Alexandre E. Eichenberger
contact information
Compiler optimization for AI, SIMD & thread parallelismThomas J. Watson Research Center, Yorktown Heights, NY USA +1
914
945
1812



links
profile
Design and development of high-performance computing systems.
My research focus is on the interaction of compiler technology and microarchitecture design. Interests include optimization for Artificial Intelligence (AI) of Deep Neural Network (DNN) for CPU and hardware accelerators using compiler optimizations, automatic/assisted parallelization using OpenMP for hybrid platforms (GPUs and shared-memory nodes), polyhedral framework, automatic SIMD code generation, multi-threading for performance and reliability, scheduling algorithms for high performance and low register requirements, support for speculative and predicated operations, memory system optimization, and exploration of highly concurrent microarchitecture designs.