Zheming Jin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Zheming Jin has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Jeffrey S. Vetter
Oak Ridge National Laboratory, Oak Ridge, TN, USA
6 shared publications - Hal Finkel
Argonne National Laboratory
4 shared publications - Keita Teranishi
Oak Ridge National Laboratory, Oak Ridge, USA
4 shared publications Hal Finkel
3 shared publications- Swaroop Pophale
Oak Ridge National Laboratory, Oak Ridge, USA
3 shared publications - Aaron Young
Oak Ridge National Laboratory, Oak Ridge, USA
2 shared publications
Latest Publications
- 24 November 2025
- Published by Springer Nature
- p. 615-625
Leveraging AI for Productive and Trustworthy HPC Software: Challenges and Research Directions
- 24 November 2025
- Published by Springer Nature
- p. 615-625
Leveraging AI for Productive and Trustworthy HPC Software: Challenges and Research Directions
- Open Access
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 524-533
Enhancing ChatPORT with CUDA-to-SYCL Kernel Translation Capability
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 524-533
Open AccessEnhancing ChatPORT with CUDA-to-SYCL Kernel Translation Capability
- Open Access
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 2129-2136
A Study of Performance Portability of Low-bit Fused Matrix-Vector Multiplication Kernels in SYCL
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 2129-2136
Open AccessA Study of Performance Portability of Low-bit Fused Matrix-Vector Multiplication Kernels in SYCL
- Open Access
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 458-474
ChatHPC: Building the Foundations for a Productive and Trustworthy AI-Assisted HPC Ecosystem
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 458-474
Open AccessChatHPC: Building the Foundations for a Productive and Trustworthy AI-Assisted HPC Ecosystem
- 1 November 2025
- Published by Springer Nature
- p. 33-45
Benchmarking Operators in Deep Neural Networks for Improving Performance Portability of SYCL
- 1 November 2025
- Published by Springer Nature
- p. 33-45
Benchmarking Operators in Deep Neural Networks for Improving Performance Portability of SYCL
- 29 September 2025
- Published by Springer Nature
- p. 197-211
ChatPORT: Fine-Tuned LLM for Easy Code {PORT}ing
- 29 September 2025
- Published by Springer Nature
- p. 197-211
ChatPORT: Fine-Tuned LLM for Easy Code {PORT}ing
- Open Access
- 1 July 2024
- Published by Office of Scientific and Technical Information (OSTI)
Evaluating Operators in Deep Neural Networks for Improving Performance Portability of SYCL
SYCL is a portable programming model for heterogeneous computing, so it is important to obtain reasonable performance portability of SYCL. Towards the goal of better understanding and improving performance portability of SYCL for machine learning workloads, we have been developing benchmarks for basic operators in deep neural networks (DNNs). These operators could be offloaded to heterogeneous computing devices such as graphics processing units (GPUs) to speed up computation. In this work, we introduce the benchmarks, evaluate the performance of the operators on GPU-based systems, and describe the causes of the performance gap between the SYCL and Compute Unified Device Architecture (CUDA) kernels. We find that the causes are related to the utilization of the texture cache for read-only data, optimization of the memory accesses with strength reduction, shared local memory accesses, and register usage per thread. We hope that the efforts of developing benchmarks for studying performance portability will stimulate discussion and interactions within the community.
- 1 July 2024
- Published by Office of Scientific and Technical Information (OSTI)
Open AccessEvaluating Operators in Deep Neural Networks for Improving Performance Portability of SYCL
SYCL is a portable programming model for heterogeneous computing, so it is important to obtain reasonable performance portability of SYCL. Towards the goal of better understanding and improving performance portability of SYCL for machine learning workloads, we have been developing benchmarks for basic operators in deep neural networks (DNNs). These operators could be offloaded to heterogeneous computing devices such as graphics processing units (GPUs) to speed up computation. In this work, we introduce the benchmarks, evaluate the performance of the operators on GPU-based systems, and describe the causes of the performance gap between the SYCL and Compute Unified Device Architecture (CUDA) kernels. We find that the causes are related to the utilization of the texture cache for read-only data, optimization of the memory accesses with strength reduction, shared local memory accesses, and register usage per thread. We hope that the efforts of developing benchmarks for studying performance portability will stimulate discussion and interactions within the community.
- Open Access
- 7 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 30-39
Experience Deploying Graph Applications on GPUs with SYCL
SYCL allows for deployment and use of accelerators across vendors’ platforms. In this work, we describe the experience of deploying graph analytics on vendors’ GPUs using SYCL. We contrast the CUDA and SYCL application programming interfaces by describing the experience of migrating the applications from CUDA to SYCL, evaluate the performance of the applications on NVIDIA and AMD GPUs, and explore performance improvement with device-level parallelism. The results show that the recent SYCL extensions facilitate functional portability, but improving code optimizations and resource usage for performance portability is needed in the compiler implementation.
- 7 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 30-39
Open AccessExperience Deploying Graph Applications on GPUs with SYCL
SYCL allows for deployment and use of accelerators across vendors’ platforms. In this work, we describe the experience of deploying graph analytics on vendors’ GPUs using SYCL. We contrast the CUDA and SYCL application programming interfaces by describing the experience of migrating the applications from CUDA to SYCL, evaluate the performance of the applications on NVIDIA and AMD GPUs, and explore performance improvement with device-level parallelism. The results show that the recent SYCL extensions facilitate functional portability, but improving code optimizations and resource usage for performance portability is needed in the compiler implementation.
- Open Access
- 25 February 2023
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Understanding Portability of Automotive Workload: A Case Study with the Points-to-Image Kernel in SYCL on Heterogeneous Computing Platforms
SYCL is a promising programming model for heterogenous computing across vendors’ devices. In this paper, we study whether SYCL can be applied to an automotive workload and its portability on heterogeneous computing platforms. We explain the automotive benchmark, describe our implementations and optimizations of the benchmark, and evaluate the performance of the benchmarks using SYCL and other programming models on heterogeneous computing devices. The study also allows us to have a better understanding of portability of the benchmark across these platforms.
- 25 February 2023
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Open AccessUnderstanding Portability of Automotive Workload: A Case Study with the Points-to-Image Kernel in SYCL on Heterogeneous Computing Platforms
SYCL is a promising programming model for heterogenous computing across vendors’ devices. In this paper, we study whether SYCL can be applied to an automotive workload and its portability on heterogeneous computing platforms. We explain the automotive benchmark, describe our implementations and optimizations of the benchmark, and evaluate the performance of the benchmarks using SYCL and other programming models on heterogeneous computing devices. The study also allows us to have a better understanding of portability of the benchmark across these platforms.
- Open Access
- 7 August 2022
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Performance portability study of epistasis detection using SYCL on NVIDIA GPU
We describe the experience of converting a CUDA implementation of a high-order epistasis detection algorithm to SYCL. The goals are for our work to be useful to application and compiler developers with a detailed description of migration paths between CUDA and SYCL. Evaluating the CUDA and SYCL applications on an NVIDIA V100 GPU, we find that the optimization of loop unrolling needs to be applied manually to the SYCL kernel for obtaining comparable performance. The performance of the SYCL group reduce function, an alternative to the CUDA warp-based reduction, depends on the problem and work group sizes. The 64-bit popcount operation implemented with tree of adders is slightly faster than the built-in popcount operation. When the number of OpenMP threads is four, the highest performance of the SYCL and CUDA applications are comparable.
- 7 August 2022
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessPerformance portability study of epistasis detection using SYCL on NVIDIA GPU
We describe the experience of converting a CUDA implementation of a high-order epistasis detection algorithm to SYCL. The goals are for our work to be useful to application and compiler developers with a detailed description of migration paths between CUDA and SYCL. Evaluating the CUDA and SYCL applications on an NVIDIA V100 GPU, we find that the optimization of loop unrolling needs to be applied manually to the SYCL kernel for obtaining comparable performance. The performance of the SYCL group reduce function, an alternative to the CUDA warp-based reduction, depends on the problem and work group sizes. The 64-bit popcount operation implemented with tree of adders is slightly faster than the built-in popcount operation. When the number of OpenMP threads is four, the highest performance of the SYCL and CUDA applications are comparable.