300 Hahn Hall
Blacksburg, VA, 24060
(515) 337-5362
EDUCATION
Ph.D., Physical Chemistry Iowa State University Nov 2012
B.S., Computer Science Valdosta State University May 2005
PRESENTATIONS AND PUBLICATIONS
* A Fast and Flexible Coupled Cluster Approach
A.Asadchev, M.S.Gordon, Journal of Chemical Theory and Computation 2013
* New Multithreaded Hybrid CPU/GPU Approach to Hartree–Fock
A.Asadchev, M.S.Gordon, Journal of Chemical Theory and Computation 2012 8 (11)
* Uncontracted Rys Quadrature Implementation of up to G Functions on GPU - Graphical Processing Units
A.Asadchev, V.Allada, J.Felder, B.M.Bode, M.S.Gordon, T.L.Windus, Journal of Chemical Theory and Computation 2010 6(3)
* Accelerating Quantum Chemistry Research using GPUs
A.Asadchev, J.Felder, GPU Technology Conference, NVIDIA, San Jose, 2009
* Performance of Electronic Structure Calculations on BG/L and XT4 Computers
A.Asadchev, B.M.Bode, M.S.Gordon, Journal of Computational and Theoretical Nanoscience, 2009 6(6)
* Uncontracted Rys Quadrature on GPU
A.Asadchev, Path to Petascale, NCSA, 2009
* Performance of Electronic Structure Calculations on Blue Gene/L and Cray XT4
A.Asadchev, Poster, IEEE/ACM Supercomputing 2008
RESEARCH EXPERIENCE
VIRGINIA TECH 2012 – PRESENT
PostDoctoral Research with Dr. Eduard Valeev
* Massively Parallel Quantum Chemistry (MPQC) package: porting to CMake build system, adding functionality for distributed arrays in memory or on the disk.
* Configuration Interaction: fully distributed CI algorithm, scalable to large number of cores.
* Two-electron integrals: an ongoing research to optimize performance of various kernels on SSE and AVX capable processors.
AMES LAB - IOWA STATE UNIVERSITY 2005 – 2012
Research Associate under Dr. Mark Gordon
* Porting of Quantum Chemistry GAMESS package to Blue Gene and Cray XT: added ARMCI remote memory runtime, parallelized linear algebra, and various other optimizations.
* Rewrite of the Rys Quadrature integral program: Implemented in C++ with optional SSE intrinsics, uses computer-generated code and C++ templates to generate fast code. Showed significant improvement over the legacy implementation.
* Parallel multi-threaded Hartree-Fock: new algorithm implementation based on the new Rys Quadrature code, showed significant speed improvement over the original.
* Port of the C++ Rys Quadrature and Hartree-Fock algorithms to GPU: implemented using CUDA C++, modified parts of Boost C++ libraries to compile with NVCC. Improvements on the order of 12-17 times.
* Scalable perturbation theory implementation: new implementation in C++ with a very low memory footprint, transparent distributed memory/HDF5 data storage backend, tensor manipulation tools integrated with BLAS math libraries. The memory and speed improvement were on various orders of magnitude over the ori...
Login or Register to view the full resume.
