Filters








5,678 Hits in 3.5 sec

Implementation of hybrid version management in hardware transactional memory

Lihang Zhao, Jeff Draper
2013 2013 IEEE International Symposium on Circuits and Systems (ISCAS2013)  
In this paper, we present a light-weight and generic implementation of the key structures to support hybrid version management in Hardware Transactional Memory (HTM).  ...  Full-system simulation demonstrates the performance advantage of our design (up to 40% improvement).  ...  In the lazy mechanism, the speculative version is not stored to memory until the transaction commits. During the execution of the transaction, pre-transaction version remains in memory.  ... 
doi:10.1109/iscas.2013.6571962 dblp:conf/iscas/ZhaoD13 fatcat:oohh4vs6sbbpdnn77jmunlv7ma

A Dynamically Adaptable Hardware Transactional Memory

Marc Lupon, Grigorios Magklis, Antonio Gonzalez
2010 2010 43rd Annual IEEE/ACM International Symposium on Microarchitecture  
While eager HTM systems store transactional state in-place in memory and resolve conflicts when they are produced, lazy HTM systems buffer the transactional state in specialized hardware and defer the  ...  Both the eager and the lazy execution modes of DynTM exhibit very high performance compared to modern HTM systems.  ...  and Intel Corporation.  ... 
doi:10.1109/micro.2010.23 dblp:conf/micro/LuponMG10 fatcat:7i232v5kwfhs7cknn5wfgwylg4

Lazy Symbolic Execution for Enhanced Learning [chapter]

Duc-Hiep Chu, Joxan Jaffar, Vijayaraghavan Murali
2014 Lecture Notes in Computer Science  
We argue that the trade-off between this 'enhanced learning' and incurring additional cost (which in principle may not be productive) is in fact in favor of speculation.  ...  This 'learning' phase consists of computing interpolants in the same spirit of no-good learning in sat solvers.  ...  Otherwise, we use the interpolant computed by speculation (line 11).  ... 
doi:10.1007/978-3-319-11164-3_27 fatcat:yopj6u4rbnhmnmywuo7fuvwpwe

Testing Speculative Work in a Lazy/Eager Parallel Functional Language [chapter]

Alberto de la Encina, Ismael Rodríguez, Fernando Rubio
2006 Lecture Notes in Computer Science  
However, in order to enable the parallel execution of processes in Eden, this feature must be disabled when new processes are instantiated.  ...  In this paper we present a framework to compare the speculated computations of an Eden program with the computations it actually requires.  ...  As a Haskell extension, Eden applies the laziness for deciding the computations to be executed in each moment.  ... 
doi:10.1007/978-3-540-69330-7_19 fatcat:sktnd7wyjrajdpsjwlktbe2i6q

LazyFP: Leaking FPU Register State using Microarchitectural Side-Channels [article]

Julian Stecklina, Thomas Prescher
2018 arXiv   pre-print
In this paper, we present an attack that exploits lazy FPU context switching and allows an adversary to recover the FPU and SIMD register set of arbitrary processes or VMs.  ...  As an optimization, the operating system can defer the context switch of the FPU and SIMD register set until the first instruction is executed that needs access to these registers.  ...  ACKNOWLEDGEMENTS The authors would like to thank Anthony Liguori and Tor Lund-Larsen for their support in the publication of this paper, and Chad Skinner for managing our relations with Intel.  ... 
arXiv:1806.07480v1 fatcat:24xg3zempzeargkvxcfo2i52ga

Optimistic evaluation

Robert Ennals, Simon Peyton Jones
2003 SIGPLAN notices  
A run-time adaption mechanism records expressions found to be unsuitable for speculative evaluation, and arranges for them to be evaluated more lazily in the future.  ...  Lazy programs are beautiful, but they are slow because they build many thunks.  ...  We would also like to thank Manuel Chakravarty, Fergus Henderson, Jan-Willem Maessen, Simon Marlow, Greg Morrisett, Alan Mycroft, Nick Nethercote, Andy Pitts, Norman Ramsey, John Reppy, and Richard Sharp  ... 
doi:10.1145/944746.944731 fatcat:vs4vh64egfcvpi72xfvilkbceq

Optimistic evaluation

Robert Ennals, Simon Peyton Jones
2003 Proceedings of the eighth ACM SIGPLAN international conference on Functional programming - ICFP '03  
A run-time adaption mechanism records expressions found to be unsuitable for speculative evaluation, and arranges for them to be evaluated more lazily in the future.  ...  Lazy programs are beautiful, but they are slow because they build many thunks.  ...  We would also like to thank Manuel Chakravarty, Fergus Henderson, Jan-Willem Maessen, Simon Marlow, Greg Morrisett, Alan Mycroft, Nick Nethercote, Andy Pitts, Norman Ramsey, John Reppy, and Richard Sharp  ... 
doi:10.1145/944705.944731 dblp:conf/icfp/EnnalsJ03 fatcat:o7ixfn3hkrddliwst36kdhxvoa

RETCON

Colin Blundell, Arun Raghavan, Milo M.K. Martin
2010 Proceedings of the 37th annual international symposium on Computer architecture - ISCA '10  
Over the past decade there has been a surge of academic and industrial interest in optimistic concurrency, i.e. the speculative parallel execution of code regions that have the semantics of isolation.  ...  We find that one common bottleneck is updates to auxiliary program data in otherwise non-conflicting operations, e.g. reference count updates and hashtable occupancy field increments.  ...  ., intruder_opt-sz) and/or reducing contention in the memory system (the source of the reduction in busy time in vacation_opt-sz and python_opt).  ... 
doi:10.1145/1815961.1815995 dblp:conf/isca/BlundellRM10 fatcat:iro2zdedafbwjj4w7anxw2tztq

RETCON

Colin Blundell, Arun Raghavan, Milo M.K. Martin
2010 SIGARCH Computer Architecture News  
Over the past decade there has been a surge of academic and industrial interest in optimistic concurrency, i.e. the speculative parallel execution of code regions that have the semantics of isolation.  ...  We find that one common bottleneck is updates to auxiliary program data in otherwise non-conflicting operations, e.g. reference count updates and hashtable occupancy field increments.  ...  ., intruder_opt-sz) and/or reducing contention in the memory system (the source of the reduction in busy time in vacation_opt-sz and python_opt).  ... 
doi:10.1145/1816038.1815995 fatcat:zj2f7cqgkfgcreh6jarsxvgizq

GPU-TLS: An Efficient Runtime for Speculative Loop Parallelization on GPUs

Chenggang Zhang, Guodong Han, Cho-Li Wang
2013 2013 13th IEEE/ACM International Symposium on Cluster, Cloud, and Grid Computing  
GPU Architecture and Execution Model The modern Graphical Processing Unit (GPU) architecture consists of two major parts: the computation sub-system and the memory sub-system.  ...  The speculative execution of a G-PU kernel has four phases: speculative execution, dependency checking, commit and mis-speculation recovery.  ... 
doi:10.1109/ccgrid.2013.34 dblp:conf/ccgrid/ZhangHW13 fatcat:gqqdzklyk5aedmt2zpzd3emlmi

Adaptive Versioning in Transactional Memory Systems

Pavan Poudel, Gokarna Sharma
2021 Algorithms  
Transactions are executed speculatively and the speculative execution is supported through data versioning mechanism.  ...  improvements as much as 6.3× for execution time and as much as 170× for number of aborts in non-persistent transactional memory systems.  ...  The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.  ... 
doi:10.3390/a14060171 fatcat:faiwls4l4zgundbbmptlxbqr5u

A Cost-based Optimizer for Gradient Descent Optimization

Zoi Kaoudi, Jorge-Arnulfo Quiane-Ruiz, Saravanan Thirumuruganathan, Sanjay Chawla, Divy Agrawal
2017 Proceedings of the 2017 ACM International Conference on Management of Data - SIGMOD '17  
Ideally, a user will specify an ML task in a high-level and easy-to-use language and the framework will invoke the appropriate algorithms and system configurations to execute it.  ...  Thus, to decouple a user specification of an ML task from its execution, a key component is a GD optimizer. We propose a cost-based GD optimizer that selects the best GD plan for a given ML task.  ...  Summary The high efficiency of our system comes from its (i) lazy transformation technique, (ii) novel sampling mechanisms, and (iii) efficient execution operators.  ... 
doi:10.1145/3035918.3064042 dblp:conf/sigmod/KaoudiQTCA17 fatcat:peqngom2z5gibd7wfbiskbccxy

Tradeoffs in buffering speculative memory state for thread-level speculation in multiprocessors

María Jesús Garzarán, Milos Prvulovic, José María Llabería, Víctor Viñals, Lawrence Rauchwerger, Josep Torrellas
2005 ACM Transactions on Architecture and Code Optimization (TACO)  
As speculative tasks run concurrently, they generate unsafe or speculative memory state that needs to be separately buffered and managed in the presence of distributed caches and buffers.  ...  In this paper, we introduce a novel taxonomy of approaches to buffer and manage multiversion speculative memory state in multiprocessors.  ...  Overall, laziness reduces the average execution time by 9% in the simpler schemes (SingleT and MultiT&SV) and by only 3% in MultiT&MV.  ... 
doi:10.1145/1089008.1089010 fatcat:ekiblury4nbrximsgoxonogxhq

Runtime Parallelization of Static and Dynamic Irregular Array of Array References

Parwat Singh Anjanaa, N. Naga Maruthia, Sagar Gujjunooria, Madhu Orugantib
2018 International Journal of Engineering & Technology  
The advancement of computer systems such as multi-core and multiprocessor systems resulted in much faster computing than earlier.  ...  In this article, we propose three different solutions to parallelize a loop having an irregularity in the array of array references, with and without dependencies.  ...  Dependence Detection and Commit In BitTLS, a lazy conflict detection mechanism is implemented, in which a thread have to wait until the commit of a less speculative thread.  ... 
doi:10.14419/ijet.v7i4.6.20452 fatcat:jprxbbpnhjfntmxn7kz4ymf3gq

EcoTM: Conflict-aware Economical Unbounded Hardware Transactional Memory

Saša Tomić, Ege Akpinar, Adrian Cristál, Osman Unsal, Mateo Valero
2013 Procedia Computer Science  
speculatively modified.  ...  We evaluate EcoTM performance by comparing it with ideal-lazy HTM, unbounded eager HTM with perfect signatures, and LogTM-SE.  ...  On such system, a naive implementation of unbounded lazy HTM can be straightforward and easy to understand.  ... 
doi:10.1016/j.procs.2013.05.190 fatcat:jkevqwwj75cpbozk7w7hfk2eby
« Previous Showing results 1 — 15 out of 5,678 results