Dead-block prediction & dead-block correlating prefetchers
By An-Chow Lai, Cem Fide, Babak Falsafi. Proceedings 28th Annual International Symposium on Computer Architecture, 2001
357 citations recorded by the source corpus.
Subfields: Computer Networks and Communications, Hardware and Architecture
View Paper: https://www.semanticscholar.org/paper/10c1bfa7fe190b3fc6d64c4909cb7ef0911a7b90
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Networks and Communications
- Onur Mutlu
- Gabriel H. Loh
- Yuejian Xie
- Moinuddin K. Qureshi
- Yale N. Patt
- Joel S. Emer
- Aamer Jaleel
- Simon C. Steely Jr.
In Hardware and Architecture
- Onur Mutlu
- Gabriel H. Loh
- Yuejian Xie
- Carole-Jean Wu
- Margaret Martonosi
- Moinuddin K. Qureshi
- Yale N. Patt
- Joel S. Emer
Citations in the corpus
- Spatial Memory Streaming (2006, background): “, [18,20,24,25,29]), few studies have demonstrated success at improving memory system performance for commercial applications.”
- Adaptive insertion policies for high performance caching (2007, background): “After the predicted last touch, the line can either be turned off [7] or be used to store prefetched data [8].”
- PIPP: promotion/insertion pseudo-partitioning of multi-core shared caches (2009, background): “Dead-line prediction (also called dead-block prediction) attempts to anticipate the last touch to a given line [1, 16, 22, 26, 29].”
- SHiP: Signature-based Hit Predictor for high performance caching (2011, methodology): “[18] proposed dead block correlating prefetchers (DBCP) that prefetch data into dead cache blocks in the L1 cache.”
- Data Cache Prefetching Using a Global History Buffer (2004, background): “Although they are simple, conventional table-based methods [4,6,7,9,10,11,12,14] are relatively inefficient, in that they reserve a fixed amount of history space per prefetch key.”