Search for: All records

Automated code generation and performance enhancements for sparse tensor algebra have become essential in many real-world applications, such as quantum computing, physical simulations, computational chemistry, and machine learning. General sparse tensor algebra compilers are not always versatile enough to generate asymptotically optimal code for sparse tensor contractions. This paper shows how to generate asymptotically better schedules for complex sparse tensor expressions using kernel fission and fusion. We present generalized loop restructuring transformations to reduce asymptotic time complexity and memory footprint. Furthermore, we present an auto-scheduler that uses a partially ordered set (poset)-based cost model that uses both time and auxiliary memory complexities to prune the search space of schedules. In addition, we highlight the use of Satisfiability Module Theory (SMT) solvers in sparse auto-schedulers to approximate the Pareto frontier of better schedules to the smallest number of possible schedules, with user-defined constraints available at compile-time. Finally, we show that our auto-scheduler can select better-performing schedules and generate code for them. Our results show that the auto-scheduler provided schedules achieve orders-of-magnitude speedup compared to the code generated by the Tensor Algebra Compiler (TACO) for several computations on different real-world tensors. 

Shorebird reproductive success monitoring often relies on surveys of nest and brood survival. However, conclusions may be inaccurate due to the challenges of gathering and interpreting evidence of nest and brood fate. We tested the efficacy of in-person versus camera- based monitoring to quantify productivity and evaluate threats to reproductive success of American Oystercatchers (Haematopus palliatus) and Piping Plovers (Charadrius melodus) at Metompkin Island, Virginia. We deployed 73 cameras using three set-ups: at nests, at brood sites, and along a transect. The same areas were also surveyed in-person approximately once per week. Camera monitoring confirmed nest fate where in-person monitors could not determine fate from field evidence and provided insight to the effectiveness of mammalian predator removal. However, cameras failed to capture causes of mortality for mobile chicks and did not consistently document chicks where in-person monitoring confirmed successful broods. Cameras produced large quantities of data requiring 63.5–315 hours to review, depending on camera set- up. We found cameras were useful for validating conclusions from in-person monitoring, highlighting threats that surveys missed, and characterizing the predator community. Managers should consider the tradeoff between potential benefits and required effort of camera monitoring when deciding which method would be effective for meeting management goals.