TY - JOUR
T1 - Leveraging Write Heterogeneity of Phase Change Memory on Supporting Self-Balancing Binary Tree
AU - Chang, Che Wei
AU - Wu, Chun Feng
AU - Chang, Yuan Hao
AU - Yang, Ming Chang
AU - Chang, Chieh Fu
N1 - Publisher Copyright:
© 1982-2012 IEEE.
PY - 2022/6/1
Y1 - 2022/6/1
N2 - With the increasing demand of massive/big data applications, nonvolatile memory (NVM), such as phase-change memory (PCM), has become a promising candidate to replace DRAM because of its low leakage power, nonvolatility, and high density. However, most of the existing memory read/write intensive algorithms and data structures are not aware of the PCM write heterogeneity in terms of both energy consumption and latency. In particular, self-balancing binary search trees, which are widely used to manage massive data in the big-data era, were designed without the consideration of PCM characteristics. Thus, the multiple rotations of the tree balancing process would degrade the memory performance. This work explores the relations among nodes and analyzes tree operations, and the node indexing and address mapping are redesigned to reduce the tree management overhead on single-level cell (SLC) PCM by decreasing the number of bit flips of tree rotations. When multilevel cell (MLC) PCM is included, our address mapping algorithm is developed to reduce the total energy consumption and latency with considerations of the heterogeneous write operations of different cell states. Experimental results show that our solution significantly outperforms the original implementation of a self-balancing binary search tree when the amount of data is large.
AB - With the increasing demand of massive/big data applications, nonvolatile memory (NVM), such as phase-change memory (PCM), has become a promising candidate to replace DRAM because of its low leakage power, nonvolatility, and high density. However, most of the existing memory read/write intensive algorithms and data structures are not aware of the PCM write heterogeneity in terms of both energy consumption and latency. In particular, self-balancing binary search trees, which are widely used to manage massive data in the big-data era, were designed without the consideration of PCM characteristics. Thus, the multiple rotations of the tree balancing process would degrade the memory performance. This work explores the relations among nodes and analyzes tree operations, and the node indexing and address mapping are redesigned to reduce the tree management overhead on single-level cell (SLC) PCM by decreasing the number of bit flips of tree rotations. When multilevel cell (MLC) PCM is included, our address mapping algorithm is developed to reduce the total energy consumption and latency with considerations of the heterogeneous write operations of different cell states. Experimental results show that our solution significantly outperforms the original implementation of a self-balancing binary search tree when the amount of data is large.
KW - Energy saving
KW - multilevel cell (MLC)
KW - nonvolatile memory (NVM)
KW - phase-change memory (PCM)
KW - tree data structure
KW - write heterogeneity
UR - http://www.scopus.com/inward/record.url?scp=85110819627&partnerID=8YFLogxK
U2 - 10.1109/TCAD.2021.3097298
DO - 10.1109/TCAD.2021.3097298
M3 - 文章
AN - SCOPUS:85110819627
SN - 0278-0070
VL - 41
SP - 1757
EP - 1770
JO - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
JF - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
IS - 6
ER -