Search for: All records

With the ever-growing popularity of Graph Neural Networks (GNNs), efficient GNN inference is gaining tremendous attention. Field-Programmable Gate Arrays (FPGAs) are a promising execution platform due to their fine-grained parallelism, low power consumption, reconfigurability, and concurrent execution. Even better, High-Level Synthesis (HLS) tools help bridge the gap between the non-trivial FPGA development efforts and rapid emergence of new GNN models. To enable investigation into how effectively modern HLS tools can accelerate GNN inference, we present GNNHLS, a benchmark suite containing a software stack for data generation and baseline deployment and FPGA implementations of 6 well-tuned GNN HLS kernels. 

We study settings where a set of identical, reusable resources must be allocated in an online fashion to arriving agents. Each arriving agent is patient and willing to wait for some period of time to be matched. When matched, each agent occupies a resource for a certain amount of time, and then releases it, gaining some utility from having done so. The goal of the system designer is to maximize overall utility given some prior knowledge of the distribution of arriving agents. We are particularly interested in settings where demand for the resources far outstrips supply, as is typical in the provision of social services, for example homelessness resources. We formulate this problem as online bipartite matching with reusable resources and patient agents. We develop new, efficient nonmyopic algorithms for this class of problems, and compare their performance with that of greedy algorithms in a variety of simulated settings, as well as in a setting calibrated to real-world data on household demand for homelessness services. We find substantial overall welfare benefits to using our nonmyopic algorithms, particularly in more extreme settings – those where agents are unwilling or unable to wait for resources, and where the ratio of resource demand to supply is particularly high. 

Abstract We confirm the planetary nature of two gas giants discovered by the Transiting Exoplanet Survey Satellite to transit M dwarfs. TOI-3714 ( V = 15.24, J = 11.74) is an M2 dwarf hosting a hot Jupiter ( M p = 0.70 ± 0.03 M J and R p = 1.01 ± 0.03 R J ) on an orbital period of 2.154849 ± 0.000001 days with a resolved white dwarf companion. TOI-3629 ( V = 14.63, J = 11.42) is an M1 dwarf hosting a hot Jupiter ( M p = 0.26 ± 0.02 M J and R p =0.74 ± 0.02 R J ) on an orbital period of 3.936551 − 0.000006 + 0.000005 days. We characterize each transiting companion using a combination of ground-based and space-based photometry, speckle imaging, and high-precision velocimetry from the Habitable-zone Planet Finder and the NEID spectrographs. With the discovery of these two systems, there are now nine M dwarfs known to host transiting hot Jupiters. Among this population, TOI-3714 b ( T eq = 750 ± 20 K and TSM = 98 ± 7) and TOI-3629 b ( T eq = 690 ± 20 K and TSM = 80 ± 9) are warm gas giants amenable to additional characterization with transmission spectroscopy to probe atmospheric chemistry and, for TOI-3714, obliquity measurements to probe formation scenarios.