Nicholas Malaya Email & Phone Number

Santa Clara, California, Us

Santa Clara, California, Us

My group and I moved from AMD Research into the newly formed Datacenter GPU Business Unit. I was a founding member of the "AMD Application Solutions Group" which is focused on application readiness and performance optimization for high performance computing and machine learning applications.My role is to act as AMD's technical lead for the Centers of Excellence (COE) for the Frontier and El Capitan Exascale Supercomputers. I am also AMD's technical lead for Exascale application performance, where I work to ensure that key benchmarks run efficiently on the world's largest computers from Day-1 of machine deployment. In this position I was the technical lead for the AMD team that worked with ORNL and HPE to break the Exascale barrier in May, 2022. In addition to my technical work, I regularly present data and recommendations to C-level executives. I enjoy working across technical departments and business units to contribute to strategy and achieve AMD's business objectives. I routinely deal with sensitive information and work closely with our legal department. AMD has certified me for customer facing roles, noting my work has financial and marketing implications for the company (i.e., SOX compliance). Finally, I remain closely engaged with my colleagues in our Research organization, and participate in forward looking research explorations to create disruptive new technologies for AMD.

Santa Clara, California, Us

AMD's technical lead for the Centers of Excellence (COE) for the Frontier and El Capitan Exascale Supercomputers. These COEs are focused on working closely with our partners DOE and HPE on ensuring application readiness, so that key workloads can run on the computers from Day-1 of machine deployment. In this role I set technical direction, manage development schedules, and act as the key point of contact both within AMD across multiple business units and outside of the company with customers and partners. I own and track KPIs for work pertaining to application readiness for future exascale systems leveraging AMD technology (CPUs and GPUs). I am AMD's representative to the COE management councils, where I work closely with HPE and DOE, represent my organization's interests, and set 6-month work schedules. I am also AMD's technical lead for performance projections. In this role, I am responsible for projecting application performance from current hardware and software to future systems. These models were key inputs for the CORAL-2 supercomputer bids that AMD participated in. This requires that I understand and communicate key algorithmic, software, and hardware trends that will occur in the HPC and AI fields. I also evaluate, contribute, and track technical roadmaps for core products. I regularly present these results for review at the executive level. I routinely deal with sensitive information and work closely with our legal department.

Santa Clara, California, Us

Researcher working broadly on: HPC Advanced Software Technology; Exascale Computing; Supercomputer Co-design; Machine Learning; Deep Learning; Uncertainty Quantification; GPU and parallel algorithms; Computer Architecture. I work extensively with HPC Software and tools, such as: CUDA/HIP, MPI, OpenMP, Slurm, Python, C/C++, Fortran, Make, CMake, Tensorflow, PyTorch, Caffe2, etc. Working with the Department of Energy's Exascale Computing Project (ECP) as part of the PathForward program.

Austin, Texas, Us

Formerly, the Institute for Computational Engineering and Sciences (ICES), now the Oden Institute for Computational Engineering and Sciences. V&V/UQ model and software development. Predictive Engineering and COmputational Science (PECOS) Center. I was funded in part by the DOE's Predictive Science Academic Alliance Program (PSAAP). Projects: Petascale Direct Numerical Simulations of Turbulent Channel Flow; Model and Software Development for Combustion Physics, Uncertainty Quantification, Bayesian Inference, Optimal Large Eddy Simulations, Software Verification* Lead developer on the open source C++ MASA library, which is an application of the Method of Manufactured Solutions (MMS) to code verification of partial differential equations. See: https://github.com/manufactured-solutions/MASALecturer at the 'Computational Physics Student Summer Workshop' at Los Alamos National Laboratory (2011-2017).* Developer on QUESO, a C++ library for doing uncertainty quantification. QUESO stands for Quantification of Uncertainty for Estimation, Simulation and Optimization. See: https://github.com/libqueso/queso* Developer on CANTERA, a suite of object-oriented software tools for problems involving chemical kinetics, thermodynamics, and/or transport processes.* Software and implementation support for the largest spectral DNS simulation ever conducted, using over 225 Million compute hours on ALCF's BG/Q Mira. This has generated approximately a PB of data. More information is available from a press release from Argonne National Laboratory: http://www.alcf.anl.gov/articles/first-mira-runs-break-new-ground-turbulence-simulations

Austin, Tx, Us

My research was directed at the development of pseudo-spectral methods for massively parallel computation, in particular algorithms for wall bounded turbulent flows that scale to Petascale/Exascale systems.I also published research in computer modeling, in particular, Optimal Large Eddy Simulations.Advisor: Professor Robert D. Moser

Gaithersburg, Md, Us

I received a competitive summer research fellowship, to conduct research in the Physics Laboratory, Optical Technology Division, Optical Thermometry and Spectral Methods Group at the National Institute of Standards and Technology (NIST). During the summer I developed methods to experimentally determine and characterize the dynamic range of InGaAs photodetectors, for use in satellite optics.Advisor: Dr. Howard W. Yoon

Doctor Of Philosophy (Phd), Computational Science

M.S.E., Mechanical Engineering, Computational Fluid Dynamics

B.S., Physics And Mathematics (Double Major)

Education record