Search

Memory tiering has received wide adoption in recent years as an effective solution to address the increasing memory demands of memory-intensive workloads. However, existing tiered memory systems often fail to meet service-level objectives (SLOs) when multiple applications share the system because they lack Quality-of-Service (QoS) support. Consequently, applications suffer severe performance drops due to local memory contention and memory bandwidth interference. In this paper, we present Mercury, a QoS-aware tiered memory system that ensures predictable performance for coexisting memory-intensive applications with different SLOs. Mercury enables per-tier page reclamation for application-level resource management and uses a proactive admission control algorithm to satisfy SLOs via per-tier memory capacity allocation and intra- and inter-tier bandwidth interference mitigation. It reacts to dynamic requirement changes via real-time adaptation. Extensive evaluations show that Mercury improves application performance by up to 53.4% and 20.3% compared to TPP and Colloid, respectively.

Deep Learning Recommendation Models (DLRMs) have become increasingly popular and prevalent in today's datacenters, consuming most of the AI inference cycles. The performance of DLRMs is heavily influenced by available bandwidth due to their large vector sizes in embedding tables and concurrent accesses. To achieve substantial improvements over existing solutions, novel approaches towards DLRM optimization are needed, especially, in the context of emerging interconnect technologies like CXL. This study delves into exploring CXL-enabled systems, implementing a process-in-fabric-switch (PIFS) solution to accelerate DLRMs while optimizing their memory and bandwidth scalability. We present an in-depth characterization of industry-scale DLRM workloads running on CXL-ready systems, identifying the predominant bottlenecks in existing CXL systems. We, therefore, propose PIFS-Rec, a PIFS-based scheme that implements near-data processing through downstream ports of the fabric switch. PIFS-Rec achieves a latency that is 3.89x lower than Pond, an industry-standard CXL-based system, and also outperforms BEACON, a state-of-the-art scheme, by 2.03x.

We present a comprehensive study on the effects of proton irradiation at different energies (150 and 590 keV) with the fluence of 1x 1012 proton/cm2 on monolayer MoS2. This study not only improves our understanding of the influence of high-energy proton beams on MoS2 but also has implications for radiation-induced changes in device processing and engineering of devices from multilayer MoS2 starting material. Increasing defect density with decreasing proton irradiation energy was observed from photoluminescence spectroscopy study. These defects are attributed to sulfur vacancies observed through x-ray photoelectron spectroscopy analysis and confirmed by transmission electron microscope imaging. Scanning electron microscopy images showed the creation of grain boundaries after proton irradiation. A higher degree of surface deformation was detected with lower irradiation energies through atomic force microscopy. Inter-defect distance is increased with the increase in proton energy irradiation as estimated by transmission electron microscopy imaging. Raman spectroscopy reveals negligible structural changes in the crystal quality after the irradiation. These deformation damages due to proton irradiation are insignificant at the MoS2 layer. Based on the overall influence of low energy proton irradiation on the material characteristics, ML-MoS2 materials can be considered robust and reliable building blocks for 2D material based devices for space applications., Comment: Figures here are not finalized yet