Simuliris: a separation logic framework for verifying concurrent program optimizations

Today’s compilers employ a variety of non-trivial optimizations to achieve good performance. One key trick compilers use to justify transformations of concurrent programs is to assume that the source program has no data races : if it does, they cause the program to have undefined behavior (UB) and give the compiler free rein. However, verifying correctness of optimizations that exploit this assumption is a non-trivial problem. In particular, prior work either has not proven that such optimizations preserve program termination (particularly non-obvious when considering optimizations that move instructions out of loop bodies), or has treated all synchronization operations as external functions (losing the ability to reorder instructions around them). In this work we present Simuliris , the first simulation technique to establish termination preservation (under a fair scheduler) for a range of concurrent program transformations that exploit UB in the source language. Simuliris is based on the idea of using ownership to reason modularly about the assumptions the compiler makes about programs with well-defined behavior. This brings the benefits of concurrent separation logics to the space of verifying program transformations: we can combine powerful reasoning techniques such as framing and coinduction to perform thread-local proofs of non-trivial concurrent program optimizations. Simuliris is built on a (non-step-indexed) variant of the Coq-based Iris framework, and is thus not tied to a particular language. In addition to demonstrating the effectiveness of Simuliris on standard compiler optimizations involving data race UB, we also instantiate it with Jung et al.’s Stacked Borrows semantics for Rust and generalize their proofs of interesting type-based aliasing optimizations to account for concurrency.

Characteristics associated with COVID-19 vaccination status among staff and faculty of a large, diverse University in Los Angeles.

Objective: This study examined characteristics associated with being unvaccinated among a sample of university staff and faculty prior to university campus reopening for in-person learning in spring-summer 2021. Methods: Staff and faculty responded to an email invitation to complete an online survey. Survey questions included demographic data (race/ethnicity, age, sex), COVID-19 knowledge and behaviors, employment specific data including division and subdivision (healthcare vs. non-healthcare related division); and self-reported vaccination status. A multivariable logistic regression analysis was performed to determine significant characteristics associated with the likelihood of being unvaccinated for COVID-19. Results: Participants identifying as Asian and Asian American, Hispanic/Latinx or Multicultural/Other had greater odds of being unvaccinated compared to Non-Hispanic White participants. Other characteristics associated with greater likelihood of being unvaccinated included working as university staff member (vs. faculty), older age, decrease in income, inability to work remotely and not traveling outside of Los Angeles area. Political affiliation as an Independent or as something else were more likely to be unvaccinated compared to participants identifying as Democrat. Conclusions: Findings suggest several factors associated with racial and social disparities may delay the uptake of COVID-19 vaccination. This study highlights the need for targeted educational interventions to promote vaccination among university staff and faculty.

Analysis of Racing Greyhound Path Following Dynamics Using a Tracking System

The University of Technology Sydney (UTS) has been working closely with the Australasian greyhound industry for more than 5 years to reduce greyhound race-related injuries. During this period, UTS has developed and deployed several different techniques including inertial measurement units, drones, high-frame-rate cameras, track geometric surveys, paw print analysis, track soil spring-force analysis, track maintenance data, race injury data, race computer simulation and modelling to assist in this task. During the period where the UTS recommendations have been adopted, the injury rate has dropped significantly. This has been achieved by animal welfare interventions that lower racing congestion, and lower transient forces and jerk rates the greyhounds experience during a race. This study investigated the use of a greyhound location tracing system where small and lightweight signal emitting devices were placed inside a pocket in the jackets of racing greyhounds. The system deployed an enhanced version of a player tracking system currently used to track the motion of human athletes. Greyhounds gallop at speeds of almost 20 m/s and are known to change their heading direction to exceed a yaw rate of 0.4 rad/s. The high magnitudes of velocity, acceleration and jerk posed significant technical challenges, as the greyhounds pushed the human tracking system beyond its original design limits. Clean race data gathered over a six-month period were analysed and presented for a typical 2-turn greyhound racing track. The data confirmed that on average, greyhounds ran along a path that resulted in the least energy wastage, which includes smooth non-linear paths that resemble easement curves at the transition between the straights to the semi-circular bends. This study also verified that the maximum jerk levels greyhounds experienced while racing were lower than the jerk levels that had been predicted with simulations and modelling for the track path. Furthermore, the results from this study show the possibility of such a systems deployment in data gathering in similar settings to greyhound racing such as thoroughbred and harness horse racing for understanding biomechanical kinematic performance.

172 Sleep Disparities in Preschool-Aged Children

Introduction Sleep is known to be associated with socioeconomic status (SES) in older children and adults with those from lower SES households often experiencing poorer sleep quality. Whether this disparity exists in early childhood is relatively unknown, despite being an important age marked by sleep transitions and the establishment of lifelong sleep habits. Furthermore, it is a critical period for cognitive development and learning, which are supported by sleep. Here, we explore associations between sleep and SES in a preschool population. We hypothesized that children from lower SES households would exhibit shorter overnight sleep, longer and more frequent naps, and shorter 24-hr sleep. Additionally, we considered racial and ethnic disparities in sleep which can be confounded with SES in some samples. Methods Child (n=441; M age=51.9mo; 45.4% female) sleep was measured objectively using actigraph watches, worn for 3-16 days (M=9.5 days). Caregivers reported child demographics and household data. Race/ethnicity of our sample was 72% White, 10.2% Black, 17.8% other or more than one race, and 28.4% identified as Hispanic. 20.1% of our sample was categorized as low SES. Effects of SES and race/ethnicity on continuous sleep measures were assessed using multiple regression models, with age and gender as covariates. Nap habituality was assessed using chi-square tests. Results Lower SES was associated with shorter nighttime sleep duration, longer nap duration, and shorter 24-hr sleep duration (p’s<.001). Children from lower SES households were also more likely to nap habitually (p=.04) as were Hispanic children (p<.001). Hispanic children also tended to have longer nap bouts (p=.002). Hispanic and Black children on average had shorter overnight sleep durations than White children (p’s<.04), but their 24-hr sleep did not differ. Conclusion SES-related sleep disparities were present in this preschool population, with lower SES children exhibiting poorer sleep. When controlling for SES, Hispanic children tended to sleep less overnight which was compensated for by longer, more frequent naps. This underscores the necessity of naps for some children to achieve adequate sleep. Future directions will explore the relationship between parenting factors and sleep, such as bedtime routines and parent knowledge surrounding child sleep needs. Support (if any) NIH R01 HL111695