Daon Park
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Daon Park has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Bernhard Egger
Seoul National University, Seoul, Republic of Korea
5 shared publications - Changyeon Jo
Seoul National University
1 shared publication - Hyunsoo Kim
Seoul National University, Seoul, Republic of Korea
1 shared publication - Jaume Mateu Cuadrat
Computer Science and Engineering, Seoul National University
1 shared publication - Sungbin Jo
Seoul National University, Seoul, Republic of Korea
1 shared publication - Youngsu Cho
Seoul National University
1 shared publication
Latest Publications
- Open Access
- 25 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Architecture and Code Optimization
- Vol. 23 (2) , 1-25
CPU-GPU Workload Distribution during Throughput-Oriented LLM Inference on Single-GPU Systems
Large language models (LLMs) have recently achieved remarkable performance in text generation, capturing the attention of a broad audience. This success, driven by the rapid growth in model parameters, comes at the expense of significantly higher operational costs and decreased processing speed. These costs, combined with privacy concerns around cloud-based deployments, have motivated research into running LLMs on commodity hardware. For example, researchers have used the memory hierarchy to boost throughput by increasing the number of batches. These studies, however, tend to overlook or inefficiently utilize the additional computational resources provided by the CPU. In this work, we present a dynamic workload allocation technique that efficiently distributes computation across all available hardware resources. The proposed method targets decoder-based models on standard general-purpose hardware, effectively minimizing idle periods for both the CPU and the GPU. Experiments show that our approach achieves up to 30% higher throughput compared to the state of the art, regardless of model architecture, LLM optimizations, and input batch sizes.
- 25 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Architecture and Code Optimization
- Vol. 23 (2) , 1-25
Open AccessCPU-GPU Workload Distribution during Throughput-Oriented LLM Inference on Single-GPU Systems
Large language models (LLMs) have recently achieved remarkable performance in text generation, capturing the attention of a broad audience. This success, driven by the rapid growth in model parameters, comes at the expense of significantly higher operational costs and decreased processing speed. These costs, combined with privacy concerns around cloud-based deployments, have motivated research into running LLMs on commodity hardware. For example, researchers have used the memory hierarchy to boost throughput by increasing the number of batches. These studies, however, tend to overlook or inefficiently utilize the additional computational resources provided by the CPU. In this work, we present a dynamic workload allocation technique that efficiently distributes computation across all available hardware resources. The proposed method targets decoder-based models on standard general-purpose hardware, effectively minimizing idle periods for both the CPU and the GPU. Experiments show that our approach achieves up to 30% higher throughput compared to the state of the art, regardless of model architecture, LLM optimizations, and input batch sizes.
- Open Access
- 13 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 233-245
Improving Throughput-oriented LLM Inference with CPU Computations
Large language models (LLMs) have recently captured the attention of a broad audience. To a large part, their exceptional performance in text generation was made possible by an exponential growth of the model parameters. This growth, however, comes at the expense of significantly higher operational costs and a decreased processing speed. Recent research has focused on running LLMs on commodity hardware, for example, by employing the memory hierarchy to augment throughput by increasing the number of batches. These studies, however, tend to overlook or inefficiently utilize the additional computational resources provided by the CPU. In this paper, we introduce a technique capable of efficiently harnessing all available computational resources through a finely tuned and dynamic workload allocation approach. This technique applies to decoder-based models on standard general-purpose hardware, effectively minimizing idle periods for both the CPU and the GPU. We conducted experiments involving various large language models, each representing distinct decoder-based architectures. Compared to the state-of-the-art, the results demonstrate a potential for an increase of up to 105% in throughput with the OPT-30B model.
- 13 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 233-245
Open AccessImproving Throughput-oriented LLM Inference with CPU Computations
Large language models (LLMs) have recently captured the attention of a broad audience. To a large part, their exceptional performance in text generation was made possible by an exponential growth of the model parameters. This growth, however, comes at the expense of significantly higher operational costs and a decreased processing speed. Recent research has focused on running LLMs on commodity hardware, for example, by employing the memory hierarchy to augment throughput by increasing the number of batches. These studies, however, tend to overlook or inefficiently utilize the additional computational resources provided by the CPU. In this paper, we introduce a technique capable of efficiently harnessing all available computational resources through a finely tuned and dynamic workload allocation approach. This technique applies to decoder-based models on standard general-purpose hardware, effectively minimizing idle periods for both the CPU and the GPU. We conducted experiments involving various large language models, each representing distinct decoder-based architectures. Compared to the state-of-the-art, the results demonstrate a potential for an increase of up to 105% in throughput with the OPT-30B model.
- Open Access
- 24 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 37-42
Improving Throughput-oriented Generative Inference with CPUs
- 24 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 37-42
Open AccessImproving Throughput-oriented Generative Inference with CPUs
- 31 March 2023
- Published by Springer Nature
- p. 132-140
A Black-Box Graph Partitioner for Generalized Deep Neural Network Parallelization
- 31 March 2023
- Published by Springer Nature
- p. 132-140
A Black-Box Graph Partitioner for Generalized Deep Neural Network Parallelization
- 9 December 2021
- Published by Springer Nature
- p. 17-26
Can VM Live Migration Improve Job Throughput? Evidence from a Real World Cluster Trace
- 9 December 2021
- Published by Springer Nature
- p. 17-26
Can VM Live Migration Improve Job Throughput? Evidence from a Real World Cluster Trace