Computational scientist with extensive experience in all aspects of operational high-performance computing. Formal education in materials science with a specialty in molecular dynamics and implementing memory-efficient algorithms. Extensive, hands-on experience in system administration, application support, debugging, performance diagnostics, parallel application development, system and user documentation, and education, outreach, and training. Also involved in workload characterization at job, system, and grid scales to provide analytical insight into future systems design to meet the current and future needs of users.
Fundamentally driven by philosophy that HPC must be usable by users to be useful. Active participant in user engagement, classroom and written training, community outreach, and collaboration with the global HPC user, vendor, and operations communities.
- Corning Foundation Fellowship (2011)
- US DOE Advanced Fuel Cycle Initiative Fellow (2008-2010)
- Innovations in Fuel Cycle Research Award (2010)
- Malcolm G. McLaren Scholarship (2005)
- Edward J. Bloustein Distinguished Scholar (2003 - 2007)
- Evaluation and co-design of burst buffer implementations
- Ongoing workload analysis of production parallel file systems
- Configured, integrated, and managed all R&D cyberinfrastructure (1200-core cluster, 700 TB of NetApp; Illumina HiSeq/NextSeq/MiSeq)
- Designed, implemented, and managed a custom continuous integration system for software group
- Designed, implemented, and maintained middleware to integrate R&D CI operations with Amazon EC2 and S3
- Provided full spectrum support for XSEDE systems (26K cores total, 1,000 unique users anually) including software integration, performance tuning, and troubleshooting node-, cluster-, and grid-level problems
- Technical consultant for strategic collaborations with industry (Johnson & Johnson, Human Longevity Inc), academia (UC San Diego, Scripps Translational Health Inst, J. Craig Venter Inst), and non-profit (San Diego Zoo)
- Technical lead for deploying Hadoop 2 and Spark on SDSC's XSEDE resources; author of myHadoop 0.2 which dramatically reduced entry barrier for Hadoop/Spark users; taught various R and Hadoop workshops and seminars
- Contributing author, editor, and key personnel on several NSF proposals including awards ACI-1341698 (deployment and operation of Comet, a 2 PF HPC; $21.6M) and IIS-1447861 (scalable middleware for integrating Hadoop/Spark on HPCs; $360K)
- Implemented a variety of novel force fields in both proprietary and open-source molecular dynamics packages
- Implemented memory-efficient algorithms to improve cache hit rate, improving performance by 50% to 200%
- Implemented efficient shared-memory parallelization of MD kernels to improve hardware utilization efficiency and reduce time to solution by up to 8x
- Procured, configured, and managed three 8- to 12-node lab clusters
Invited presentation at the Joint Facilities User Forum on Data-Intensive Computing, sponsored by US Department of Energy. Oakland, CA.
Invited panelist at Leverage BIGDATA'14, sponsored by HPCwire/Tabor Communications. Carlsbad, CA.
Presented at Are you ready for the HPC Storage and Big Data deluge?, sponsored by Dell. La Jolla, CA.
Presented in the SDSC Industrial Partnership Program webinar series, sponsored by SDSC. La Jolla, CA.
Presented at the 36th International Conference and Exposition on Advanced Ceramics and Composites (ICACC). Daytona Beach, FL.
Presented at the 2011 Materials Research Society Fall Meeting & Exhibit. Boston, MA.
Presented at the Materials Science & Technology 2011 Conference & Exhibition (MS&T'11). Columbus, OH.
Poster presented at the 2010 Gordon Research Conference on Water & Aqueous Solutions. Plymouth, NH.
Poster presented at the Synthesis and Processing Sciences Contractors Meeting, sponsored by US Department of Energy. Warrenton, VA.
G. K. Lockwood, R. Wagner, and M. Tatineni. Proceedings of the 2015 Annual Conference on Extreme Science and Engineering Discovery Environment - XSEDE'15 (2015).
G. K. Lockwood and S. H. Garofalini. J. Phys. Chem. C 118, 29750 (2014).
G. K. Lockwood, M. Tatineni, and R. Wagner. Proceedings of the 2014 Annual Conference on Extreme Science and Engineering Discovery Environment - XSEDE'14 (2014).
D. J. Choi, G. K. Lockwood, R. S. Sinkovits, M. Tatineni. Proceedings of the 2014 Annual Conference on Extreme Science and Engineering Discovery Environment - XSEDE'14 (2014).
J. A. Tracey, J. K. Sheppard, G. K. Lockwood, A. Chourasia, M. Tatineni, R. N. Fisher, and R. S. Sinkovits. Proceedings of the 2014 Annual Conference on Extreme Science and Engineering Discovery Environment - XSEDE'14 (2014).
M. Kagan, G. K. Lockwood and S. H. Garofalini. Phys. Chem. Chem. Phys. 16, 9294 (2014).
G. K. Lockwood and S. H. Garofalini. J. Phys. Chem. B 117, 4089 (2013).
G. K. Lockwood and S. H. Garofalini. J. Nucl. Mater. 430, 239 (2012).
Y. Ma, G. K. Lockwood and S. H. Garofalini. J. Phys. Chem. C 115, 24198 (2011).
G. K. Lockwood and S. H. Garofalini. J. Nucl. Mater 400, 73 (2010).
G. K. Lockwood and S. H. Garofalini. J. Chem. Phys. 131, 074703 (2009).
G. K. Lockwood, S. Zhang, and S. H. Garofalini. J. Am. Ceram. Soc 91, 3536 (2008).
One of three instructors for a two-day, hands-on workshop illustrating the essential concepts of Hadoop provided through UCSD Extension, UCSD's continuing education program. In addition to teaching, coordinated online logistics of course material.
Volunteered to present the 65-minute "Parallel Options for R / Introduction to Hadoop" talk (see below) to the San Diego R User Group as an outreach event on behalf of SDSC. Orchestrated logistics of hosting user group meetings at SDSC facilities.
Provided a 45-minute presentation on parallelization libraries (via MPI and Hadoop) for the R language at SDSC's 2013 Summer Institute. Supplemented with a 25-minute conceptual overview of Hadoop based on earlier attendee feedback.
Contributed slides for a half-day tutorial on using Hadoop on traditional supercomputers and provided individual support to attendees.
Local facilitator for videconferenced course; provided individual support to attendees.
Instructed a three-credit, core curriculum graduate course for fifteen students at Picatinny Arsenal (U.S. Army) on behalf of Rutgers University. All Ph.D. students who took this course went on to qualify for candidacy.