Matches in SemOpenAlex for { <https://semopenalex.org/work/W2409751210> ?p ?o ?g. }
- W2409751210 abstract "The semiconductor industry is experiencing a shift from computation-bound design to communication-bound design. Many future systems will use one or many chip multiprocessors (CMPs) and support shared memory, as CMP-based systems can provide high-performance, cost-effective computing for workloads with abundant thread-level parallelism. One of the biggest challenges in CMP designs is to employ efficient cache coherence protocols to maintain coherence. Most commercial products implement directory-based protocols to maintain coherence. Directory-based protocols avoid the bandwidth and electrical limits of a centralized interconnect by tracking the global coherence state of cache lines via a structure. Directory-based protocols are preferred in future systems due to their ability to exploit arbitrary point-to-point interconnects. However, existing directory-based protocols are not optimum in specific contexts. For example, existing protocols cannot optimize the coherence traffic according to the sharing patterns shown in the applications or the varying latency, bandwidth needs of different coherence messages, which leads to suboptimal performance and extra power consumption. To overcome these limitations of conventional directory-based protocols, we propose two cache coherence protocols that optimize the coherence traffic based on context knowledge. After automatically detecting a stable producer-consumer pattern in an application, our sharing pattern-aware coherence protocol uses delegation to delegate the home directory of a cache line to the producer node, thereby converting 3-hop coherence operations into 2-hop operations. Then, the producer employs speculative updates to push the data to where it might soon be consumed. When the producer correctly predicts when and where to send updates, these 3-hop misses become local misses, effectively eliminating the impact of remote memory latency. Our interconnect-aware protocol can exploit a heterogenous interconnect comprised of wires with varying latency, bandwidth, and energy characteristics. By intelligently mapping critical messages to wires optimized for delay and noncritical messages to wires optimized for low power, our interconnect-aware protocol can achieve performance improvement and power reduction at the same time. We demonstrate the performance advantage of the proposed mechanisms through architecture-level simulation. The producer-consumer sharing-aware protocol reduces the average remote miss rate by 40%, reduces network traffic by 15%, and improves performance by 21% on seven benchmark programs that exhibit producerconsumer sharing using a cycle-accurate simulator of a future 16-node SGI multiprocessor. The interconnect-aware protocol yields a performance improvement of 11% and energy reduction of 22% on a set of scientific and commercial benchmarks for a future 16-core CMP system." @default.
- W2409751210 created "2016-06-24" @default.
- W2409751210 creator A5014611111 @default.
- W2409751210 creator A5084000289 @default.
- W2409751210 date "2007-01-01" @default.
- W2409751210 modified "2023-09-27" @default.
- W2409751210 title "Context-aware coherence protocols for future processors" @default.
- W2409751210 cites W1496710542 @default.
- W2409751210 cites W1548312885 @default.
- W2409751210 cites W1555915743 @default.
- W2409751210 cites W1556480701 @default.
- W2409751210 cites W1558805549 @default.
- W2409751210 cites W1589943916 @default.
- W2409751210 cites W1850405760 @default.
- W2409751210 cites W1879996931 @default.
- W2409751210 cites W1882088395 @default.
- W2409751210 cites W1966141505 @default.
- W2409751210 cites W1966887456 @default.
- W2409751210 cites W1969243435 @default.
- W2409751210 cites W1973468635 @default.
- W2409751210 cites W1976186908 @default.
- W2409751210 cites W1978980371 @default.
- W2409751210 cites W1986185670 @default.
- W2409751210 cites W1995546379 @default.
- W2409751210 cites W2001490269 @default.
- W2409751210 cites W2019198721 @default.
- W2409751210 cites W2022740893 @default.
- W2409751210 cites W2028292906 @default.
- W2409751210 cites W2034919323 @default.
- W2409751210 cites W2037531986 @default.
- W2409751210 cites W2039358455 @default.
- W2409751210 cites W2044643398 @default.
- W2409751210 cites W2051265155 @default.
- W2409751210 cites W2054739713 @default.
- W2409751210 cites W2059736952 @default.
- W2409751210 cites W2060616331 @default.
- W2409751210 cites W2066691060 @default.
- W2409751210 cites W2068713803 @default.
- W2409751210 cites W2073776178 @default.
- W2409751210 cites W2082881490 @default.
- W2409751210 cites W2088310619 @default.
- W2409751210 cites W2095644119 @default.
- W2409751210 cites W2098537124 @default.
- W2409751210 cites W2098737795 @default.
- W2409751210 cites W2099706037 @default.
- W2409751210 cites W2100720297 @default.
- W2409751210 cites W2100893446 @default.
- W2409751210 cites W2102081731 @default.
- W2409751210 cites W2103695893 @default.
- W2409751210 cites W2104606146 @default.
- W2409751210 cites W2106500842 @default.
- W2409751210 cites W2108583124 @default.
- W2409751210 cites W2112121929 @default.
- W2409751210 cites W2113338652 @default.
- W2409751210 cites W2113676942 @default.
- W2409751210 cites W2114827297 @default.
- W2409751210 cites W2114995630 @default.
- W2409751210 cites W2117324528 @default.
- W2409751210 cites W2118033476 @default.
- W2409751210 cites W2118383989 @default.
- W2409751210 cites W2118543332 @default.
- W2409751210 cites W2120369192 @default.
- W2409751210 cites W2120635877 @default.
- W2409751210 cites W2122096916 @default.
- W2409751210 cites W2123199776 @default.
- W2409751210 cites W2130562747 @default.
- W2409751210 cites W2130611655 @default.
- W2409751210 cites W2133985894 @default.
- W2409751210 cites W2134182085 @default.
- W2409751210 cites W2135673469 @default.
- W2409751210 cites W2139239342 @default.
- W2409751210 cites W2139397536 @default.
- W2409751210 cites W2140370341 @default.
- W2409751210 cites W2141386638 @default.
- W2409751210 cites W2142002613 @default.
- W2409751210 cites W2143246663 @default.
- W2409751210 cites W2145021036 @default.
- W2409751210 cites W2145252892 @default.
- W2409751210 cites W2148740960 @default.
- W2409751210 cites W2152390090 @default.
- W2409751210 cites W2152450729 @default.
- W2409751210 cites W2152654963 @default.
- W2409751210 cites W2153075575 @default.
- W2409751210 cites W2153904572 @default.
- W2409751210 cites W2154904877 @default.
- W2409751210 cites W2155350341 @default.
- W2409751210 cites W2156476900 @default.
- W2409751210 cites W2158630919 @default.
- W2409751210 cites W2159535674 @default.
- W2409751210 cites W2160010046 @default.
- W2409751210 cites W2162109942 @default.
- W2409751210 cites W2162192855 @default.
- W2409751210 cites W2162826731 @default.
- W2409751210 cites W2163581305 @default.
- W2409751210 cites W2164264749 @default.
- W2409751210 cites W2166213383 @default.
- W2409751210 cites W2167722125 @default.
- W2409751210 cites W2167848093 @default.
- W2409751210 cites W2168682824 @default.
- W2409751210 cites W2169387321 @default.