Come and join us to share your experiences and interest in the LLVM ecosystem.
Agenda
18.45 Doors Open
19.00 Welcome
19.05 "The Guts of the AMD Next Generation Fortran Compiler", Michael Klemm (AMD)
20.00 Social w/ food and drinks
22.30 Tentative End
The meetup is hosted in room 7 of building S1|03.
Take the main entrance from Hochschulstraße into the building, then turn left and head down the hallway. The door will be open.
Abstract
Today’s supercomputers are heterogeneous systems and GPUs have been established as accelerators to boost the execution of data parallel algorithms. As a directive-based programming model, the OpenMP API provides productive and high-performance features to use accelerators without a need to port code to vendor-specific programming languages.
This talk has three parts. Firstly, it will introduce target offloading using
the OpenMP API. Secondly, we will introduce the AMD Instinct(tm) Accelerator MI300A with unified shared memory and explain the impact on software development for accelerated code. The final topic will be an overview of the inner workings of the AMD Next Generation Fortran Compiler, a modern compiler based on LLVM Flang.
