Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 43 for “"Processing in memory (PIM)"”.

  1. Integrating Processing In-Memory (PIM) Technology into General Purpose Graphics Processing Units (GPGPU) for Energy Efficient Computing

    Processing-in-memory (PIM) offers a viable solution to overcome the memory wall crisis that has been plaguing memory system for decades. Due to advancements in 3D stacking technology in recent years, PIM provides an opportunity to reduce both energy and data movement overheads, which are the …

    houston Repository record for Integrating Processing In-Memory (PIM) Technology into General Purpose Graphics Processing Units (GPGPU) for Energy Efficient Computing (opens in a new tab)

  2. Enhancing Sparse Data Processing through Energy-Efficient Heterogeneous Architectures

    Efficiently processing sparse data is a significant challenge in many high-impact domains, including machine learning, graph analytics, and scientific computing. Sparse data representations, such as compressed sparse row (CSR), often suffer from substantial overheads during indexing and irregular …

    unt Repository record for Enhancing Sparse Data Processing through Energy-Efficient Heterogeneous Architectures (opens in a new tab)

  3. Variability-aware training and self-tuning of highly quantized DNNs for analog PIM

    DNNs deployed on analog processing in memory (PIM) architectures are subject to fabrication-time variability. We developed a new joint variability- and quantization-aware DNN training algorithm for highly quantized analog PIM-based models that is significantly more effective than prior work. It …

    texas Repository record for Variability-aware training and self-tuning of highly quantized DNNs for analog PIM (opens in a new tab)

  4. Accelerating irregular applications via efficient synchronization and data access techniques

    Irregular applications comprise an increasingly important workload domain for many fields, including bioinformatics, chemistry, graph analytics, physics, social sciences and machine learning. Therefore, achieving high performance and energy efficiency in the execution of emerging irregular …

    greece Repository record for Accelerating irregular applications via efficient synchronization and data access techniques (opens in a new tab)

  5. Exploring the potential for accelerating sparse matrix-vector product on a Processing-in-Memory architecture

    As the importance of memory access delays on performance has mushroomed over the past few decades, researchers have begun exploring Processing-in-Memory (PIM) technology, which offers higher memory bandwidth, lower memory latency, and lower power consumption. In this study, we investigate whether …

    rice Repository record for Exploring the potential for accelerating sparse matrix-vector product on a Processing-in-Memory architecture (opens in a new tab)

  6. Impact of HKMG and FDSOI FeFET Drain Current Variation in Processing-in-Memory Architectures

    In this thesis, we analyze the impact of drain current variation in 28 nm high-K metal-gate and 22 nm fully-depleted silicon-on-insulator Ferroelectric FET devices on processing-in-memory (PIM) deep neural network (DNN) accelerators. When performing repeated read operations on several devices at …

    gatech Repository record for Impact of HKMG and FDSOI FeFET Drain Current Variation in Processing-in-Memory Architectures (opens in a new tab)

  7. BREAKING MEMORY WALLS: EFFICIENT MEMORY MANAGEMENT FOR THE POST-MOORE’S ERA

    … transistors near their limits, computing systems face growing challenges—especially in memory performance—while demand for power-hungry applications like large language models (LLMs) continues to rise. This thesis addresses four major “memory walls” limiting modern systems: …

    nus Repository record for BREAKING MEMORY WALLS: EFFICIENT MEMORY MANAGEMENT FOR THE POST-MOORE’S ERA (opens in a new tab) Download file for BREAKING MEMORY WALLS: EFFICIENT MEMORY MANAGEMENT FOR THE POST-MOORE’S ERA (opens in a new tab)

  8. A Bit-Exact Hierarchical SystemC Model of pPIM for Verification and Performance Modeling

    The semiconductor industry has seen tremendous advancement lately and is expected to continue to grow since multiple industries like automobile, health care, meteorology, etc are developing dedicated chips with advancements in Artificial Intelligence (AI). These improvements also come with …

    rit Repository record for A Bit-Exact Hierarchical SystemC Model of pPIM for Verification and Performance Modeling (opens in a new tab)

  9. Implementation and Evaluation of Deep Neural Networks in Commercially Available Processing in Memory Hardware

    … computations. Therefore, data-centric computing paradigms, such as Processing in Memory (PIM), are being widely explored for CNN acceleration applications. A recent PIM architecture, developed and commercialized by the UPMEM company, has demonstrated impressive performance boost over …

    rit Repository record for Implementation and Evaluation of Deep Neural Networks in Commercially Available Processing in Memory Hardware (opens in a new tab)

  10. Energy-aware DNN Quantization for Processing-In-Memory Architecture

    With increasing computational cost of deep neural network (DNN), many efforts to develop energy-efficient intelligent system have been proposed from dedicated hardware platforms to model compression algorithms. Recently, hardware-aware quantization algorithms have shown further improvement in the …

    gatech Repository record for Energy-aware DNN Quantization for Processing-In-Memory Architecture (opens in a new tab)

  11. Reliable processing-in-memory

    Processing-in-memory (PIM) architectures integrate compute units within memory, enhancing performance and efficiency but introducing significant reliability challenges. These challenges arise from the inherent conflict between localized data accesses, which minimize data movement, and the necessity …

    texas Repository record for Reliable processing-in-memory (opens in a new tab)

  12. Processing-in-Memory based CNN Acceleration: Application Characterization and Simulation

    … neural networks (CNN) are widely used in different machine learning tasks, especially computer vision. Large amounts of computation in CNN cause intensive data movement, which makes traditional compute-centric architectures (CPU and GPU) less capable. There is a demand for new …

    chalmers Repository record for Processing-in-Memory based CNN Acceleration: Application Characterization and Simulation (opens in a new tab)

  13. Towards finding the right workload to process in memory

    In modern computing systems, the memory wall challenge poses a significant obstacle to performance optimization. This challenge refers to the widening gap between CPU processing speeds and memory access speeds, which has become increasingly pronounced with advancements in CPU technology. As CPUs …

    ubc Repository record for Towards finding the right workload to process in memory (opens in a new tab) Download PDF for Towards finding the right workload to process in memory (opens in a new tab)

  14. PIM-gem5 : a system simulator for Processing-in-Memory design space exploration

    Processing-in-Memory (PIM) has been recently revisited to address the issues of memory and power wall, mainly due to the maturity of 3D-stacking manufacturing technology and the increasing demand for bandwidth and parallel access in emerging data-centric applications. Recent studies have shown a …

    brazil-ufrgs Repository record for PIM-gem5 : a system simulator for Processing-in-Memory design space exploration (opens in a new tab)

  15. Towards Energy-Efficient In-Memory Computing Systems Using Electronic Design Automation

    The exponential growth in digital data has fueled the emergence of data-intensive applications like large language models, digital twins, and metaverse. These applications have unprecedented computing demands that exceed the capabilities of today's high-performance computing systems. To address …

    ucf Repository record for Towards Energy-Efficient In-Memory Computing Systems Using Electronic Design Automation (opens in a new tab)

  16. A Reconfigurable Architecture for Dense and Sparse Matrix Multiplication in Processing-in-Memory

    Traditional Von-Neumann computing architectures suffer from a fundamental limitation known as the memory wall, where performance is constrained by the cost of data movement between memory and processing units rather than by computational capability. This limitation is particularly pronounced in …

    rit Repository record for A Reconfigurable Architecture for Dense and Sparse Matrix Multiplication in Processing-in-Memory (opens in a new tab)

  17. Accelerator Architectures for Deep Learning and Graph Processing

    <p>Deep learning and graph processing are two big-data applications and they are widely applied in many domains. The training of deep learning is essential for inference and has not yet been fully studied. With data forward, error backward, and gradient calculation, deep learning training is a more …

    duke Repository record for Accelerator Architectures for Deep Learning and Graph Processing (opens in a new tab)

  18. Power-efficient Spiking Neuromorphic Designs using CMOS and Emerging Devices

    <p>The artificial intelligence (AI) algorithms have played critical roles in a variety of application scenarios in our daily life. The size of state-of-the-art large-scale AI models widely adopted in different domains have been proliferating to tens of billions of parameters. The dedicated AI …

    duke Repository record for Power-efficient Spiking Neuromorphic Designs using CMOS and Emerging Devices (opens in a new tab)

  19. Offloading embedding lookups to processing-in-memory for deep learning recommender models

    Recommender systems are an essential part of many industries and businesses. Generating accurate recommendations is critical for user engagement and business revenue. Currently, deep learning recommender models are commonly used, but they face challenges in processing and representing categorical …

    ubc Repository record for Offloading embedding lookups to processing-in-memory for deep learning recommender models (opens in a new tab) Download PDF for Offloading embedding lookups to processing-in-memory for deep learning recommender models (opens in a new tab)

  20. Code generation and adaptive control divergence management for light weight SIMT processors

    The energy costs of data movement are limiting the performance scaling of future generations of high performance computing architectures targeted to data intensive applications. The result has been a resurgence in the interest in processing-in-memory (PIM) architectures. This challenge has spawned …

    gatech Repository record for Code generation and adaptive control divergence management for light weight SIMT processors (opens in a new tab)

Page 1 of 3