Daniel Waddington
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Daniel Waddington has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Sarel Cohen
Reichman University, Herzliya, Israel
5 shared publications - Moshik Hershcovitch
IBM Research
4 shared publications - Andrew Wood
Department of Accounting and Financial Management, University of Portsmouth, Portsmouth, United Kingdom
2 shared publications - Dalit Naor
Academic College Tel Aviv-Yaffo, Tel Aviv, Israel
2 shared publications - Daniel Cohen
The Academic College of Tel, Aviv-Yaffo, Israel
2 shared publications - Moshik Hershcovitch
IBM Research
2 shared publications
Latest Publications
- Open Access
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 30-45
DPUF: DPU-accelerated Near-storage Secure Filtering
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 30-45
Open AccessDPUF: DPU-accelerated Near-storage Secure Filtering
- Open Access
Speeding up Model Loading with fastsafetensors
The rapid increases in model parameter sizes introduces new challenges in pre-trained model loading. Currently, machine learning code often deserializes each parameter as a tensor object in host memory before copying it to device memory. We found that this approach underutilized storage throughput and significantly slowed down loading large models with a widely-used model file formats, safetensors. In this work, we present fastsafetensors, a Python library designed to optimize the deserialization of tensors in safetensors files. Our approach first copies groups of on-disk parameters to device memory, where they are directly instantiated as tensor objects. This design enables further optimization in low-level I/O and high-level tensor preprocessing, including parallelized copying, peer-to-peer DMA, and GPU offloading. Experimental results show performance improvements of 4.8x to 7.5x in loading models such as Llama (7, 13, and 70 billion parameters), Falcon (40 billion parameters), and the Bloom (176 billion parameters).
Open AccessSpeeding up Model Loading with fastsafetensors
The rapid increases in model parameter sizes introduces new challenges in pre-trained model loading. Currently, machine learning code often deserializes each parameter as a tensor object in host memory before copying it to device memory. We found that this approach underutilized storage throughput and significantly slowed down loading large models with a widely-used model file formats, safetensors. In this work, we present fastsafetensors, a Python library designed to optimize the deserialization of tensors in safetensors files. Our approach first copies groups of on-disk parameters to device memory, where they are directly instantiated as tensor objects. This design enables further optimization in low-level I/O and high-level tensor preprocessing, including parallelized copying, peer-to-peer DMA, and GPU offloading. Experimental results show performance improvements of 4.8x to 7.5x in loading models such as Llama (7, 13, and 70 billion parameters), Falcon (40 billion parameters), and the Bloom (176 billion parameters).
- Open Access
- 8 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 8-14
Dictionary Based Cache Line Compression
Active-standby mechanisms for VM high-availability demand frequent synchronization of memory and CPU state, involving the identification and transfer of "dirty" memory pages to a standby target. Building upon the granularity offered by CXL-enabled memory devices, as discussed by Waddington et al. [21], this paper proposes a dictionary-based compression method operating on 64-byte cache lines to minimize snapshot volume and synchronization latency. The method aims to transmit only necessary information required to reconstruct the memory state at the standby machine, augmented by byte grouping and cache-line partitioning techniques. We assess the compression benefits on memory access patterns across 20 benchmarks snapshots and compare our approach to standard off-the-shelf compression methods. Our findings reveal significant improvements across nearly all benchmarks, with some experiencing over a twofold enhancement compared to standard compression, while others show more moderate gains. We conduct an in-depth experimental analysis on the contribution of each method and examine the nature of the benchmarks. We ascertain that the repeating nature of cache lines across snapshots (caused by transient memory changes) and their concise representation contributes most to the size reduction, accounting for 92% of the gains. Our work paves the way for further reduction in the data transferred to standby machines, thereby enhancing VM high-availability and reducing synchronization latency.
- 8 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 8-14
Open AccessDictionary Based Cache Line Compression
Active-standby mechanisms for VM high-availability demand frequent synchronization of memory and CPU state, involving the identification and transfer of "dirty" memory pages to a standby target. Building upon the granularity offered by CXL-enabled memory devices, as discussed by Waddington et al. [21], this paper proposes a dictionary-based compression method operating on 64-byte cache lines to minimize snapshot volume and synchronization latency. The method aims to transmit only necessary information required to reconstruct the memory state at the standby machine, augmented by byte grouping and cache-line partitioning techniques. We assess the compression benefits on memory access patterns across 20 benchmarks snapshots and compare our approach to standard off-the-shelf compression methods. Our findings reveal significant improvements across nearly all benchmarks, with some experiencing over a twofold enhancement compared to standard compression, while others show more moderate gains. We conduct an in-depth experimental analysis on the contribution of each method and examine the nature of the benchmarks. We ascertain that the repeating nature of cache lines across snapshots (caused by transient memory changes) and their concise representation contributes most to the size reduction, accounting for 92% of the gains. Our work paves the way for further reduction in the data transferred to standby machines, thereby enhancing VM high-availability and reducing synchronization latency.
- Open Access
- 5 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 145
Cache Line Deltas Compression
Synchronization of replicated data and program state is an essential aspect of application fault-tolerance. Current solutions use virtual memory mapping to identify page writes and replicate them at the destination. This approach has limitations because the granularity is restricted to a minimum of 4KiB per page, which may result in more data being replicated. Motivated by the emerging CXL hardware, we expand on the work Waddington, et al. [SoCC 22] by evaluating popular compression algorithms on VM snapshot data at cache line granularity. We measure the compression ratio vs. the compression time and present our conclusions.
- 5 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 145
Open AccessCache Line Deltas Compression
Synchronization of replicated data and program state is an essential aspect of application fault-tolerance. Current solutions use virtual memory mapping to identify page writes and replicate them at the destination. This approach has limitations because the granularity is restricted to a minimum of 4KiB per page, which may result in more data being replicated. Motivated by the emerging CXL hardware, we expand on the work Waddington, et al. [SoCC 22] by evaluating popular compression algorithms on VM snapshot data at cache line granularity. We measure the compression ratio vs. the compression time and present our conclusions.
- 11 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Micro
- Vol. 43 (2) , 58-65
CXL-Enabled Enhanced Memory Functions
The arrival of the Compute Express Link (CXL) protocol is a significant milestone for the systems community. CXL provides a standardized, cache-coherent memory protocol that can be used to attach devices and memory to a system, while maintaining memory coherency with the host processor. CXL enables accelerators (e.g., GPU, DPU) to both have direct load/store access to the host memory and the ability to make their own on-device memory likewise accessible to the host CPU. Because CXL allows technology interposition on the memory data plane, it opens up the possibility of “pushing down” functions into the memory subsystem. In this paper, we introduce the concept of Enhanced Memory Functions (EMFs). We then describe two use-cases, one prototyped using an FPGA-based Intelligent Memory Controller (IMC) platform. Finally, we show initial experimental results indicating that EMFs could present valuable solutions to problems that are difficult to solve within existing computer architectures.
- 11 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Micro
- Vol. 43 (2) , 58-65
CXL-Enabled Enhanced Memory Functions
The arrival of the Compute Express Link (CXL) protocol is a significant milestone for the systems community. CXL provides a standardized, cache-coherent memory protocol that can be used to attach devices and memory to a system, while maintaining memory coherency with the host processor. CXL enables accelerators (e.g., GPU, DPU) to both have direct load/store access to the host memory and the ability to make their own on-device memory likewise accessible to the host CPU. Because CXL allows technology interposition on the memory data plane, it opens up the possibility of “pushing down” functions into the memory subsystem. In this paper, we introduce the concept of Enhanced Memory Functions (EMFs). We then describe two use-cases, one prototyped using an FPGA-based Intelligent Memory Controller (IMC) platform. Finally, we show initial experimental results indicating that EMFs could present valuable solutions to problems that are difficult to solve within existing computer architectures.
- 10 June 2022
- Published by Springer Nature
- p. 369-382
Biologically Plausible Complex-Valued Neural Networks and Model Optimization
- 10 June 2022
- Published by Springer Nature
- p. 369-382
Biologically Plausible Complex-Valued Neural Networks and Model Optimization
- Open Access
Non-Volatile Memory Accelerated Geometric Multi-Scale Resolution Analysis
Dimensionality reduction algorithms are standard tools in a researcher's toolbox. Dimensionality reduction algorithms are frequently used to augment downstream tasks such as machine learning, data science, and also are exploratory methods for understanding complex phenomena. For instance, dimensionality reduction is commonly used in Biology as well as Neuroscience to understand data collected from biological subjects. However, dimensionality reduction techniques are limited by the von-Neumann architectures that they execute on. Specifically, data intensive algorithms such as dimensionality reduction techniques often require fast, high capacity, persistent memory which historically hardware has been unable to provide at the same time. In this paper, we present a re-implementation of an existing dimensionality reduction technique called Geometric Multi-Scale Resolution Analysis (GMRA) which has been accelerated via novel persistent memory technology called Memory Centric Active Storage (MCAS). Our implementation uses a specialized version of MCAS called PyMM that provides native support for Python datatypes including NumPy arrays and PyTorch tensors. We compare our PyMM implementation against a DRAM implementation, and show that when data fits in DRAM, PyMM offers competitive runtimes. When data does not fit in DRAM, our PyMM implementation is still able to process the data.
Open AccessNon-Volatile Memory Accelerated Geometric Multi-Scale Resolution Analysis
Dimensionality reduction algorithms are standard tools in a researcher's toolbox. Dimensionality reduction algorithms are frequently used to augment downstream tasks such as machine learning, data science, and also are exploratory methods for understanding complex phenomena. For instance, dimensionality reduction is commonly used in Biology as well as Neuroscience to understand data collected from biological subjects. However, dimensionality reduction techniques are limited by the von-Neumann architectures that they execute on. Specifically, data intensive algorithms such as dimensionality reduction techniques often require fast, high capacity, persistent memory which historically hardware has been unable to provide at the same time. In this paper, we present a re-implementation of an existing dimensionality reduction technique called Geometric Multi-Scale Resolution Analysis (GMRA) which has been accelerated via novel persistent memory technology called Memory Centric Active Storage (MCAS). Our implementation uses a specialized version of MCAS called PyMM that provides native support for Python datatypes including NumPy arrays and PyTorch tensors. We compare our PyMM implementation against a DRAM implementation, and show that when data fits in DRAM, PyMM offers competitive runtimes. When data does not fit in DRAM, our PyMM implementation is still able to process the data.
- Open Access
Non-Volatile Memory Accelerated Posterior Estimation
Bayesian inference allows machine learning models to express uncertainty. Current machine learning models use only a single learnable parameter combination when making predictions, and as a result are highly overconfident when their predictions are wrong. To use more learnable parameter combinations efficiently, these samples must be drawn from the posterior distribution. Unfortunately computing the posterior directly is infeasible, so often researchers approximate it with a well known distribution such as a Gaussian. In this paper, we show that through the use of high-capacity persistent storage, models whose posterior distribution was too big to approximate are now feasible, leading to improved predictions in downstream tasks.
Open AccessNon-Volatile Memory Accelerated Posterior Estimation
Bayesian inference allows machine learning models to express uncertainty. Current machine learning models use only a single learnable parameter combination when making predictions, and as a result are highly overconfident when their predictions are wrong. To use more learnable parameter combinations efficiently, these samples must be drawn from the posterior distribution. Unfortunately computing the posterior directly is infeasible, so often researchers approximate it with a well known distribution such as a Gaussian. In this paper, we show that through the use of high-capacity persistent storage, models whose posterior distribution was too big to approximate are now feasible, leading to improved predictions in downstream tasks.
- Open Access
An Architecture for Memory Centric Active Storage (MCAS)
The advent of CPU-attached persistent memory technology, such as Intel's Optane Persistent Memory Modules (PMM), has brought with it new opportunities for storage. In 2018, IBM Research Almaden began investigating and developing a new enterprise-grade storage solution directly aimed at this emerging technology. MCAS (Memory Centric Active Storage) defines an evolved network-attached key-value store that offers both near-data compute and the ability to layer enterprise-grade data management services on shared persistent memory. As a converged memory-storage tier, MCAS moves towards eliminating the traditional separation of compute and storage, and thereby unifying the data space. This paper provides an in-depth review of the MCAS architecture and implementation, as well as general performance results.
Open AccessAn Architecture for Memory Centric Active Storage (MCAS)
The advent of CPU-attached persistent memory technology, such as Intel's Optane Persistent Memory Modules (PMM), has brought with it new opportunities for storage. In 2018, IBM Research Almaden began investigating and developing a new enterprise-grade storage solution directly aimed at this emerging technology. MCAS (Memory Centric Active Storage) defines an evolved network-attached key-value store that offers both near-data compute and the ability to layer enterprise-grade data management services on shared persistent memory. As a converged memory-storage tier, MCAS moves towards eliminating the traditional separation of compute and storage, and thereby unifying the data space. This paper provides an in-depth review of the MCAS architecture and implementation, as well as general performance results.