Hsiang-Pang Li
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
Research Subjects
Hsiang-Pang Li has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Tei-Wei Kuo
National Taiwan University
2 shared publications - Yuan-Hao Chang
National Taiwan University
2 shared publications - Chia-Chun Chien
National Taiwan University, Taipei, Taiwan
1 shared publication - Chia-Lin Yang
Department of Computer Science and Information Engineering, National Taiwan University, Taipei, Taiwan
1 shared publication - Chih-Yuan Lu
Macronix International Company Ltd., Hsinchu, Taiwan
1 shared publication - Hang-Ting Lue
Emerging Central Lab., Macronix Int. Co., Ltd., Hsinchu
1 shared publication
Latest Publications
- Open Access
- 20 January 2025
- Published by Association for Computing Machinery (ACM)
- p. 1008-1015
In-Storage Read-Centric Seed Location Filtering Using 3D-NAND Flash for Genome Sequence Analysis
Read mapping is a critical bottleneck in genome sequence analysis, requiring costly approximate string matching to identify potential matches between reads and a reference genome. Pre-alignment filtering methods aim to mitigate this issue by filtering out unnecessary mapping locations, and implementing them with processing-in-memory (PIM) approaches offers potential benefits by offloading filtering from the computing unit. However, the sparse number of potential mapping locations for each read limits the utilization of PIM's parallel computing capabilities, thereby hindering the overlapping of filtering and sequence alignment to hide filtering latency overheads. In this paper, we propose a 3D NAND-based in-storage pre-alignment filtering approach. Leveraging the read depth property, we introduce a read-centric pre-alignment filtering method that enables parallel comparison of multiple reads. We co-design software and hardware for in-situ processing of read-centric pre-alignment filtering within the storage, capitalizing on 3D NAND Flash's approximate parallel search capability. When integrating with a representative read mapping accelerator, our design achieves an average 1.36x performance improvement with comparable energy consumption. Compared to the state-of-the-art (SOTA) PIM solution, our design results in 123.8x and 53.3x performance gain and energy efficiency improvement.
- 20 January 2025
- Published by Association for Computing Machinery (ACM)
- p. 1008-1015
Open AccessIn-Storage Read-Centric Seed Location Filtering Using 3D-NAND Flash for Genome Sequence Analysis
Read mapping is a critical bottleneck in genome sequence analysis, requiring costly approximate string matching to identify potential matches between reads and a reference genome. Pre-alignment filtering methods aim to mitigate this issue by filtering out unnecessary mapping locations, and implementing them with processing-in-memory (PIM) approaches offers potential benefits by offloading filtering from the computing unit. However, the sparse number of potential mapping locations for each read limits the utilization of PIM's parallel computing capabilities, thereby hindering the overlapping of filtering and sequence alignment to hide filtering latency overheads. In this paper, we propose a 3D NAND-based in-storage pre-alignment filtering approach. Leveraging the read depth property, we introduce a read-centric pre-alignment filtering method that enables parallel comparison of multiple reads. We co-design software and hardware for in-situ processing of read-centric pre-alignment filtering within the storage, capitalizing on 3D NAND Flash's approximate parallel search capability. When integrating with a representative read mapping accelerator, our design achieves an average 1.36x performance improvement with comparable energy consumption. Compared to the state-of-the-art (SOTA) PIM solution, our design results in 123.8x and 53.3x performance gain and energy efficiency improvement.
- Open Access
- 5 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
LUTIN: Efficient Neural Network Inference with Table Lookup
DNN models are becoming increasingly large and complex, but they are also being deployed on commodity devices that require low power and latency but lack specialized accelerators. We introduce LUTIN (LUT-based INference), which reduces the amount of matrix multiplication in DNN inference by converting it into table lookups. LUTIN's innovation is its use of hyperparameter optimization to refine the quantization process and vector partitioning, allowing it to run efficiently on a variety of hardware. By reducing off-chip memory lookups and designing a cache-efficient data layout, LUTIN reduces energy consumption while increasing the use of available CPU cache, even on devices with limited processing power. Our approach goes beyond the traditional limitations of 8-bit quantization, investigating lower bit-widths to further reduce LUT size while meeting accuracy requirements. Experimental results show that LUTIN achieves up to a 2.34x speedup in latency and a 2.04x improvement in energy efficiency over full-precision models.
- 5 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessLUTIN: Efficient Neural Network Inference with Table Lookup
DNN models are becoming increasingly large and complex, but they are also being deployed on commodity devices that require low power and latency but lack specialized accelerators. We introduce LUTIN (LUT-based INference), which reduces the amount of matrix multiplication in DNN inference by converting it into table lookups. LUTIN's innovation is its use of hyperparameter optimization to refine the quantization process and vector partitioning, allowing it to run efficiently on a variety of hardware. By reducing off-chip memory lookups and designing a cache-efficient data layout, LUTIN reduces energy consumption while increasing the use of available CPU cache, even on devices with limited processing power. Our approach goes beyond the traditional limitations of 8-bit quantization, investigating lower bit-widths to further reduce LUT size while meeting accuracy requirements. Experimental results show that LUTIN achieves up to a 2.34x speedup in latency and a 2.04x improvement in energy efficiency over full-precision models.
- 1 May 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
System Performance Analysis of Bit-Alterable 3D NAND Flash Devices for High-Performance Solid-State Drive (SSD) Applications
NAND Flash is not bit-alterable in the erase operation, and the erase is often carried out in a very large block size. This deteriorates the overall system performances. In this work, we propose a novel bit-alterable 3D NAND Flash that can overcome the drawback of large block size for erase. Such bit-alterable 3D NAND device is possible by adopting novel dual-channel floating-body 3D NAND devices, where both +FN programming and -FN erasing are selectable with suitable inhibit methods. Because erase can be carried out in a much smaller page unit rather than a large block, our evaluation indicates significant improvements in garbage collection (GC) and effective programming time. Moreover, the write amplification effect is reduced so that the data life time is expanded.
- 1 May 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
System Performance Analysis of Bit-Alterable 3D NAND Flash Devices for High-Performance Solid-State Drive (SSD) Applications
NAND Flash is not bit-alterable in the erase operation, and the erase is often carried out in a very large block size. This deteriorates the overall system performances. In this work, we propose a novel bit-alterable 3D NAND Flash that can overcome the drawback of large block size for erase. Such bit-alterable 3D NAND device is possible by adopting novel dual-channel floating-body 3D NAND devices, where both +FN programming and -FN erasing are selectable with suitable inhibit methods. Because erase can be carried out in a much smaller page unit rather than a large block, our evaluation indicates significant improvements in garbage collection (GC) and effective programming time. Moreover, the write amplification effect is reduced so that the data life time is expanded.
- 1 October 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1662-1667
Channel-aware construction of polar codes
The conventional construction of polar codes is designed in advance for a specific channel, the performance of polar codes deteriorates if the construction does not match the current channel. In other words, we suffer from performance loss when either a time-variant channel or an inaccurate estimation of channel are considered. To put polar codes into practice, a channel-aware construction of polar codes, which optimizes the construction on the fly, is proposed in this paper. Unlike universal polar codes, the receiver of the proposed scheme does not need to know any knowledge of actual channel. In addition, the proposed algorithm constructs polar codes only based on decoding results, it is well suited to not only i.i.d. channels but general channels. Simulation results show that the proposed schemes over a time-variant parallel channel (i.e. multi-channel) are always achieved a relatively good performance.
- 1 October 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1662-1667
Channel-aware construction of polar codes
The conventional construction of polar codes is designed in advance for a specific channel, the performance of polar codes deteriorates if the construction does not match the current channel. In other words, we suffer from performance loss when either a time-variant channel or an inaccurate estimation of channel are considered. To put polar codes into practice, a channel-aware construction of polar codes, which optimizes the construction on the fly, is proposed in this paper. Unlike universal polar codes, the receiver of the proposed scheme does not need to know any knowledge of actual channel. In addition, the proposed algorithm constructs polar codes only based on decoding results, it is well suited to not only i.i.d. channels but general channels. Simulation results show that the proposed schemes over a time-variant parallel channel (i.e. multi-channel) are always achieved a relatively good performance.
- Open Access
- 7 June 2015
- Published by Association for Computing Machinery (ACM)
- Vol. 5, 1-6
Achieving SLC performance with MLC flash memory
Although the Multi-Level-Cell technique is widely adopted by flash-memory vendors to boost the chip density and to lower the cost, it results in serious performance and reliability problems. Different from the past work, a new cell programming method is proposed to not only significantly improve the chip performance but also reduce the potential bit error rate. In particular, a Single-Level-Cell-like programming style is proposed to better explore the threshold-voltage relationship to denote different Multi-Level-Cell bit information, which in turn drastically provides a larger window of threshold voltage similar to that found in Single-Level-Cell chips. It could result in less programming iterations and simultaneously a much less reliability problem in programming flash-memory cells. In the experiments, the new programming style could accelerate the programming speed up to 742% and even reduce the bit error rate up to 471% for Multi-Level-Cell pages.
- 7 June 2015
- Published by Association for Computing Machinery (ACM)
- Vol. 5, 1-6
Open AccessAchieving SLC performance with MLC flash memory
Although the Multi-Level-Cell technique is widely adopted by flash-memory vendors to boost the chip density and to lower the cost, it results in serious performance and reliability problems. Different from the past work, a new cell programming method is proposed to not only significantly improve the chip performance but also reduce the potential bit error rate. In particular, a Single-Level-Cell-like programming style is proposed to better explore the threshold-voltage relationship to denote different Multi-Level-Cell bit information, which in turn drastically provides a larger window of threshold voltage similar to that found in Single-Level-Cell chips. It could result in less programming iterations and simultaneously a much less reliability problem in programming flash-memory cells. In the experiments, the new programming style could accelerate the programming speed up to 742% and even reduce the bit error rate up to 471% for Multi-Level-Cell pages.