This presentation was one of 13 presentations in the one-day forum, "Excellence in Computer Simulation," which brought together a broad set of experts to reflect on the future of computational science and engineering.
Novel nanoelectronic devices such as quantum dots, nanowires, and ultra-scaled quantum wells are expected to significantly enhance existing nanoelectronic technologies. The behavior of carriers and their interaction with their environment need to be fundamentally explained at a quantum mechanical level. Modeling efforts that are targeted to enhance the theoretical understanding of these devices are underway worldwide. Most of these device level descriptions utilize an effective mass approach. However, the concepts of device and material meet at the nanometer scale. The new device is really a new material and vice versa. A representation of the constituent materials at the atomic resolution is needed to quantitatively model devices with a countable number of atoms. While atomistic representations are novel to device physicists, the concept of finite devices that are not infinitely periodic is novel in the semiconductor materials modeling community.
This presentation provides a two-fold overview of lessons-learned: 1) 1-D high bias transport simulations that matched and predicted experimental data in resonant tunneling diodes (NEMO 1D), and 2) 3-D multimillion atom simulations that electronic structure in disordered Si/SiGe quantum wells (NEMO 3D). NEMO 1D has been recently shown to scale to 23,000 processors and NEMO 3D to 8,192 on ORNL Jaguar. The two code concepts are now being combined towards peta-scale computing.
To truly have impact on the research, experimental, and educational efforts of the community, relevant tools must be put into the hands of experimentalists and educators. NEMO 3-D offers the opportunity to engage both educators and advanced researchers, utilizing a single code. An educational version of NEMO 3-D has been released on the nanoHUB and user-friendly versions that enable large-scale parallel executions are under development. Over 567 users ran over 4,300 simulations using the educational NEMO 3-D version embodied in the nanoHUB quantum dot lab in the past 12 months. Over 26,100 users interacted with the nanoHUB in the 12 months leading to September. 2007. Over 5,900 users have launched over 226,000 simulations in that time frame.
Gerhard Klimeck is the Technical Director of the Network for Computational Nanotechnology at Purdue University and a Professor of Electrical and Computer Engineering since Dec. 2003. He was the Technical Group Supervisor of the High Performance Computing Group and continues to hold his Principal member position at the NASA Jet Propulsion Laboratory on a part time faculty basis. His research interest is in the modeling of nanoelectronic devices, parallel cluster computing, and genetic algorithms. Gerhard led the development of the Nanoelectronic Modeling tool (NEMO 3-D) for multimillion atom simulations which has demonstrated parallel scaling to 8,192 CPUs. Previously he was a member of technical staff at the Central Research Lab of Texas Instruments where he served as manager and principal architect of the Nanoelectronic Modeling (NEMO 1-D) program, which has demonstrated to scale to 23,000 CPUs. Dr. Klimeck received his Ph.D. in 1994 from Purdue University and his German electrical engineering degree (summa cum laude) in 1990 from Ruhr-University Bochum. Dr. Klimeck's work is documented in over 170 peer-reviewed publications and over 290 conference presentations. He is a senior member of IEEE and member of APS, HKN and TBP.
The Center of Integrated Nanomechanical Systems (COINS)
The Molecular Foundry at Lawrence Berkeley National Lab
Bancroft Hotel, University of California at Berkeley, Berkeley, CA