Cache partitioning and sharing is critical to the effective utilization of multicore processors. However, almost all existing studies have been evaluated by simulation that often has several limitations, such as excessive time, absence OS activities proneness inaccuracy. To address these issues, we taken an efficient software approach supporting both static dynamic cache in through memory mapping. We comprehensively representative schemes with different optimization objectives, including...

The widespread use of multicore processors has dramatically increased the demand on high memory bandwidth and large capacity. As DRAM subsystem designs stretch to meet demand, power consumption is now approaching that processors. However, conventional architecture prevents any meaningful performance trade-offs for memory-intensive workloads. We propose a novel idea called mini-rank DDRx (DDR/DDR2/DDR3) DRAMs, which uses small bridge chip each DIMM break rank into multiple smaller mini-ranks...

Magnetic flux leakage (MFL) detection robots are broadly employed in acquiring MFL signals to detect pipeline defects. However, influenced by the complex environments, accuracy of defect under is undesirable. To resolve this problem, a cascaded attention feature fusion network (CAFF-Net) proposed article. First, novel enhancement method named bidirectional compression enhance expression Second, multilevel extraction module presented, which consists two cascade networks with different...

Deep learning methods have demonstrated promising performance in magnetic flux leakage (MFL) defect detection under adequate amounts of labeled samples. However, industrial occasions, obtaining samples is time-consuming and expensive, applying only limited can lead to unsatisfactory accuracy. To address the above issues, a method named semisupervised circular teacher network (SSCT-Net) proposed this article. First, parallel feature extraction with hybrid attention SSCT-Net so that useful...

To improve the validity of magnetic flux leakage (MFL) multisensor signals, anomaly detection has become a significant part MFL signal processing. The anomalies in are uncertain and have no prior information or labels. Therefore, location difficult issue. Regarding abovementioned problem, we propose an unsupervised method called cycle-supervised convolutional neural network (CsCNN). CsCNN is built including multiple CNNs with same structure part. proposed model realizes through for first...

The widespread use of multicore processors has dramatically increased the demands on high bandwidth and large capacity from memory systems. In a conventional DDR2/DDR3 DRAM system, bus devices run at same data rate. To improve bandwidth, we propose new system design called decoupled DIMM that allows to operate rate much higher than devices. design, synchronization buffer is added relay between slow fast bus; access scheduling revised avoid conflicts ranks. not only improves beyond what can...

Performance degradation of memory-intensive programs caused by the LRU policy's inability to handle weak-locality data accesses in last level cache is increasingly serious for two reasons. First, last-level remains CPU's critical path, where only simple management mechanisms, such as LRU, can be used, precluding some sophisticated hardware mechanisms address problem. Second, commonly used shared structure multi-core processors has made this path even more performance-sensitive due intensive...

Accurate weld defect detection is a critical signal processing technique for intelligent magnetic flux leakage (MFL) system. However, the MFL signals between and are coupled with each other, features two difficult to distinguish, resulting in unsatisfactory accuracy of detection. To address this issue, novel method named contextual relation network (CRNet) proposed achieve high-accuracy First, based on strong coupling signals, feature extraction sparse attention extract high-quality manner....

Cancer is the second leading cause of death worldwide. Differences in drug resistance and treatment response caused by heterogeneity cancer cells are primary reasons for poor therapy outcomes patients. In addition, current vitro anticancer drug-screening methods rely on two-dimensional monolayer-cultured cells, which cannot accurately predict behavior vivo. Therefore, a powerful tool to study produce effective tumor models warranted leverage research. Droplet microfluidics has become...

With increasing speed and power density, high-performance memories, including FB-DIMM (Fully Buffered DIMM) DDR2 DRAM, now begin to require dynamic thermal management(DTM) as processors hard drives did. The DTM of nevertheless, is different in that it should take the processor performance consumption into consideration. Existing schemes have ignored that. In this study, we investigate a new approach controls memory issues from source generating activities - processor. It will smooth program...

The management of shared caches in multicore processors is a critical and challenging task. Many hardware OS-based methods have been proposed. However, they may be hardly adopted practice due to their non-trivial overheads, high complexities, and/or limited abilities handle increasingly complicated scenarios cache contention caused by many-cores.

Magnetic flux leakage (MFL) testing is an effective method of pipeline defect detection to guarantee the health submarine pipelines. However, due interference strong signals and complex nondefect signals, MFL weak defects a tough issue currently. Focusing on this problem, based two-stage heterogeneous mutual supervision network (THMS-Net) proposed. THMS-Net includes two deep networks for feature extraction axial radial them enhances discrimination between signals. There are stages in...

Safe transmission of oil pipelines is one the guarantees national defense and environmental protection. Magnetic flux leakage (MFL) testing critical to safety inspection in-service pipelines. In detection process, incompleteness MFL data affects defect location inversion severely. This article proposes an recovery method based on multifeature condition risk, which can effectively handle block gap problem. First, a preprocessing mechanism proposed determine boundaries uniformly interpolate...

Defect inversion estimates the defect size quantitatively, which is a significant process in magnetic flux leakage (MFL) detection system. The actual measured defects are irregular due to complex environment, leads signal deformation. Moreover, sufficient and accurate features of hard extract, so it difficult invert MFL precisely. Therefore, an end-to-end method based on heterogeneous multi-class feature fusion (HMFF) proposed quantify pipeline for HMFF extracts abstract external signals,...

Thermal management of DRAM memory has become a critical issue for server systems. We have done, to our best knowledge, the first study software thermal subsystem on real machines. Two recently proposed DTM (Dynamic Management) policies been improved and implemented in Linux OS evaluated two multicore servers, Dell PowerEdge 1950 customized Intel SR1500AL testbed. The experimental results confirm that system-level policy may significantly improve system performance power efficiency, compared...