Persistent Memory Research Articles

Localized Validation Accelerates Distributed Transactions on Disaggregated Persistent Memory

Persistent memory (PM) disaggregation significantly improves the resource utilization and failure isolation to build a scalable and cost-effective remote memory pool in modern data centers. However, due to offering limited computing power and overlooking the bandwidth and persistence properties of real PMs, existing distributed transaction schemes, which are designed for legacy DRAM-based monolithic servers, fail to efficiently work on the disaggregated PM. In this article, we propose FORD, a F ast O ne-sided R DMA-based D istributed transaction system for the new disaggregated PM architecture. FORD thoroughly leverages one-sided remote direct memory access to handle transactions for bypassing the remote CPU in the PM pool. To reduce the round trips, FORD batches the read and lock operations into one request to eliminate extra locking and validations for the read-write data. To accelerate the transaction commit, FORD updates all remote replicas in a single round trip with parallel undo logging and data visibility control. Moreover, considering the limited PM bandwidth, FORD enables the backup replicas to be read to alleviate the load on the primary replicas, thus improving the throughput. To efficiently guarantee the remote data persistency in the PM pool, FORD selectively flushes data to the backup replicas to mitigate the network overheads. Nevertheless, the original FORD wastes some validation round trips if the read-only data are not modified by other transactions. Hence, we further propose a localized validation scheme to transfer the validation operations for the read-only data from remote to local as much as possible to reduce the round trips. Experimental results demonstrate that FORD significantly improves the transaction throughput by up to 3× and decreases the latency by up to 87.4% compared with state-of-the-art systems.

Gemini nanoparticles-based quadruple therapy (GNQT) achieved effective tumor immunotherapy by comprehensive regulation of tumor microenvironment

Even though immune checkpoint blockade (ICB) therapy has advanced cancer immunotherapy greatly due to its quick development, and scientists have also tried a variety of combination therapies to further amplify the therapeutic impact, the therapeutic outcome is still limited because of the complex immunosuppressive microenvironment of tumor tissue. To improve the antitumor effect of immunotherapy, developing more comprehensive and effective tumor microenvironment regulation strategy is necessary. In this work, we prepared Gemini NPs composed of drug-loaded nanoparticles DI NPs (Doxorubicin and Ibrutinib were simultaneously encapsulated by PLG-g-mPEG) and gene-loaded nanoparticles PPD NPs (pSpam1 and pshPD-L1 were simultaneously encapsulated by PEI1.8k-RT), and proposed a Gemini nanoparticles-based quadruple therapy (GNQT). Compared with triple therapy and four-drug combination therapy, GNQT demonstrated comprehensive modulation of the tumor microenvironment and had exceptional anticancer ability in melanoma model without obvious toxicity. In addition, by establishing a persistent immunological memory effect, GNQT could also successfully prevent tumor lung metastasis in mice. The GNQT proposed here provided new insights into tumor immunotherapy and had important implications for the research of preclinical antitumor immunotherapy.

Memento: A Framework for Detectable Recoverability in Persistent Memory

Persistent memory (PM) is an emerging class of storage technology that combines the performance of DRAM with the durability of SSD, offering the best of both worlds. This had led to a surge of research on persistent objects in PM. Among such persistent objects, concurrent data structures (DSs) are particularly interesting thanks to their performance and scalability. One of the most widely used correctness criteria for persistent concurrent DSs is detectable recoverability , ensuring both thread safety (for correctness in non-crashing concurrent executions) and crash consistency (for correctness in crashing executions). However, the existing approaches to designing detectably recoverable concurrent DSs are either limited to simple algorithms or suffer from high runtime overheads. We present Memento: a general and high-performance programming framework for detectably recoverable concurrent DSs in PM. To ensure general applicability to various DSs, Memento supports primitive operations such as checkpoint and compare-and-swap and their composition with control constructs. To ensure high performance, Memento employs a timestamp-based recovery strategy that requires fewer writes and flushes to PM than the existing approaches. We formally prove that Memento ensures detectable recoverability in the presence of crashes. To showcase Memento, we implement a lock-free stack, list, queue, and hash table, and a combining queue that detectably recovers from random crashes in stress tests and performs comparably to existing hand-tuned persistent DSs with and without detectable recoverability.

Reducing the stickiness of negative memory retrieval through positive memory training in adolescents

Background and objectivesIndividuals at risk for depression exhibit a decreased ability to disengage from negative memory retrieval during times of mental distress, partly because they have difficulty retrieving positive memories to repair sad mood. In this study, we tested whether this persistent tendency for negative memory retrieval could be reduced in adolescents through repeated practice to retrieve positive autobiographical memories, namely Positive Memory Specificity Training (PMST). Further, we examined the impact of this intervention on secondary outcomes, including depressive symptoms, emotion regulation strategies, and fear of positive emotions. MethodsAdolescents (n = 68) between 16 and 18 years old were randomly allocated to either PMST or bogus control training. Persistent negative memory retrieval was assessed following the training using a behavioral decision-making task (Emotional Reversal Learning Task). Additionally, participants completed self-report measurements (e.g., depressive symptoms) before and two weeks after the training. ResultsWe found preliminary supportive evidence for a significant training effect such that adolescents following PMST showed less persistence in negative memory retrieval compared to those in the control group. Only for anhedonia a significant training effect was found, indicating a possible adverse effect of the intervention. LimitationsThe primary outcome was assessed only at post-intervention to prevent a potential learning effect due to repeated measurements. We cannot exclude the possibility that baseline individual differences contaminated our results. To examine possible adverse effects of PMST, larger sample are needed. ConclusionsPMST may help to reduce persistent negative memory retrieval in adolescents. Recommendations for future studies are addressed.

Cooperative Buffer Management With Fine-Grained Data Migrations for Hybrid Memory Systems

Hybrid memory composed of DRAM and persistent memory (PM) offers a promising way to realize large-capacity main memory supporting in-memory data storage and computing. However, traditional buffer management schemes focus on improving the hit ratio but lack awareness of the limitations of PM, e.g., slower write time and lower write endurance than DRAM. Therefore, developing new buffer management policies that can reduce costly write-backs of PM blocks while maintaining high performance for the hybrid buffer, is of paramount importance. Existing approaches mainly use a page-grained buffering policy, which will cause unnecessary data migrations between DRAM and PM, leading to a high number of disk I/Os and PM writes. Aiming to reduce I/O costs and PM writes, we propose a new buffer manager named HiBuffer for DRAM/PM-based hybrid memory systems. HiBuffer presents several novel ideas. First, it adopts multi-grained data layouts to manage the hybrid buffer cooperatively. In addition to the page granularity, we introduce Lines for the DRAM buffer and Sectors to the PM buffer, forming a buffer with three granularities, including Line, Sector, and Page. We prove that the multi-grained cooperative buffer management can deliver higher performance than existing page-grained schemes. Second, we propose a sector-grained method to migrate data from DRAM to PM, which can avoid unnecessary data movements and reduce PM writes. Third, we use an out-of-place updating mechanism to absorb updates in DRAM, which can further reduce the writes to PM. We compare HiBuffer with three existing schemes, including LRU, CLOCK-DWF, and MiniPage, on five synthetic workloads and the YCSB benchmark using real Intel Optane DC Persistent Memory. The results in terms of various metrics, including running time, PM writes, hit ratio, and disk I/Os, suggest the efficiency of HiBuffer. In particular, HiBuffer reduces the running time by up to 37.8% and the writes to PM by up to 83% compared to the competitors when evaluated on the YCSB benchmark.

NVM: Is it Not Very Meaningful for Databases?

Persistent or Non Volatile Memory (PMEM) offers expanded memory capacity and faster access to persistent storage. However, there is no comprehensive empirical analysis of existing database engines under different PMEM modes, to understand how databases can benefit from the various hardware configurations. To this end, we analyze multiple different engines under common benchmarks with PMEM in AppDirect mode and Memory mode. Our results show that PMEM in Memory mode does not offer any clear performance advantage despite the larger volatile memory capacity. Also, using PMEM as persistent storage usually speeds up query execution, but with some caveats as the I/O path is not fully optimized and therefore does not always justify the additional cost. We show this to be the case through a comprehensive evaluation of different engines and database configurations under different workloads.

Classification of transient epileptic amnesia attacks: Two types of amnestic seizures, the pure amnesia type and partial amnesia type.

Transient epileptic amnesia (TEA) is a type of mesial temporal lobe epilepsy characterized by recurrent amnesia attacks. In 1998, Zeman et al. proposed the following diagnostic criteria for TEA: (1) recurrent, witnessed episodes of amnesia (TEA attacks); (2) other cognitive functions remain intact during attacks; and (3) evidence of epilepsy. It was also reported that patients with TEA often demonstrate two other types of memory symptoms: accelerated long-term forgetting (ALF) and autobiographical amnesia (AbA). Both ALF and AbA are persistent memory disorders, but transient epileptic seizures are not. We encountered two cases of TEA associated with two types of amnesia attacks. Therefore, we reviewed TEA cases in the literature to clarify the type of TEA attacks that occurred. Based on the extracted TEA cases, including our two cases, we found that there are two main types of TEA attacks, and we discussed their clinical features. We propose two main types of TEA attacks; that is, pure amnesia-type and partial amnesia-type seizures. Furthermore, we also propose that topographical amnesia mainly manifests as a type of amnesia attack, rather than as a chronic memory disturbance, such as ALF or AbA.

H2M: Exploiting Heterogeneous Shared Memory Architectures

Over the past decades, the performance gap between the memory subsystem and compute capabilities continued to spread. However, scientific applications and simulations show increasing demand for both memory speed and capacity. To tackle these demands, new technologies such as high-bandwidth memory (HBM) or non-volatile memory (NVM) emerged, which are usually combined with classical DRAM. The resulting architecture is a heterogeneous memory system in which no single memory is “best”. HBM is smaller but offers higher bandwidth than DRAM, whereas NVM provides larger capacity than DRAM at a reasonable cost and less energy consumption. Despite that, in several cases, DRAM still offers the best latency out of all three technologies.In order to use different kinds of memory, applications typically have to be modified to a great extent. Consequently, vendor-agnostic solutions are desirable. First, they should offer the functionality to identify kinds of memory, and second, to allocate data on it. In addition, because memory capacities may be limited, decisions about data placement regarding the different memory kinds have to be made. Finally, in making these decisions, changes over time in data that is accessed, and the actual access pattern, should be considered for initial data placement and be respected in data migration at run-time.In this paper, we introduce a new methodology that aims to provide portable tools and methods for managing data placement in systems with heterogeneous memory. Our approach allows programmers to provide traits (hints) for allocations that describe how data is used and accessed. Combined with characteristics of the platforms’ memory subsystem, these traits are exploited by heuristics to decide where to place data items. We also discuss methodologies for analyzing and identifying memory access characteristics of existing applications, and for recommending allocation traits.In our evaluation, we conduct experiments with several kernels and two proxy applications on Intel Knights Landing (HBM + DRAM) and Intel Ice Lake with Intel Optane DC Persistent Memory (DRAM + NVM) systems. We demonstrate that our methodology can bridge the performance gap between slow and fast memory by applying heuristics for initial data placement.

When Tree Meets Hash: Reducing Random Reads for Index Structures on Persistent Memories

Indexing structures are widely used in modern data-processing applications to support high-performance queries, and there are a variety of recent designs specifically optimized for the newly available persistent memory (PM). The primary focus of previous PM indexes is on reducing the expensive PM writes for persisting data. However, we find that in tree-based PM indexes, because of the smaller performance gap between writes and random reads on real PM devices, the read-intensive tree traversal phase dominates the overall latency. This observation calls for further optimizations on existing indexing structures for PM. In this paper, we propose Extendible Radix Tree (ERT), an efficient indexing structure for PM that significantly reduces tree heights to minimize random reads, while still maintaining fast in-node search speed. The key idea is to use extendible hashing for each node in a radix tree. This design allows us to have a relatively large fanout of the radix tree to keep the tree height small, and also to realize constant-time lookups within a node. Using extendible hashing also allows for incremental node modification without excessive writes during inserts and updates. Range queries are efficiently and robustly handled by enforcing partial ordering among the keys in the hash table of each node without introducing more hash collisions. Our experiments on both synthetic and real-world data sets demonstrate that ERT achieves up to 2.65×, 4.41×, and 2.43× speedups for search, insert, and range queries over the respectively state-of-the-art PM index.

Pea Hash: A Performant Extendible Adaptive Hashing Index

Hashing index is widely used to support efficient point operations. We observe that there is a conflict between performance and memory utilization goals. Existing hashing indices often have to trade off hash table access latency for better memory utilization. Moreover, many designs support only unique keys, and their performance is often suboptimal with skew workloads. In this paper, we propose Pea Hash with two techniques to address the above two problems: (i) adaptive hashing strategy that holistically optimizes both access latency and memory utilization, and (ii) data-aware adaptive buckets that accommodate unique keys, and keys with various numbers of duplicates. We develop both an NVM-optimized Pea Hash and a DRAM-based Pea Hash index. Experiments on a machine equipped with Intel Optane DC Persistent memory show that compared to state-of-the-art NVM-optimized hashing indices, the NVM-optimized Pea Hash achieves up to 13.8x performance improvements with similar memory utilization. The DRAM-based Pea Hash outperforms existing in-DRAM hashing index designs, showing the generality of the proposed techniques.

Persistent charge storage and memory operation of top-gate transistors solely based on two-dimensional molybdenum disulfide

We fabricate top-gate transistors on the three-layer molybdenum disulfide (MoS2) with three, two, and one layers in the source and drain regions using atomic layer etching (ALE). In the presence of ALE, the device at zero gate voltage can exhibit high and low levels of drain current under the forward and reverse gate bias, respectively. The hysteresis loop on the transfer curve of transistor indicates that two distinct charge states exist in the device within a range of gate bias. A long retention time of the charge is observed. Unlike conventional semiconductor memories with transistors and capacitors, the two-dimensional (2D) material itself plays two parts in the current conduction and charge storage. The persistent charge storage and memory operation of the multilayer MoS2 transistors with thicknesses of a few atomic layer will further expand the device application of 2D materials with reduced linewidths.

Evaluation of the cognitive outcome after out-of-hospital cardiac arrest: The role of thalamus.

Cardiac arrest survivors develop a variety of neuropsychological impairments and neuroanatomical lesions. The goal of this study is to evaluate if brain voxel-based morphometry and lesional Magnetic Resonance Imaging (MRI) analyses performed in the acute phase of an Out-of-Hospital Cardiac Arrest (OHCA) can be sensitive enough to predict the persistence of neuropsychological disorders beyond 3 months. Survivors underwent a prospective brain MRI during the first month after an OHCA and performed neuropsychological assessments at 1 and 3 months. According to the second neuropsychological assessment, survivors were separated into two subgroups, a deficit subgroup with persistent memory, executive functions, attention and/or praxis disorders (n = 11) and a preserved subgroup, disorders free (n = 14). Brain vascular lesion images were investigated, and volumetric changes were compared with healthy controls. Correlations were discussed between brain MRI results, OHCA data and the second neuropsychological assessment. Analyses of acute ischemic lesions did not reveal significant differences between the two subgroups (p= .35), and correlations with cognitive impairments could not be assessed. voxel-based morphometry analyses revealed a global cerebral volume reduction for the two subgroups and a clear decrease of the right thalamic volume for the deficit subgroup. It was associated with a cognitive dysexecutive syndrome represented by four executive indexes according to the 'Groupe de Réflexion pour l'Evaluation des Fonctions EXécutives' criteria. The right thalamus atrophy seems to be more predictive than the vascular lesions and more specific than a global cerebral volume reduction of post-OHCA neuropsychological executive disorders.

NICFS: a file system based on persistent memory and SmartNIC

NICFS: a file system based on persistent memory and SmartNIC

Few-Shot Temporal Sentence Grounding via Memory-Guided Semantic Learning

Temporal sentence grounding (TSG) is an important yet challenging task in video-based information retrieval. Given an untrimmed video input, it requires the machine to predict the interested video segment semantically related to a given sentence query. Most existing TSG methods train well-designed deep networks to align the semantic between video-query pairs for activity grounding with a large amount of data. However, we argue that these works easily capture the selection biases of video-query pairs in a dataset rather than showing the robust reasoning abilities to handle the rarely appeared pairs ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e</i> ., few-shot contents). To alleviate such limitation of the off-balance data distribution during the network training, in this paper, we propose a novel memory-augmented network called Memory-Guided Semantic Learning Network (MGSL-Net) to handle the few-shot TSG task for enhancing the model generalization ability. Specifically, given the matched video-query input, we first employ a graph attentive cross-modal interaction module to align their semantics in a cycle-consistent manner. Then, we develop the memory modules in both video and query domains to record the cross-modal shared semantic features in the domain-specific persistent memory. At last, a heterogeneous attention module is utilized to integrate the memory-enhanced multi-modal features in both video and query domains with further feature calibration. During training, the memory modules are dynamically associated with both common and rare cases to memorize all appeared contents, alleviating the issue of forgetting the few-shot contents. Therefore, in testing, the rare cases can be enhanced by retrieving the stored memories, improving the generalization ability of the model. Experimental results on three benchmarks (ActivityNet Caption, Charades-STA and TACoS) show the superiority of our method on both effectiveness and efficiency.

An efficient wear-leveling-aware multi-grained allocator for persistent memory file systems

An efficient wear-leveling-aware multi-grained allocator for persistent memory file systems

Whole-Transcriptome Analysis of Repeated Low-Level Sarin-Exposed Rat Hippocampus and Identification of Cerna Networks to Investigate the Mechanism of Sarin-Induced Cognitive Impairment.

Sarin is a potent organophosphorus nerve agent that causes cognitive dysfunction, but its underlying molecular mechanisms are poorly understood. In this study, a rat model of repeated low-level sarin exposure was established using the subcutaneous injection of 0.4 × LD50 for 21 consecutive days. Sarin-exposed rats showed persistent learning and memory impairment and reduced hippocampal dendritic spine density. A whole-transcriptome analysis was applied to study the mechanism of sarin-induced cognitive impairment, and a total of 1035 differentially expressed mRNA (DEmRNA), including 44 DEmiRNA, 305 DElncRNA, and 412 DEcircRNA, were found in the hippocampus of sarin-treated rats. According to Gene Ontology (GO) annotation, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, and Protein-Protein Interaction (PPI) analysis, these DERNAs were mainly involved in neuronal synaptic plasticity and were related to the pathogenesis of neurodegenerative diseases. The circRNA/lncRNA-miRNA-mRNA ceRNA network was constructed, in which Circ_Fmn1, miR-741-3p, miR-764-3p, miR-871-3p, KIF1A, PTPN11, SYN1, and MT-CO3 formed one circuit, and Circ_Cacna1c, miR-10b-5p, miR-18a-5p, CACNA1C, PRKCD, and RASGRP1 constituted another circuit. The balance between the two circuits was crucial for maintaining synaptic plasticity and may be the regulatory mechanism by which sarin causes cognitive impairment. Our study reveals the ceRNA regulation mechanism of sarin exposure for the first time and provides new insights into the molecular mechanisms of other organophosphorus toxicants.

Teaching NeuroImages: Persistent anterograde amnesia due to sequential, bilateral vascular damage to the Papez circuit

A 62-year-old previously cognitively intact man presented to the emergency department with acute impairment of anterograde memory (ability to form new memories).1 Head CT did not show acute changes; however, amnesia persisted. MRI disclosed an acute infarct in the left posterior hippocampus and a nonacute lacunar stroke in the right anterior thalamus (figure). Memory testing showed severe deficits in verbal and visual memory, persisting after 2 months. Bilateral lesions located anywhere in the Papez circuit (hippocampal formations, mammillary bodies, anterior thalamic nuclei, cingulate and parahippocampal cortices, and their connections via fornix and mammillothalamic tracts) may be implicated in marked, persistent anterograde memory deficits.1,2 The hippocampus receives its blood supply from collateral branches of the posterior cerebral artery and the anterior choroidal artery.3 The anterior thalamus is supplied by the tuberothalamic artery.4

Analysis of global DNA methylation and epigenetic modifiers (DNMTs and HDACs) in human foetal endothelium exposed to gestational and type 2 diabetes

ABSTRACT Foetuses exposed to maternal gestational diabetes (GDM) and type 2 diabetes (T2D) have an increased risk of adverse perinatal outcomes. Epigenetic mechanisms, including DNA methylation and histone modifications, may act as mediators of persistent metabolic memory in endothelial cells (ECs) exposed to hyperglycaemia, even after glucose normalization. Therefore, we investigated alterations in global DNA methylation and epigenetic modifier expression (DNMT1, DNMT3a, DNMT3b, HDAC1, and HDAC2) in human umbilical vein ECs (HUVECs) from the umbilical cords of mothers with GDM (n = 8) and T2D (n = 3) compared to that of healthy mothers (n = 6). Global DNA alteration was measured using a 5-methylation cytosine colorimetric assay, followed by quantitative real-time polymerase chain reaction to measure DNA methyltransferase and histone acetylase transcript expression. We revealed that DNA hypermethylation occurs in both GDM- and T2D-HUVECs compared to that in Control-HUVECs. Furthermore, there was a significant increase in HDAC2 mRNA levels in GDM-HUVECs and increase in DNMT3b mRNA levels in T2D-HUVECs. Overall, our results suggest that GDM and T2D are associated with global DNA hypermethylation in foetal endothelial cells under normoglycemic conditions and the aberrant mRNA expression of HDAC2 and DNMT3b could play a role in this dysregulation.

A critical perspective on updating drug memories through the integration of memory editing and brain stimulation.

Addiction is a persistent, recurring condition characterized by repeated relapses despite the desire to control drug use or maintain sobriety. The attainment of abstinence is hindered by persistent maladaptive drug-associated memories, which drive drug-seeking and use behavior. This article examines the preliminary evidence supporting the combination of non-invasive brain stimulation (NIBS) techniques and memory editing (or reconsolidation) interventions as add-on forms of treatment for individuals with substance-related disorders (SUD). Studies have shown that NIBS can modestly reduce drug use and craving through improved cognitive control or other undetermined reasons. Memory reconsolidation, a process by which a previously consolidated memory trace can be made labile again, can potentially erase or significantly weaken SUD memories underpinning craving and the propensity for relapse. This approach conveys enthusiasm while also emphasizing the importance of managing boundary conditions and null results for interventions found on fear memory reconsolidation. Recent studies, which align with the state-dependency and activity-selectivity hypotheses, have shown that the combination of NIBS and behavioral interventions holds promise for treating SUD by reducing self-reported and physiological aspects of craving. Effective long-term outcomes for this procedure require better identification of critical memories, a deeper understanding of the brain mechanisms underlying SUD and memory reconsolidation and overcoming any boundary conditions of destabilized memories. This will enable the procedure to be personalized to the unique needs of individual patients.

Sex and age affect acute and persisting COVID-19 illness

Long COVID is associated with neurological and neuropsychiatric manifestations. We conducted an observational study on 97 patients with prior SARS-CoV-2 infection and persisting cognitive complaints that presented to the University Health Network Memory Clinic between October 2020 and December 2021. We assessed the main effects of sex, age, and their interaction on COVID-19 symptoms and outcomes. We also examined the relative contribution of demographics and acute COVID-19 presentation (assessed retrospectively) on persistent neurological symptoms and cognition. Among our cohort, males had higher hospitalization rates than females during the acute COVID-19 illness (18/35 (51%) vs. 15/62 (24%); P = .009). Abnormal scores on cognitive assessments post-COVID were associated with older age (AOR = 0.84; 95% CI 0.74–0.93) and brain fog during initial illness (AOR = 8.80; 95% CI 1.76–65.13). Female sex (ARR = 1.42; 95% CI 1.09–1.87) and acute shortness of breath (ARR = 1.41; 95% CI 1.09–1.84) were associated with a higher risk of experiencing more persistent short-term memory symptoms. Female sex was the only predictor associated with persistent executive dysfunction (ARR = 1.39; 95% CI 1.12–1.76) and neurological symptoms (ARR = 1.66; 95% CI 1.19–2.36). Sex differences were evident in presentations and cognitive outcomes in patients with long COVID.