检索结果-内蒙古大学图书馆

High performance rapid single-flux-quantum bit-slice arithmetic logic unit

Superconductivity 2024年第3期11卷 73-79页

作者： Jing Ren Pei-Yao Qu Jia-Hong Yang Xiang-Yu Zheng Hui Zhang Jie Ren Guang-Ming Tang State Key Lab of Processors Institute of Computing TechnologyChinese Academy of SciencesBeijing 100190China University of Chinese Academy of Sciences Beijing 100049China Shanghai Institute of Microsystem and Information Technology Chinese Academy of SciencesShanghai 200233China

Two optimization technologies, namely, bypass and carry-control optimization, were demonstrated for enhancing the performance of a bit-slice Arithmetic Logic Unit (ALU) in 2n-bit Rapid Single-Flux-Quantum (RSFQ) microprocessors. These technologies can not only shorten the calculation time but also solve data hazards. Among them, the proposed bypass technology is applicable to any 2n-bit ALU, whether it is bit-serial, bit-slice or bit-parallel. The high performance bit-slice ALU was implemented using the 6 kA/cm^(2) Nb/AlOx/Nb junction fabrication process from Superconducting Electronics Facility of Shanghai Institute of Microsystem and Information technology. It consists of 1693 Josephson junctions with an area of 2.46 0.81 mm^(2). All ALU operations of the MIPS32 instruction set are implemented, including two extended instructions, i.e., addition with carry (ADDC) and subtraction with borrow (SUBB). All the ALU operations were successfully obtained in SFQ testing based on OCTOPUX and the measured DC bias current margin can reach 86% - 104%. The ALU achieves a 100 utilization rate, regardless of carry/borrow read-after-write correlations between instructions.

关键词： High performance Rapid Single-Flux-Quantum(RSFQ) Arithmetic Logic Unit(ALU) Optimization technologies Superconducting integrated circuits

来源：评论

学校读者我要写书评

暂无评论

Near-term quantum computing techniques: Variational quantum algorithms, error mitigation, circuit compilation, benchmarking and classical simulation

引用

science china(Physics,Mechanics & Astronomy) 2023年第5期66卷 23-72页

作者： He-Liang Huang Xiao-Yue Xu Chu Guo Guojing Tian Shi-Jie Wei Xiaoming Sun Wan-Su Bao Gui-Lu Long Henan Key Laboratory of Quantum Information and Cryptography Zhengzhou 450000China State Key Lab of Processors Institute of Computing TechnologyChinese Academy of SciencesBeijing 100190China School of Computer Science and Technology University of Chinese Academy of SciencesBeijing 100049China Beijing Academy of Quantum Information Sciences Beijing 100193China Department of Physics Tsinghua UniversityBeijing 100084China

Quantum computing is a game-changing technology for global academia,research centers and industries including computational science,mathematics,finance,pharmaceutical,materials science,chemistry and *** it has seen a major boost in the last decade,we are still a long way from reaching the maturity of a full-fledged quantum *** said,we will be in the noisy-intermediate scale quantum(NISQ)era for a long time,working on dozens or even thousands of qubits quantum computing *** outstanding challenge,then,is to come up with an application that can reliably carry out a nontrivial task of interest on the near-term quantum devices with non-negligible quantum *** address this challenge,several near-term quantum computing techniques,including variational quantum algorithms,error mitigation,quantum circuit compilation and benchmarking protocols,have been proposed to characterize and mitigate errors,and to implement algorithms with a certain resistance to noise,so as to enhance the capabilities of near-term quantum devices and explore the boundaries of their ability to realize useful ***,the development of near-term quantum devices is inseparable from the efficient classical sim-ulation,which plays a vital role in quantum algorithm design and verification,error-tolerant verification and other *** review will provide a thorough introduction of these near-term quantum computing techniques,report on their progress,and finally discuss the future prospect of these techniques,which we hope will motivate researchers to undertake additional studies in this field.

关键词： quantum computing noisy-intermediate scale quantum variational quantum algorithms error mitigation circuit com-pilation benchmarking protocols classical simulation

来源：评论

学校读者我要写书评

暂无评论

Achieving high strength in laser powder-bed fusion processed AlFeCuZr alloy via dual-nanoprecipitations and grain boundary segregation

引用

Journal of Materials science & technology 2023年第6期137卷 56-66页

作者： Jing-Yu Xu Cheng Zhang Li-Xue Liu Rong Guo Ming-Jun Sun Lin Liu School of Materials Science and Engineering State Key Lab for Materials Processing and Die&Mold TechnologyHuazhong University of Science and TechnologyWuhan 430074China

Additive manufacturing of aluminum alloys has received significant attention in the aerospace industry;however,achieving sufficient high strength,especially at elevated temperatures,remains ***,a crack-free and near-full dense Al-1Fe-0.6Cu-1.3Zr alloy was fabricated by the laser powder bed fusion(LPBF)*** Al-Fe-Cu-Zr alloy exhibits heterogeneous microstructures with two distinct *** is the so-called coarse-grain zones(CGZs)with an average grain size of 0.95μm,where(Al,Cu)Fe_(3) nanoparticles precipitate in the Al matrix and Fe and Cu cosegregate at the grain boundaries(GBs).The other is fine-grain zones(FGZs)with an average grain size of 0.45μm,where an Al 3 Zr nanoparti-cle precipitates in each of theα-Al grains(serves as the nuclei),and Fe-rich nanoprecipitates and Fe/Cu cosegregation appear at the *** a result,the LPBF Al-Fe-Cu-Zr alloy,with these unique heteroge-neous structures,displays high strength at both room temperature and elevated temperatures,e.g.,with high yield strengths of 500 MPa at room temperature,and 163 MPa at 573 K,both are higher than those of additive manufactured Al-based alloys reported thus *** is suggested that the high strength over a wide temperature range of the current LPBF Al alloy is mainly attributed to the combination of the precipitation strengthening mechanism and grain-boundary strengthening mechanism.

关键词： Laser powder bed fusion Aluminum alloy Heterogeneous microstructure Nanoprecipitation High strength

来源：评论

学校读者我要写书评

暂无评论

Nanopore-rich NiFe LDH targets the formation of the high-valent nickel for enhanced oxygen evolution reaction

引用

Nano Research 2023年第2期16卷 2286-2293页

作者： Qunlei Wen Shuzhe Wang Ruiwen Wang Danji Huang Jiakun Fang Youwen Liu Tianyou Zhai State Key Laboratory of Materials Processing and Die&Mould Technology School of Materials Science and EngineeringHuazhong University of Science and TechnologyWuhan 430074China State Key Lab of Advanced Electromagnetic Engineering and Technology School of Electrical and Electronic EngineeringHuazhong University of Science and TechnologyWuhan 430074China

Nickel-iron layered double hydroxides(NiFe LDHs)represent a promising candidate for oxygen evolution reaction(OER),however,are still confronted with insufficient activity,due to the slow kinetics of electrooxidation of Ni^(2+)cations for the high-valent active ***,nanopore-rich NiFe LDH(PR-NiFe LDH)nanosheets were proposed for enhancing the OER activity together with *** the designed catalyst,the confined nanopores create abundant unsaturated Ni sites at edges,and decrease the migration distance of protons down to the scale of their mean free path,thus promoting the formation of high-valent Ni^(3+)/^(4+)active *** unique configuration further improves the OER stability by releasing the lattice stress and accelerating the neutralization of the local acidity during the phase ***,the optimized PR-NiFe LDH catalysts exhibit an ultralow overpotential of 278 mV at 10 mA∙cm^(−2)and a small Tafel slope of 75 mV∙dec^(−1),which are competitive among the advanced LDHs based ***,the RP-NiFe LDH catalyst was implemented in anion exchange membrane(AEM)water electrolyzer devices and operated steadily at a high catalytic current of 2 A over 80 *** results demonstrated that PR-NiFe LDH could be a viable candidate for the practical *** concept also provides valuable insights into the design of other catalysts for OER and beyond.

关键词： nickel-iron layered double hydroxides(NiFe LDHs) water oxidation nanopores high-valent sites

来源：评论

学校读者我要写书评

暂无评论

Nd^(3)+:YAG-Al_(2)O_(3) nanocrystalline transparent ceramics with high inflexion concentration quenching of Nd^(3+) prepared by amorphous crystallization

引用

Journal of Advanced Ceramics 2024年第8期13卷 1242-1253页

作者： Xiujuan Wan Guangfan Tan Liang Cai Jie Fu Jianqiang Li Yingchun Zhang School of Materials Science and Engineering University of Science and Technology BeijingBeijing 100083China State Key Laboratory of New Ceramics and Fine Processing School of Materials Science and EngineeringTsinghua UniversityBeijing 100080China State Key Lab for Mesoscience and Engineering Institute of Process EngineeringChinese Academy of SciencesBeijing 100190China

Rare earth ion-doped Y_(3)Al_(5)O_(12)(YAG)-based transparent ceramics have been used as important laser gain media for a long time,yet the doping concentration of active ions is limited due to concentration quenching,wherein the inflexion concentration quenching of Nd^(3+)is recognized as 1.0 at%.In this work,YAG-Al_(2)O_(3) nanocrystalline transparent ceramics with a concentration of Nd^(3+)(O-5.0 at%)were fabricated via amorphous crystallization,and the crystal structure evolution,morphology,and optical properties were systematically investigated by differential scanning calorimetry(DSC),X-ray powder diffraction(XRD),transmission electron microscopy(TEM),magnetic resonation(MAS),nuclear magnetic resonation(NMR),and fluorescence *** doping of Nd^(3+)can promote the transition of Al[5]and Al[6]to Al[14],indicating improvements in the ability of the amorphous material to form Nd^(3+):Y_(2)O_(3)-Al_(2)O_(3) vitrified beads,and 1.5 at%Nd^(3+):YAG-Al_(2)O_(3) nanocrystalline transparent ceramics can be obtained by crystallization at 1050℃ with a matrix composed of YAG and concomitant δ-Al_(2)O_(3) and θ-Al_(2)O_(3).The nanocrystalline transparent ceramics show an internal transmitance of 89.56%at 1064 nm,and the strongest emission peak corresponds to the energy transfer from 4F_(3/2) to 4l_(11/2) of Nd^(3+)with a fluorescence lifetime of 231μs when pumped by an 808 nm ***,spectral broadening begins to occur,indicating the onset of concentration quenching,when the concentration of Nd^(3+)exceeds 1.5 at%,substantially higher than the 1.0 at% observed in YAG ***-Al_(2)O_(3) nanocrystalline transparent ceramics obtained by amorphous crystalization can be utilized as the matrix to increase the inflexion point of doping concentration quenching of Nd^(3+),and this material may have great potential as a laser gain medium.

关键词： Y_(3)Al_(5)O_(12)(YAG)-Al_(2)O_(3) nanocrystalline transparent ceramics laser gain medium Nd^(3+)

来源：评论

学校读者我要写书评

暂无评论

High-temperature creep rupture behavior of dissimilar welded joints of RAFM steel and 316L steel by friction stir welding

引用

Journal of Iron and Steel Research International 2024年第10期31卷 2438-2447页

作者： Xin Ji Bin He Wei Guan Chen-xi Liu Hui-jun Li Lei Cui Yong-chang Liu Tianjin Key Laboratory of Advanced Joining Technology School of Materials Science and EngineeringTianjin UniversityTianjin300354China State Key Lab of Hydraulic Engineering Simulation and Safety School of Materials Science and EngineeringTianjin UniversityTianjin300350China

Creep rupture of the reduced activation ferritic/martensitic(RAFM)steel and 316L steel dissimilar joint by friction stir welding was *** creep rupture time of the dissimilar joint was 1941 h at 600℃/100 MPa and 120 h at 650℃/100 *** creep fracture occurred in heat affect zone(HAZ)of RAFM steel side where coarse Laves phase was *** formation and coarsening of the Laves phase particles in HAZ of RAFM steel side were the main reasons that caused the creep fracture of the dissimilar *** Laves phase particles nucleated adjacent to the large M_(23)C_(6) particles at the grain boundaries where W element segregated and grew fast during creep *** large Laves phase would deteriorate the pinning effect of M_(23)C_(6) carbides and weaken the solid solution strengthening ***,the size of the Laves phase in HAZ of RAFM steel side was bigger than that in stir zone of RAFM steel *** reasons explain the creep fracture in HAZ of RAFM steel side of dissimilar joint.

关键词： Creep behavior Dissimilar joint Friction stir welding Laves phase M23C6 carbide

来源：评论

学校读者我要写书评

暂无评论

SmartATPG: Learning-based Automatic Test Pattern Generation with Graph Convolutional Network and Reinforcement Learning 24

SmartATPG: Learning-based Automatic Test Pattern Generation ...

引用

61st ACM/IEEE Design Automation Conference, DAC 2024

作者： Li, Wenxing Lyu, Hongqin Liang, Shengwen Wang, Tiancheng Li, Huawei State Key Lab of Processors Institute of Computing Technology Cas Beijing China University of Chinese Academy of Sciences Beijing China Castest Beijing China

ISBN: (纸本)9798400706011

Automatic test pattern generation (ATPG) is a critical technology in integrated circuit testing. It searches for effective test vectors to detect all possible faults in the circuit as entirely as possible, thereby ensuring chip yield and improving chip quality. However, the process of searching for test vectors is NP-complete. At the same time, the large amount of backtracking generated during the search for test vectors can directly affect the performance of ATPG. In this paper, a learning-based ATPG framework, SmartATPG, is proposed to search for high-quality test vectors, reduce the number of backtracking during the search process, and thereby improve the performance of ATPG. SmartATPG utilizes graph convolutional network (GCN) to fully extract circuit feature information and efficiently explore the ATPG search space through reinforcement learning (RL). Experimental results indicate that the proposed SmartATPG outperforms traditional and artificial neural network (ANN)-based heuristic strategies on most benchmark circuits. © 2024 Copyright is held by the owner/author(s). Publication rights licensed to ACM.

关键词： Integrated circuit testing

来源：评论

学校读者我要写书评

暂无评论

Aceso: Efficient Parallel DNN Training through Iterative Bottleneck Alleviation 24

Aceso: Efficient Parallel DNN Training through Iterative Bot...

引用

19th European Conference on Computer Systems, EuroSys 2024

作者： Liu, Guodong Miao, Youshan Lin, Zhiqi Shi, Xiaoxiang Maleki, Saeed Yang, Fan Bao, Yungang Wang, Sa State Key Lab of Processors Institute of Computing Technology Cas Univ. of Chinese Academy of Sciences China Microsoft Research United States University of Science and Technology of China China Shanghai Jiao Tong University China

ISBN: (纸本)9798400704376

Many parallel mechanisms, including data parallelism, tensor parallelism, and pipeline parallelism, have been proposed and combined together to support training increasingly large deep neural networks (DNN) on massive GPU devices. Given a DNN model and GPU cluster, finding the optimal configuration by combining these parallelism mechanisms is an NP-hard problem. Widely adopted mathematical programming approaches have been proposed to search in a configuration subspace, but they are still too costly when scaling to large models over numerous devices. Aceso is a scalable parallel-mechanism auto-configuring system that operates iteratively. For a given parallel configuration, Aceso identifies a performance bottleneck and then, by summarizing all possible configuration adjustments with their resource consumption changes, infers their performance impacts to the bottleneck and selects one that mitigates the bottleneck. This process repeats for many iterations until a desired final configuration is found. Unlike mathematical programming approaches that examine the configurations subspace to find the optimal solution, Aceso searches in the configuration space in a stochastic approach by repeatedly identifying and alleviating bottlenecks. Aceso significantly reduces configuration searching cost by taking the approach of resolving one bottleneck at a time. This allows Aceso to find configurations that would be usually missed in subspace search approaches. We implemented and tested Aceso on representative DNN models. Evaluations show that it can scale to 1K-layer models. Compared to state-of-the-art systems, Aceso achieves up to 1.33× throughput improvement with less than 5% of the searching cost. © 2024 ACM.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

ProXplore: A GPU-Enhanced Protein Discovery Engine

ProXplore: A GPU-Enhanced Protein Discovery Engine

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Sun, Zewen Zhang, Peiheng Fu, Xingzhuo Li, Xueqi Cas Institute of Intelligent Computing Technology Suzhou China University of Chinese Academy of Sciences China Cas State Key Lab of Processors Institute of Computing Technology China

ISBN: (纸本)9798350386226

Language models have achieved unprecedented success in natural language processing tasks and have recently been adapted for biological sequences. However, GPUs still encounter significant performance bottlenecks when running BERT-style natural language models. Moreover, due to the vastly greater number of tokens in protein sequences compared to human languages (as illustrated in Figure 1), directly transferring language models from human languages to proteins results in significant changes in runtime behavior. Consequently, optimizations designed for short-input BERT models are less effective for protein language *** this paper, we propose a novel GPU-based cross-layer optimization strategy. From an architectural perspective, our approach leverages intra- and inter-operator parallelism, multilevel data computation and warp parallelism to fully utilize GPU features. From an algorithmic perspective, we address the issue of varying lengths in protein input sequences and reduce memory overhead through padding removal. Experimental results demonstrate that ProXplore achieves a 1.8× increase in inference speed on the TAPE benchmark without a significant loss in prediction accuracy. This improvement effectively overcomes performance bottlenecks and provides substantial benefits for protein sequence analysis and other bioinformatics applications. © 2024 IEEE.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

SRender: Boosting Neural Radiance Field Efficiency via Sensitivity-Aware Dynamic Precision Rendering 57

SRender: Boosting Neural Radiance Field Efficiency via Sensi...

引用

57th Annual IEEE/ACM International Symposium on Microarchitecture, MICRO 2024

作者： Song, Zhuoran He, Houshu Liu, Fangxin Hao, Yifan Song, Xinkai Jiang, Li Liang, Xiaoyao Shanghai Jiao Tong University Department of Computer Science and Engineering Shanghai China Institute of Computing Technology Chinese Academy of Sciences State Key Lab of Processors Beijing China

ISBN: (纸本)9798350350579

Neural Radiance Field (NeRF) holds immense promise for generating photo-realistic images and videos. How-ever, the computation and memory demands significantly impede its applicability in real-Time tasks such as virtual reality and augmented reality. While accelerators dedicatedly designed for NeRF have emerged, they may not be the optimal solution as they only consider the optimization of full-precision NeRF models, leading to unsatisfactory memory and computation saving. This paper proposes Sensitivity-Aware Dynamic Precision Rendering, abbreviated as SRender, an algorithm-hardware co-design framework that borrows the idea of Adaptive Rendering commonly used in conventional graphics rendering to accelerate the rendering process of NeRF exploiting the fine-grained per-ray and per-point levels sensitivity. The fundamental idea behind SRender is that rays and points with high sensitivity require high-precision data types, while the insensitive rays and points do not, allowing them to be sacrificed for greater acceleration. Accordingly, the algorithm part of SRender dynamically configures the precision of the hash table and the multilayer perceptron (MLP), and the hardware part features a specialized accelerator that exploits such mixed-precision memory access and computation patterns. Specifically, we predict the sensitive rays and points based on low-precision hash encoding and feature computation. Then, we trigger a high-precision rendering again on the sensitive ones to guarantee the rendering quality. The hardware design of SRender includes a sensitivity prediction engine, coarse-Then-fine-grained encoding engines, and a recovery engine, which translates the computational saving in the algorithm to real speedup. Using the coarse-Then-fine-grained encoding engine, we maximize the data reuse on-chip, thereby reducing the stress of off-chip memory bandwidth. Experiments show that the edge-level SRender can achieve 3.94x and 34.06x speedup over state-of-The-Art NeRF

关键词： Encoding (symbols)

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：