The National Energy Research Scientific Computing Center (NERSC) at Berkeley Lab seeks a highly motivated postdoctoral fellow to join the Application Performance team as part of NERSC's Exascale Science Acceleration Program (NESAP). NESAP postdocs collaborate with scientific teams to enable the solution of deep, meaningful problems across all program areas funded by the Department of Energy Office of Science.
The Perlmutter Supercomputer is NERSC's first production GPU-based system. Many workflows running at NERSC need to be adapted or optimized to run efficiently on GPUs. At the same time, solutions that put GPU performance in users' hands need to be portable. In this role, as part of a multidisciplinary team composed of computational and domain scientists, you'll develop and apply cutting-edge computer science, advanced performance analysis and software engineering to meet these challenges. In order to carry out this work, you will collaborate closely with one or more NESAP teams.
What You Will Do: Contribute to one or more NESAP team's scientific workflows, targeting the use of HPC resources in the following topics:Climate modelingMolecular DynamicsQuantum ChemistryGeophysics (Seismology)Metagenome AnalysisComputational Fluid DynamicsContribute to the identification, implementation, and application of new or advanced methods and technologies to improve the performance, portability, or productivity of scientific software.Publish and present results in peer-reviewed journals and conferences.Required Qualifications: Ph.D. in Computer Science, Computational Science, Applied Mathematics or an equivalent/related field awarded within the last five years.Experience in one or more HPC programming languages/ frameworks/ models:Python, C++, FortranMPI, Kokkos, OpenMP Offload, SYCL, HIP/CUDAImplementing/Optimizing/Integrating math libraries (fft, cufft, etc)Ability to work in an interdisciplinary team.Demonstrated track record of written and verbal communication of candidate-led results.Desired Qualifications: Knowledge of GPU architecture.Experience with programming GPUs.Experience with performance analysis and profiling tools (Perftools, NVIDIA Nsight, AMD uProf, Omniperf).Experience debugging distributed memory parallel applications.Experience with container technologies (e.g. Docker/ Podman, shifter, or equivalent).Knowledge of git version control and modern software development practices (e.g. Unit Testing, CI/CD, collaborative development).Notes: This is a full-time, 2 year, postdoctoral appointment with the possibility of renewal based upon satisfactory job performance, continuing availability of funds and ongoing operational needs. You must have less than 3 years of paid postdoctoral experience.The salary range for this position is $77,172 - $103,704 and is expected to start at $94,116 or above. Postdoctoral positions are paid on a step schedule per union contract and salaries will be predetermined based on postdoctoral step rates. Each step represents one full year of completed post-Ph.D. postdoctoral experience.This position is represented by a union for collective bargaining purposes.This position may be subject to a background check. Any convictions will be evaluated to determine if they directly relate to the responsibilities and requirements of the position. Having a conviction history will not automatically disqualify an applicant from being considered for employment.Work will be primarily performed at Lawrence Berkeley National Lab, 1 Cyclotron Road, Berkeley, CA. Position is eligible for an on-site or hybrid work schedule.
#J-18808-Ljbffr
