Application slowdowns that accumulate on the critical path significantly impact performance. Identifying them requires visualizing inter-thread wait-for dependencies, a task neglected by existing tools, especially regarding ad-hoc, busy-waiting synchronization. We present Tapestry, which efficiently traces these dependencies by dynamically switching between hardware watchpoints and software breakpoints. Using three use cases, we show that Tapestry is able to reveal NUMA effects and explain performance anomalies in oversubscribed and virtualized environments.

To best support highly parallel applications, Linux's CFS scheduler tends to spread tasks across the machine on task creation and wakeup. It has been observed, however, that in a server environment, such a strategy leads to tasks being unnecessarily placed on long-idle cores that are running at lower frequencies, reducing performance, and to tasks being unnecessarily distributed across sockets, consuming more energy. In this paper, we propose to exploit the principle of core reuse, by constructing a nest of cores to be used in priority for task scheduling, thus obtaining higher frequencies and using fewer sockets. We implement the Nest scheduler in the Linux kernel. While performance and energy usage are comparable to CFS for highly parallel applications, for a range of applications using fewer tasks than cores, Nest improves performance 10%--2× and can reduce energy usage.