Pointer cache assisted prefetching

Collins, Jamison D.; Sair, Suleyman; Calder, Brad; Tullsen, Dean M.

doi:10.1109/micro.2002.1176239

Cited by 60 publications

(52 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Focusing on misses, these schemes target the subset of the future working set which is not currently cached. Dependence-following schemes [2,10] prefetch by following dependence chains through memory. While valuable, these techniques alone are of limited utility immediately after a migration, due to their serial progression.…”

Section: Background and Related Workmentioning

confidence: 99%

“…Pointer, Pointer-chase: These capture active pointers and pointer traversals, respectively, by detecting when the data output of a load matches the address of a later memory access, similar to pointer-cache [10] and dependence- based [28] prefetching. Pointer tracks loaded values used as addresses without following them, while Pointer-chase replaces each entry with the target value when a load match is observed.…”

Section: Memory Loggermentioning

confidence: 99%

See 1 more Smart Citation

Fast thread migration via cache working set prediction

Brown

Porter

Tullsen

2011

2011 IEEE 17th International Symposium on High Performance Computer Architecture

View full text Add to dashboard Cite

show abstract

Section: Background and Related Workmentioning

confidence: 99%

Section: Memory Loggermentioning

confidence: 99%

Fast thread migration via cache working set prediction

Brown

Porter

Tullsen

2011

2011 IEEE 17th International Symposium on High Performance Computer Architecture

View full text Add to dashboard Cite

show abstract

“…However, they are not available in commercial hardware today. Proposals include to either use hardware only pointer prefetchers [18,28], or to use hardware pointer prefetchers assisted by the software [16]. Examining some of these techniques for R can potentially minimize the number of cache misses due to pointer-chasing.…”

Section: Hardware Enhancements: Research Opportunitiesmentioning

confidence: 99%

Profiling R on a contemporary processor

Sridharan

Patel

2014

Proc. VLDB Endow.

View full text Add to dashboard Cite

R is a popular data analysis language, but there is scant experimental data characterizing the run-time profile of R programs. This paper addresses this limitation by systematically cataloging where time is spent when running R programs. Our evaluation using four different workloads shows that when analyzing large datasets, R programs a) spend more than 85% of their time in processor stalls, which leads to slower execution times, b) trigger the garbage collector frequently, which leads to higher memory stalls, and c) create a large number of unnecessary temporary objects that causes R to swap to disk quickly even for datasets that are far smaller than the available main memory. Addressing these issues should allow R programs to run faster than they do today, and allow R to be used for analyzing even larger datasets. As outlined in this paper, the results presented in this paper motivate a number of future research investigations in the database, architecture, and programming language communities. All data and code that is used in this paper (which includes the R programs, and changes to the R source code for instrumentation) can be found at: http://quickstep.cs.wisc.edu/dissecting-R/.

show abstract

“…Sair, et,al [1] survey several prefetchers, and introduce a method for classifying memory access behaviours in hardware: the branch access stream is matched against the behaviour specific table in parallel. We utilize such a similar technique in the listing the branches in the programming environment.…”

Section: Back Ground and Related Workmentioning

confidence: 99%