%0 Conference Proceedings
%T Efficient Management for Hybrid Memory in Managed Language Runtime
%+ Institute of Computing Technology [Beijing] (ICT)
%+ University of Chinese Academy of Sciences [Beijing] (UCAS)
%+ The University of Sydney
%+ State Key Laboratory of Mathematical Engineering and Advanced Computing
%A Wang, Chenxi
%A Cao, Ting
%A Zigman, John
%A Lv, Fang
%A Zhang, Yunquan
%A Feng, Xiaobing
%Z Part 1: Memory: Non-Volatile, Solid State Drives, Hybrid Systems
%< avec comité de lecture
%( Lecture Notes in Computer Science
%B 13th IFIP International Conference on Network and Parallel Computing (NPC)
%C Xi'an, China
%Y Guang R. Gao
%Y Depei Qian
%Y Xinbo Gao
%Y Barbara Chapman
%Y Wenguang Chen
%I Springer International Publishing
%3 Network and Parallel Computing
%V LNCS-9966
%P 29-42
%8 2016-10-28
%D 2016
%R 10.1007/978-3-319-47099-3_3
%K Hybrid memory
%K Managed runtime
%K JVM
%K Memory management
%Z Computer Science [cs]Conference papers
%X Hybrid memory, which leverages the benefits of traditional DRAM and emerging memory technologies, is a promising alternative for future main memory design. However popular management policies through memory-access recording and page migration may invoke non-trivial overhead in execution time and hardware space. Nowadays, managed language applications are increasingly dominant in every kind of platform. Managed runtimes provide services for automatic memory management. So it is important to adapt them for the underlying hybrid memory.This paper explores two opportunities, heap partition placement and object promotion, inside managed runtimes for allocating hot data in a fast memory space (fast-space) without any access recording or data migration overhead. For heap partition placement, we quantitatively analyze LLC miss density and performance effect for each partition. Results show that LLC misses especially store misses mostly hit nursery partitions. Placing nursery in fast-space, which is 20 % total memory footprint of tested benchmarks on average, causes only 10 % performance difference from all memory footprint in fast-space. During object promotion, hot objects will be directly allocated to fast-space. We develop a tool to analyze the LLC miss density for each method of workloads, since we have found that the LLC misses are mostly triggered by a small percentage of the total set of methods. The objects visited by the top-ranked methods are recognized as hot. Results show that hot objects do have higher access density, more than 3 times of random distribution for SPECjbb and pmd, and placing them in fast-space further reduces their execution time by 6 % and 13 % respectively.
%G English
%Z TC 10
%Z WG 10.3
%2 https://inria.hal.science/hal-01647992/document
%2 https://inria.hal.science/hal-01647992/file/432484_1_En_3_Chapter.pdf
%L hal-01647992
%U https://inria.hal.science/hal-01647992
%~ IFIP-LNCS
%~ IFIP
%~ IFIP-TC
%~ IFIP-TC10
%~ IFIP-NPC
%~ IFIP-WG10-3
%~ IFIP-LNCS-9966