A*STAR Research
IHPC - Performance portability of CFD codes Traineeship [GRIT@Gov]
A*STAR Research, Montgomery, Alabama, United States
IHPC - Performance portability of CFD codes Traineeship [GRIT@Gov]
Posting Date : 05 Jan 2026 | Closing Date :02 Feb 2026
What the role is As a Trainee at the Institute of High Performance Computing (IHPC), you will contribute to improving the performance and portability of Computational Fluid Dynamics (CFD) codes across different high-performance computing (HPC) architectures.
What will you be working on Understand some of the existing implementation of CPU-based CFD codes and help refactor these codes to utilize GPU accelerators.
What we are looking for
B.S. in CS/ME/Aero (or similar) with strong programming fundamentals.
Solid experience in C/C++, data structures, and debugging on Linux.
Coursework or experience in numerical methods / linear algebra (e.g., PDE discretization, iterative solvers).
Intro to parallel programming: OpenMP, basic CUDA/HIP/SYCL or a clear self-driven project showing GPU curiosity.
Learning and Development Focus
GPU programming models & portability layers: Practical fluency in some higher-level portability frameworks like Kokkos or RAJA.
Profiling & optimization: Using Nsight, rocprof, VTune/perf to find computational bottleneck.
MPI+X scaling: Single- and multi-GPU runs, overlapping communication/computation.
Software engineering for HPC in the context of CFD.
The above eligibility criteria are not exhaustive. A*STAR may include additional selection criteria based on its prevailing recruitment policies. These policies may be amended from time to time without notice. We regret that only shortlisted candidates will be notified.
#J-18808-Ljbffr
Posting Date : 05 Jan 2026 | Closing Date :02 Feb 2026
What the role is As a Trainee at the Institute of High Performance Computing (IHPC), you will contribute to improving the performance and portability of Computational Fluid Dynamics (CFD) codes across different high-performance computing (HPC) architectures.
What will you be working on Understand some of the existing implementation of CPU-based CFD codes and help refactor these codes to utilize GPU accelerators.
What we are looking for
B.S. in CS/ME/Aero (or similar) with strong programming fundamentals.
Solid experience in C/C++, data structures, and debugging on Linux.
Coursework or experience in numerical methods / linear algebra (e.g., PDE discretization, iterative solvers).
Intro to parallel programming: OpenMP, basic CUDA/HIP/SYCL or a clear self-driven project showing GPU curiosity.
Learning and Development Focus
GPU programming models & portability layers: Practical fluency in some higher-level portability frameworks like Kokkos or RAJA.
Profiling & optimization: Using Nsight, rocprof, VTune/perf to find computational bottleneck.
MPI+X scaling: Single- and multi-GPU runs, overlapping communication/computation.
Software engineering for HPC in the context of CFD.
The above eligibility criteria are not exhaustive. A*STAR may include additional selection criteria based on its prevailing recruitment policies. These policies may be amended from time to time without notice. We regret that only shortlisted candidates will be notified.
#J-18808-Ljbffr