Search

Systolic arrays with promising attributes, such as high degree of concurrent computation and high data-reuse rate, are attractive solutions for dense linear algebra. Recently, systolic arrays have been used for accelerating the inference of deep neural networks (DNNs). However, as sparsification mechanisms are applied to DNNs during or after training, DNN inference is usually a sparse problem. Therefore, it cannot fully benefit from the fundamental advantages offered by systolic arrays. To solve this challenge, we propose Eridanus, an approach to structured pruning that produces DNNs compatible with the synchronous and rhythmic flow of data from memory to systolic arrays. [ABSTRACT FROM AUTHOR]

The poem "Bright Rags" by Kim Hyesoon is presented. First Line: I want to lie down beneath the crotches of the women carrying water jars; Last Line: The large pink rags held up by the women are bright.

The poem "Moonrise" by Kim Hyesoon and translated by Don Mee Choi is presented. First Line: The night sky bowed down like a black well; Last Line: finally let out a sigh of relief, trembling.

Scientific computations with a wide range of applications in domains such as developing vaccines, forecasting the weather, predicting natural disasters, simulating aerodynamics of spacecraft, and exploring oil resources, create the main workloads of supercomputers. The key integration of such scientific computations is modeling physical phenomena that are done with the aid of partial differential equations (PDEs). Solving PDEs on supercomputers, even with those equipped with GPUs, consumes a large amount of power and yet is not as fast as desired. The main reason behind such slow processing is data dependency. The key challenge is that software techniques cannot resolve these dependencies, therefore, such applications cannot benefit from the parallelism provided by processors such as GPUs. Our key insight to address this challenge is that although we cannot resolve the dependencies, we can reduce their negative impacts by using hardware/software co-optimization. To this end, we propose breaking down the data-dependent operations into two groups of operations: a majority of parallelizable and the minority of data-dependent operations. We execute these two groups in the desired order: first, we put together all parallelizable operations and execute them all, subsequently; then, we switch to execute the small data-dependent part. As long as the data-dependent part is small, we can accelerate them by using fast hardware mechanisms. Besides, our proposed hardware mechanisms guarantee quickly switching between the two groups of operations. To follow the same order of execution, dictated by our software mechanism, and implemented in hardware, we also propose a new low-overhead compression format – sparsity is another attribute of PDEs that require compression. Furthermore, the core generic architecture of our proposed hardware allows the execution of other applications including sparse matrix-vector multiplication (SpMV) and graph algorithms. The key feature of the proposed hardware is partial reconfigurability, which on one hand, facilitates the execution of data-dependent computations, and on the other hand, allows executing broad application without changing the entire configuration. Our evaluations show that compared to GPUs, we achieve an average speedup of 15.6× for scientific computations while consuming 14× less energy. [ABSTRACT FROM AUTHOR]

With the advent of die stacking technology and big data applications, Processing-in-memory (PIM) is regaining attention as a promising technology for improving performance and energy efficiency. Although various PIM techniques have been proposed in recent studies for effectively offloading computation from the host, the thermal impacts of PIM offloading have not been fully explored. This paper investigates the thermal constraints of PIM and proposes techniques to enable thermal awareness for efficient PIM offloading. To understand the thermal effects of 3D-stacked designs, we measure the temperature of a real Hybrid Memory Cube (HMC) prototype and observe that compared to conventional DRAM, HMC reaches a significantly higher operating temperature, which causes thermal shutdowns with a passive cooling solution. Even with a commodity-server cooling solution, when in-memory processing is highly utilized, HMC fails to maintain the temperature of the memory dies within the normal operating range. In this paper, we propose a collection of software- and hardware-based techniques to enable thermal-aware PIM offloading by controlling the intensity of PIM offloading at runtime. Our evaluation results show that the proposed techniques achieve up to 1.4 × and 1.37 × speedups compared to non-offloading and naïve offloading scenarios. • We provide a thermal analysis to understand the thermal constraints of a real-world 3D memory. • Our thermal analysis shows that DRAM layers can exceed the normal operating temperature when PIM offloading is used. • We propose a collection of thermal-aware source throttling techniques that dynamically controls the intensity of PIM offloading. [ABSTRACT FROM AUTHOR]

Developmental endothelial locus-1 (Del-1) is an endogenous anti-inflammatory molecule that is highly expressed in the lung and the brain and limits leukocyte migration to these tissues. We previously reported that the expression of Del-1 is positively regulated by p53 in lung endothelial cells. Although several reports have implicated the altered expression of Del-1 gene in cancer patients, little is known about its role in tumor cells. We here investigated the effect of Del-1 on the features of human lung carcinoma cells. Del-1 mRNA was found to be significantly decreased in the human lung adenocarcinoma cell lines A549 (containing wild type of p53), H1299 (null for p53) and EKVX (mutant p53), compared to in human normal lung epithelial BEAS-2B cells and MRC-5 fibroblasts. The decrease of Del-1 expression was dependent on the p53 activity in the cell lines, but not on the expression of p53. Neither treatment with recombinant human Del-1 protein nor the introduction of adenovirus expressing Del-1 altered the expression of the apoptosis regulators BAX, PUMA and Bcl-2. Unexpectedly, the adenovirus-mediated overexpression of Del-1 gene into the lung carcinoma cell lines promoted proliferation and invasion of the lung carcinoma cells, as revealed by BrdU incorporation and transwell invasion assays, respectively. In addition, overexpression of the Del-1 gene enhanced features of epithelial–mesenchymal transition (EMT), such as increasing vimentin while decreasing E-cadherin in A549 cells, and increases in the level of Slug, an EMT-associated transcription regulator. Our findings demonstrated for the first time that there are deleterious effects of high levels of Del-1 in lung carcinoma cells, and suggest that Del-1 may be used as a diagnostic or prognostic marker for cancer progression, and as a novel therapeutic target for lung carcinoma. [ABSTRACT FROM AUTHOR]

Objectives: This study investigated quality of life in Korean juvenile inmates with attention-deficit/hyperactivity disorder (ADHD) and the impact of behavioral and emotional problems on quality of life., Methods: In total, 200 inmates were evaluated using the Korean version of the Mini-International Neuropsychiatric Interview (K-MINI) and the Korean version of the Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime (K-SADS-PL-K). We extracted the inmates with ADHD and evaluated their quality of life, behavioral problems, and emotional problems with the Pediatric Quality of Life Inventory (PedsQL) and the Korean Youth Self Report (K-YSR) scale. Descriptive statistics, Pearson correlation analysis, and multiple regression analysis were conducted., Results: Among the 200 total inmates, 68 were diagnosed with ADHD by the K-SADS-PL-K. Most of the correlations between PedsQL scores and K-YSR items were significant. Multiple regression analysis showed that PedsQL could be predicted by affective problems (among the DSM-oriented scales of the K-YSR) and attention problems (among the syndrome scales of the K-YSR)., Conclusion: Our results demonstrate that, among juvenile inmates with ADHD, quality of life was negatively correlated with most behavioral and emotional problems. Meanwhile, the significant influence of affective and attention problems on inmates' quality of life suggests the necessity of comprehensive treatments for this group., Competing Interests: Conflicts of Interest The authors have no financial conflicts of interest., (Copyright: © The Korean Society for Applied Biological Chemistry.)

Inflammation in the central nervous system (CNS) and disruption of its immune privilege are major contributors to the pathogenesis of multiple sclerosis (MS) and of its rodent counterpart, experimental autoimmune encephalomyelitis (EAE). We have previously identified developmental endothelial locus-1 (Del-1) as an endogenous anti-inflammatory factor, which inhibits integrin-dependent leukocyte adhesion. Here we show that Del-1 contributes to the immune privilege status of the CNS. Intriguingly, Del-1 expression decreased in chronic-active MS lesions and in the inflamed CNS in the course of EAE. Del-1-deficiency was associated with increased EAE severity, accompanied by increased demyelination and axonal loss. As compared with control mice, Del-1(-/-) mice displayed enhanced disruption of the blood-brain barrier and increased infiltration of neutrophil granulocytes in the spinal cord in the course of EAE, accompanied by elevated levels of inflammatory cytokines, including interleukin-17 (IL-17). The augmented levels of IL-17 in Del-1-deficiency derived predominantly from infiltrated CD8(+) T cells. Increased EAE severity and neutrophil infiltration because of Del-1-deficiency was reversed in mice lacking both Del-1 and IL-17 receptor, indicating a crucial role for the IL-17/neutrophil inflammatory axis in EAE pathogenesis in Del-1(-/-) mice. Strikingly, systemic administration of Del-1-Fc ameliorated clinical relapse in relapsing-remitting EAE. Therefore, Del-1 is an endogenous homeostatic factor in the CNS protecting from neuroinflammation and demyelination. Our findings provide mechanistic underpinnings for the previous implication of Del-1 as a candidate MS susceptibility gene and suggest that Del-1-centered therapeutic approaches may be beneficial in neuroinflammatory and demyelinating disorders.

The gltX gene from Xanthomonas oryzae pv. oryzae (Xoo1504) encodes glutamyl-tRNA synthetase (GluRS), one of the most important enzymes involved in bacterial blight (BB), which causes huge production losses of rice worldwide. GluRS is a class I-type aminoacyl-tRNA synthetase (aaRS) that is primarily responsible for the glutamylation of tRNA(Glu). It plays an essential role in protein synthesis, as well as the regulation of cells, in all organisms. As it represents an important target for the development of new antibacterial drugs against BB, determination of the three-dimensional structure of GluRS is essential in order to understand its catalytic mechanism. In order to analyze its structure and function, the gltX gene was cloned and the GluRS enzyme was expressed, purified and then crystallized. A GluRS crystal belonging to the monoclinic space group C2 diffracted to 2.8 A resolution and had unit-cell parameters a = 186.8, b = 108.4, c = 166.1 A, beta = 96.3 degrees . The unit-cell volume of the crystal allowed the presence of six to eight monomers in the asymmetric unit, with a corresponding Matthews coefficient (V(M)) range of 2.70-2.02 A(3) Da(-1) and a solvent-content range of 54.5-39.3%.

Xanthomonas oryzae pv. oryzae (Xoo) causes bacterial blight of rice (Oryza sativa L.), one of the most devastating diseases of rice in most rice-growing countries. XometC, a cystathionine gamma-lyase (CGL) like protein that is an antibacterial drug-target protein against Xoo, was cloned, expressed, purified and crystallized. CGL catalyzes the second step in the reverse-transsulfuration pathway, which is essential for the metabolic interconversion of the sulfur-containing amino acids cysteine and methionine. Crystals of two different shapes, plate-shaped and pyramid-shaped, diffracted to 2.9 and 3.2 A resolution and belonged to the primitive orthogonal space group P2(1)2(1)2(1) and the tetragonal space group P4(1) (or P4(3)), with unit-cell parameters a = 73.0, b = 144.9, c = 152.3 A and a = b = 78.2, c = 300.7 A, respectively. For the P2(1)2(1)2(1) crystals, three or four monomers exist in the asymmetric unit with a corresponding V(M) of 3.02 or 2.26 A(3) Da(-1) and a solvent content of 59.3 or 45.7%. For the P4(1) (or P4(3)) crystals, four or five monomers exist in the asymmetric unit with a corresponding V(M) of 2.59 or 2.09 A(3) Da(-1) and a solvent content of 52.5 or 40.6%.