Engineering topics
Green, Roland
Publications and source records attributed to Green, Roland.
Balar: A SST GPU Component for Performance Modeling and Profiling
Programmable accelerators have become commonplace in modern computing systems. Advances in programming models and the availability of massive amounts of data have created a space for massively parallel accelerators capable of maintaining context for thousands of concurrent threads resident on-chip. These threads are grouped and interleaved on a cycle-by-cycle basis among several massively parallel computing cores. One path for the design of future supercomputers relies on an ability to model the performance of these massively parallel cores at scale. The SST framework has been proven to scale up to run simulations containing tens of thousands of nodes. A previous report described the initial integration of the open-source, execution-driven GPU simulator, GPGPU-Sim, into the SST framework. This report discusses the results of the integration and how to use the new GPU component in SST. It also provides examples of what it can be used to analyze and a correlation study showing how closely the execution matches that of a Nvidia V100 GPU when running kernels and mini-apps.
SST-GPU: An Execution -Driven CUDA Kernel Scheduler and Streaming-Multiprocessor Compute Model
Programmable accelerators have become commonplace in modern computing systems. Advances in programming models and the availability of massive amounts of data have created a space for massively parallel acceleration where the context for thousands of concurrent threads are resident on-chip. These threads are grouped and interleaved on a cycle-by-cycle basis among several massively parallel computing cores. The design of future supercomputers relies on an ability to model the performance of these massively parallel cores at scale. To address the need for a scalable, decentralized GPU model that can model large GPUs, chiplet-based GPUs and multi-node GPUs, this report details the first steps in integrating the open-source, execution driven GPGPU-Sim into the SST framework. The first stage of this project, creates two elements: a kernel scheduler SST element accepts work from SST CPU models and schedules it to an SM-collection element that performs cycle-by-cycle timing using SSTs Mem Hierarchy to model a flexible memory system.
Methods for the synthesis of arrays of DNA probes
The synthesis of arrays of DNA probes sequences, polypeptides, and the like is carried out using a patterning process on an active surface of a substrate. An image is projected onto the active surface of the substrate utilizing reflective projection optics. The projection optics project a light image onto the active surface of the substrate to deprotect linker molecules thereon. A first level of bases may then be applied to the substrate, followed by development steps, and subsequent exposure of the substrate utilizing a different light image, with further repeats until the elements of a two dimensional array on the substrate surface have an appropriate base bound thereto.
CUBIC - A non-dispersive Diffuse X-ray Background spectrometer
The Cosmic Unresolved X-ray Background Instrument using CCDs (CUBIC) is designed to obtain spectral observations of the Diffuse X-ray Background (DXRB) with moderate spectral resolution over the energy range 0.2-10 keV, using mechanically-collimated CCDs. At this time, it is the only planned satellite payload devoted to the study of the spectrum of the DXRB. Over the anticipated 3 year lifetime of the satellite, CUBIC will be able to study up to 50 percent of the sky with 5 x 5 deg spatial resolution for the subkilovolt Galactic diffuse background, and with 10 x 10 deg spatial resolution for the extragalactic diffuse background above 2 keV. CUBIC will obtain high quality nondispersive spectra of soft X-ray emission from the interstellar medium, supernova remnants, and some bright sources, and will make a sensitive seach for line emission or other features in the extragalactic cosmic X-ray background from 2-10 keV.