8,206 Hits in 3.5 sec

Cache replacement based on reuse-distance prediction

Georgios Keramidas, Pavlos Petoumenos, Stefanos Kaxiras
2007 2007 25th International Conference on Computer Design  
Several cache management techniques have been proposed that indirectly try to base their decisions on cacheline reuse-distance, like Cache Decay which is a postdiction of reuse-distances: if a cacheline  ...  In this work, we propose to directly predict reuse-distances via instruction-based (PC) prediction and use this information for cache level optimizations.  ...  This guarantees that if we have reliable information about large reuse-distances, the replacement decision will be based on those; otherwise we revert to LRU replacement.  ... 
doi:10.1109/iccd.2007.4601909 dblp:conf/iccd/KeramidasPK07 fatcat:3tqipqswarfrhcavcylri3hayu

Instruction-based reuse-distance prediction for effective cache management

Pavlos Petoumenos, Georgios Keramidas, Stefanos Kaxiras
2009 2009 International Symposium on Systems, Architectures, Modeling, and Simulation  
In this work, we quantify the temporal characteristics of the cache block at run time by predicting the cache block reuse distances (measured in intervening cache accesses), based on the access patterns  ...  We show that an instruction-based reused distance predictor is very accurate and allows approximation of optimal replacement decisions, since we can "see" the future.  ...  PRACTICAL REUSE-DISTANCE COLLECTION AT RUN-TIME The predictor structure discussed in the previous section can deliver predictions based on its stored history information about reuse distances.  ... 
doi:10.1109/icsamos.2009.5289241 dblp:conf/samos/PetoumenosKK09 fatcat:mdatwpznjnahfpoodhkgjnxhzu

Throttling Capacity Sharing Using Life Time and Reuse Time Prediction in Private L2 Caches of Chip Multiprocessors

Young-Sik EOM, Jong Wook KWAK, Seong Tae JHANG, Chu Shik JHON
2012 IEICE transactions on information and systems  
TCS determines whether to spill a replaced block by predicting reuse possibility, based on life time and reuse time.  ...  To deal with this problem, private caches can share capacity through spilling replaced blocks to other private caches.  ...  reuse time as follows, since threshold value is based on them.relativeError = |Time real − Time prediction | MAX(Time real , Time prediction ) × 100 (6)  ... 
doi:10.1587/transinf.e95.d.1676 fatcat:wj5ssyvqurh6ffiosa6hql5epu


We show that the ACR policy performs better than recently proposed thread-aware dynamic re-reference interval prediction (TA-DRRIP) and protecting distance based (PDP) techniques for various 2-core, 4-  ...  Such multicores handle the large bandwidth requirement from the processing cores by employing multiple levels of caches with one or two levels of private caches along with a shared last-level cache (LLC  ...  Reuse distance predictor based on PC is used to drive the replacement policy in [11, 12] .  ... 
doi:10.34218/ijcet.10.2.2019.023 fatcat:p2ca7ukxszgjficjlltibqywsy

A modeling framework for reuse distance-based estimation of cache performance

Xiaoyue Pan, Bengt Jonsson
2015 2015 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)  
We develop an analytical modeling framework for efficient prediction of cache miss ratios based on reuse distance distributions.  ...  The predictions are based on a uniform modeling framework which can be specialized for a variety of cache replacement policies, including Random, LRU, PLRU, and MRU (aka. bit-PLRU), and for arbitrary values  ...  An efficient alternative is to base cache performance predictions on reuse distance distributions.  ... 
doi:10.1109/ispass.2015.7095785 dblp:conf/ispass/PanJ15 fatcat:i4xksqrn4ve7devkkeuwj5r52y

Learning Forward Reuse Distance [article]

Pengcheng Li, Yongbin Gu
2020 arXiv   pre-print
We find that a powerful LSTM-based recurrent neural network model can provide high prediction accuracy based on only a cache trace as input.  ...  The recent advancement of deep learning techniques enables the design of novel intelligent cache replacement policies.  ...  This paper strives to predict forward reuse distance based on only the past cache trace by applying a novel learningaided approach.  ... 
arXiv:2007.15859v1 fatcat:qn5topsvfrhvzc4dqavkm656oy

An Imitation Learning Approach for Cache Replacement [article]

Evan Zheran Liu, Milad Hashemi, Kevin Swersky, Parthasarathy Ranganathan, Junwhan Ahn
2020 arXiv   pre-print
To increase cache hits, we focus on the problem of cache replacement: choosing which cache line to evict upon inserting a new line.  ...  Program execution speed critically depends on increasing cache hits, as cache hits are orders of magnitude faster than misses.  ...  We also thank Chelsea Finn, Lisa Lee, and Amir Yazdanbakhsh for their comments on a draft of this paper.  ... 
arXiv:2006.16239v2 fatcat:v2gm2hjanrdhnafysbswk7dsku

Tighter WCET analysis of input dependent programs with classified-cache memory architecture

Yanhui Li, Shakith Devinda Fernando, Heng Yu, Xiaolei Chen, Yajun Ha, Teng Tiow Tay
2008 2008 15th IEEE International Conference on Electronics, Circuits and Systems  
Our work classifies predictable and unpredictable accesses and allocates them into predictable caches and unpredictable caches respectively, using the CME (Cache Miss Equations) and reuse-distance based  ...  Several works have studied the data cache impacts on the WCET of programs, but they can only handle programs with no input dependent data accesses.  ...  Further research explored the method of reducing the reuse distance and cache misses based on reuse distance visualization.  ... 
doi:10.1109/icecs.2008.4674877 dblp:conf/icecsys/LiFYCHT08 fatcat:b4ou52jjlzhrlngkqucczgtsxy

Improving Cache Management Policies Using Dynamic Reuse Distances

Nam Duong, Dali Zhao, Taesu Kim, Rosario Cammarota, Mateo Valero, Alexander V. Veidenbaum
2012 2012 45th Annual IEEE/ACM International Symposium on Microarchitecture  
Cache management policies such as replacement, bypass, or shared cache partitioning have been relying on data reuse behavior to predict the future.  ...  A new replacement policy is proposed which prevents replacing a cache line until a certain number of accesses to its cache set, called a Protecting Distance (PD).  ...  In summary, replacement based on protecting distance reduces cache pollution. A combination of replacement and bypass further improves reuse and reduces pollution.  ... 
doi:10.1109/micro.2012.43 dblp:conf/micro/DuongZKCVV12 fatcat:mzjuzn4mm5fu5jsnrvynavxvgm

ReMAP: Reuse and memory access cost aware eviction policy for last level cache management

Akhil Arunkumar, Carole-Jean Wu
2014 2014 IEEE 32nd International Conference on Computer Design (ICCD)  
In this paper, we show that in addition to the recency information provided by the cache replacement policy, post eviction reuse distance (PERD) and main memory access latency cost are useful to make betterinformed  ...  Such dependence on the last-level cache (LLC) has motivated numerous innovations in cache management schemes.  ...  Though RRIP and EAF-cache predict the reuse behavior of cache lines, their predictions are limited to insertion time. On the contrary, ReMAP uses post eviction reuse distance prediction.  ... 
doi:10.1109/iccd.2014.6974670 dblp:conf/iccd/ArunkumarW14 fatcat:ips52qvzr5fstooqkbmtnbazqu

Optimal bypass monitor for high performance last-level caches

Lingda Li, Dong Tong, Zichao Xie, Junlin Lu, Xu Cheng
2012 Proceedings of the 21st international conference on Parallel architectures and compilation techniques - PACT '12  
In the last-level cache, large amounts of blocks have reuse distances greater than the available cache capacity.  ...  Cache performance and efficiency can be improved if some subset of these distant reuse blocks can reside in the cache longer.  ...  A recent proposal guides replacement by explicitly predicting the reuse distance with a PC based predictor [17] .  ... 
doi:10.1145/2370816.2370862 dblp:conf/IEEEpact/LiTXLC12 fatcat:mio4na2nbbdxpfve56tquje54m

Characterizing the impact of last-level cache replacement policies on big-data workloads

Alexandre Valentin Jamet, Lluc Alvarez, Daniel A. Jimenez, Marc Casas
2020 2020 IEEE International Symposium on Workload Characterization (IISWC)  
The first proposed replacement policy, Multi-Sampler Multiperspective (MS-MPPPB), uses multiple samplers instead of a single one and dynamically selects the best-behaving sampler to drive reuse distance  ...  state-of-the-art LLC replacement policies such as Multiperspective Reuse Prediction (MPPPB), Glider, Hawkeye, SHiP, DRRIP and SRRIP.  ...  Reuse Distance Prediction As reuse distance is a crucial concept when it comes to cache replacement, recent works focus on proposing new techniques to build run-time approximations of the distance to the  ... 
doi:10.1109/iiswc50251.2020.00022 dblp:conf/iiswc/JametAJC20 fatcat:klljuqur3fd25ezm5okkg52sfi

Relative Performance of a Multi-level Cache with Last-Level Cache Replacement: An Analytic Review [article]

Bijay Paikaray
2013 arXiv   pre-print
Current day processors employ multi-level cache hierarchy with one or two levels of private caches and a shared last-level cache (LLC).  ...  This paper analyzes some of the existing replacement techniques on the LLC with their performance assessment.  ...  On a cache miss, if the predicted reuse distance of the memory reference is higher than the reuse distance seen by all cache lines in the set, the requested data is directly sent to the processor without  ... 
arXiv:1307.6406v1 fatcat:xswji4f5pzbnfdfg24gygqzoqu

Exploiting compressed block size as an indicator of future reuse

Gennady Pekhimenko, Tyler Huberty, Rui Cai, Onur Mutlu, Phillip B. Gibbons, Michael A. Kozuch, Todd C. Mowry
2015 2015 IEEE 21st International Symposium on High Performance Computer Architecture (HPCA)  
• Observations: -Importance of the cache block depends on its compressed size -Block size is indicative of reuse behavior in some applications • Key Idea: Use compressed block size in making cache replacement  ...  In a compressed cache, compressed block size is an additional dimension • Problem: How can we maximize cache performance utilizing both block reuse and compressed size?  ...  • Observations: -Importance of the cache block depends on its compressed size -Block size is indicative of reuse behavior in some applications • Key Idea: Use compressed block size in making cache replacement  ... 
doi:10.1109/hpca.2015.7056021 dblp:conf/hpca/PekhimenkoHCMGK15 fatcat:rt4v5lglajb6dp3itxkhzg6wwi

Modeling cache performance beyond LRU

Nathan Beckmann, Daniel Sanchez
2016 2016 IEEE International Symposium on High Performance Computer Architecture (HPCA)  
It uses absolute reuse distances instead of stack distances, and models replacement policies as abstract ranking functions. These innovations let us model arbitrary age-based replacement policies.  ...  Accurate predictions of cache performance enable many optimizations in multicore systems.  ...  Instruction-based reuse distance prediction [27] uses PC-based classification, but does not do protection.  ... 
doi:10.1109/hpca.2016.7446067 dblp:conf/hpca/BeckmannS16 fatcat:pcbfqiafrba5fhrjqup7pituga
« Previous Showing results 1 — 15 out of 8,206 results