بعد أكاديمية كاوست
Muhand is studying at King AbdulAziz University. In summer 2026, Muhand worked as a summer student co-op at CERN on the CMS experiment, porting part of a reconstruction chain to run on both CPU and GPU, and presented the work at the KAUST Academy closing ceremony in September 2026.
المشروع الصيفي
تخصص الذكاء الاصطناعي · دفعة 2026
Porting the CMS Trackster to a GPU-Friendly Data Layout with Alpaka
Reconstruction software for the CMS experiment must run on GPUs to keep pace with the data rates of the High Luminosity LHC. Efficient GPU execution requires data laid out as a Structure of Arrays (SoA) rather than an Array of Structs. This project converts the Trackster a core data structure of the TICL reconstruction framework to an SoA layout using the CMSSW SoA templates and the Alpaka performance-portability layer, with host and device collections and a back-pointer design for its variable-length members. The layout is verified by execution: data written on the host, transferred to an NVIDIA GPU and read back is unchanged in every field. The layout is then applied to the pattern recognition step of TICL, whose interface was changed to return the collection rather than fill a growable vector, since an SoA must know its sizes before allocation. That conversion compiles for CUDA, ROCm and serial backends; runtime verification continues under CMS supervision. No performance benchmark was made.
مع أكاديمية كاوست
- دفعة 2026
المنظمة الأوروبية للأبحاث النووية (CERN) · التدريب الصيفي في الذكاء الاصطناعي
هل تريد تحديث بياناتك أو حذفها؟ تواصل معنا.