Training large AI models on numerous GPUs consumes a massive amount of energy,
making power delivery one of the largest limiting factors in building and operating …

Operating server components beyond their voltage and power design limit (ie, overclocking)
enables improving performance and lowering cost for cloud workloads. However …

Meta maintains additional capacity within its infrastructure to ensure high availability for
business workloads, accommodating user growth, temporal traffic variations, and …

Cloud platforms commonly exploit workload temporal flexibility to reduce their carbon
emissions. They suspend/resume workload execution for when and where the energy is …

Cloud platforms' rapid growth raises significant concerns about their electricity consumption
and resulting carbon emissions. Power capping is a known technique for limiting the power …

Conclusion In this paper, we propose a Fluctuation-Aware Power Auction Strategy (FLAPS),
which aims to reduce power usage overload. FLAPS identifies previously overlooked job …

Power throttling with dynamic voltage and frequency scaling (DVFS) is important to the
dynamic thermal management (DTM) of processors. With increasing numbers of on-chip …

The hardware-software abstraction enables programmers to write high-level algorithms
without delving into low-level microarchitectural details. Compilers, positioned at the …

The ever-increasing demands of data-intensive applications and the rapid evolution of
computer architectures have posed significant challenges in memory performance and …