos\nC++ abstraction layers: SYCL\, Thrust \, and standard library parallel algorithms\nPerformance analysis: profili ng with Nsight Systems and Nsight Compute\, and common optimization patter ns\nHands-on programming challenge: porting STREAM\, a 2D stencil\, and a conjugate-gradient solver across multiple approaches\nComparative evaluati on: portability\, performance\, and practical considerations for framework selection\n\nLearning Outcomes\nAfter completing this course\, you will b e able to:\n\nDescribe the key abstractions and execution model of each ma jor GPU programming framework: CUDA/HIP\, SYCL\, OpenACC\, OpenMP offloadi ng\, Kokkos\, Thrust\, and standard C++ parallel algorithms\nImplement a r epresentative kernel in multiple frameworks and compare the resulting code \nEvaluate each approach across the dimensions of portability\, performanc e\, and programming effort\nSelect the most appropriate GPU programming mo del for a given combination of hardware targets and application requiremen ts\nProfile a GPU application with Nsight Systems and Nsight Compute and r elate observed performance to the choice of approach\nIdentify NHR@FAU cou rses and resources for deepening expertise in any specific framework\n\nRe gistration\, Wait List and Withdrawal Policy\nRegistration\nPlease registe r at the bottom of this page. Registration is open until a few days before the course starts\, or until the course is fully booked.\nPrices and Elig ibility\nThis course is open and free of charge for participants affiliate d with academic institutions in European Union (EU) member states and Hori zon 2020-associated countries.\nWait List\nIf the course reaches its maxim um capacity\, you can request to join the wait list by sending an email to nhr-training@fau.de. Please include your name and university affiliation in the message.\nWithdrawal Policy\nPlease only register if you are commit ted to attending the course. No-shows will be blacklisted and excluded fro m future events.\nIf you need to withdraw your registration\, please eithe r cancel it directly through the registration system or send an email to n hr-training@fau.de.\nAdditional Courses\nYou can find an up-to-date list o f all courses offered by NHR@FAU at https://hpc.fau.de/teaching/tutorials- and-courses/.\n\nhttps://indico.ecap.work/event/209/ LOCATION:Online URL:https://indico.ecap.work/event/209/ END:VEVENT END:VCALENDAR