Forgive-TM: Supporting Lazy Conflict Detection In Eager Hardware Transactional Memory

被引:6
|
作者
Park, Sunjae [1 ]
Hughes, Christopher J. [2 ]
Prvulovic, Milos [1 ]
机构
[1] Georgia Inst Technol, Atlanta, GA 30332 USA
[2] Intel, Santa Clara, CA USA
关键词
Parallel processing; Multithreading; COHERENCE; VERIFICATION; MANAGEMENT;
D O I
10.1109/PACT.2019.00023
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Commercial hardware transactional memory (TM) systems commonly use coherence messages to detect data conflicts. When a core inside a transaction receives a coherence request for data, it uses this information to determine whether there was a data conflict. Inherent in this behavior is the fact that data conflicts are detected eagerly, i.e., as soon as possible, and even while both sides of the conflict are speculative. Although it has been shown that lazy conflict detection can lead to better performance, this approach precludes lazy detection. In this paper, we describe a mechanism that allows conventional hardware to support lazy conflict detection, while still keeping the coherence protocol intact. Under Forgive-TM, speculative writes are done immediately to a special buffer, without first obtaining global write permission. The write permission is acquired later, when the transaction is about to commit. In other words, it "acts first, and asks forgiveness later." This effectively allows conflict detection to be done lazily. Using this scheme, ForgiveTM is able to provide 19% overall performance improvement in STAMP.
引用
下载
收藏
页码:192 / 204
页数:13
相关论文
共 41 条
  • [31] Yet Another Waiting Mechanism based on Conflict Prediction for Hardware Transactional Memory
    Mashita, Keisuke
    Miyake, Sho
    Yamada, Ryohei
    Tsumura, Tomoaki
    PROCEEDINGS OF 2015 THIRD INTERNATIONAL SYMPOSIUM ON COMPUTING AND NETWORKING (CANDAR), 2015, : 400 - 403
  • [32] Two-Phase Conflict Detection for Transactional Memory on Clusters
    Wang, Ruibo
    Lu, Kai
    Lu, Xicheng
    2009 IEEE INTERNATIONAL CONFERENCE ON CLUSTER COMPUTING AND WORKSHOPS, 2009, : 574 - 576
  • [33] PleaseTM: Enabling Transaction Conflict Management in Requester-wins Hardware Transactional Memory
    Park, Sunjae
    Prvulovic, Milos
    Hughes, Christopher J.
    PROCEEDINGS OF THE 2016 IEEE INTERNATIONAL SYMPOSIUM ON HIGH-PERFORMANCE COMPUTER ARCHITECTURE (HPCA-22), 2016, : 285 - 296
  • [34] TxRace: Efficient Data Race Detection Using Commodity Hardware Transactional Memory
    Zhang, Tong
    Lee, Dongyoon
    Jung, Changhee
    ACM SIGPLAN NOTICES, 2016, 51 (04) : 159 - 173
  • [35] Embedded-TM: Energy and complexity-effective hardware transactional memory for embedded multicore systems
    Ferri, Cesare
    Wood, Samantha
    Moreshet, Tali
    Bahar, R. Iris
    Herlihy, Maurice
    JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 2010, 70 (10) : 1042 - 1052
  • [36] Aware conflict detection of non-uniform memory access system and prevention for transactional memory
    Wang Rui-bo
    Lu Kai
    Lu Xi-cheng
    JOURNAL OF CENTRAL SOUTH UNIVERSITY, 2012, 19 (08) : 2266 - 2271
  • [37] Aware conflict detection of non-uniform memory access system and prevention for transactional memory
    Rui-bo Wang
    Kai Lu
    Xi-cheng Lu
    Journal of Central South University, 2012, 19 : 2266 - 2271
  • [38] Aware conflict detection of non-uniform memory access system and prevention for transactional memory
    王睿伯
    卢凯
    卢锡城
    Journal of Central South University, 2012, 19 (08) : 2266 - 2271
  • [39] LosaTM: A Hardware Transactional Memory Integrated With a Low-Overhead Scenario-Awareness Conflict Manager
    Fu, Chao
    Wan, Li
    Han, Jun
    IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, 2022, 33 (12) : 4849 - 4862
  • [40] Extending hardware transactional memory capacity via rollback-only transactions and suspend/resumePOWER8 TM
    Shady Issa
    Pascal Felber
    Alexander Matveev
    Paolo Romano
    Distributed Computing, 2020, 33 : 327 - 348