Thermal balancing policy for streaming computing on multiprocessor architectures

Abstract

As feature sizes decrease, power dissipation and heat generation density exponentially increase. Thus, temperature gradients in Multiprocessor Systems on Chip (MPSoCs) can seriously impact system performance and reliability. Thermal balancing policies based on task migration have been proposed to modulate power distribution between processing cores to achieve temperature flattening. However, in the context of MPSoC for multimedia streaming computing, where timeliness is critical, the impact of migration on quality of service must be carefully analyzed. In this paper we present the design and implementation of a lightweight thermal balancing policy that reduces on-chip temperature gradients via task migration. This policy exploits run-time temperature and load information to balance the chip temperature. Moreover, we assess the effectiveness of the proposed policy for streaming computing architectures using a cycle-accurate thermal-aware emulation infrastructure. Our results using a real-life software defined radio multitask benchmark show that our policy achieves thermal balancing while keeping migration costs bounded.

Thermal-aware task scheduling in 3D chip multiprocessor with real-time constrained workloads

“…Several temperature-aware algorithms are present in Coskun et al [2008] Mulas et al [2008, Lin et al [2009], Ayoub and Rosing [2009], Zhu et al [2008], Zhou et al [2010], and recently. The Adapt3D approach in Coskun et al [2009] assigns the upcoming job to the coolest core to achieve thermal balance.…”

Dynamic thermal management by greedy scheduling algorithm

“…[24], the thread with less task size tends to be migrated easier than other threads to reduce the performance overhead caused by migration, but the task size does not necessarily mean the memory usage, especially the cache usage in the core. Moreover, the difference of migration cost between tasks with distinct memory usages is only within 10 ms, and migration cost mainly comes from the task suspension and resumption.…”

Exploiting power budgeting in thermal-aware dynamic placement for reconfigurable systems

“…Online DTM techniques apply several mechanisms (voltage/frequency scaling [2] or task migration [3,4]) to reduce temperature. Reactive systems in general require the use of thermal sensors and actuators to detect a thermal emergency and activate the appropriate thermal management scheme, which can in turn increase the complexity of the design.…”

“…In order to cope with the increasing demand for higher computation capacities in reconfigurable systems, several dynamic thermal management (DTM) schemes have been proposed [2,3,4]. In general, DTM techniques try to buy in more power saving by compromising the computation speed.…”