Abstract | ||
---|---|---|
Current processors carry out a Page Table Walk to translate a virtual address into a physical address. This process requires several memory accesses to retrieve the entry from each of the levels of the Page Table Tree. As each main memory access has a large latency, the address translation process may penalize significantly the system performance. Translation Lookaside Buffers (TLBs) avoid completely these accesses by storing the most recent translations. But on a TLB miss a Page Table Walk is required. An effective way to accelerate it is introducing a cache inside/near the MMU to store partial translations. Recent works have shown the advantages of using a Page Walk Cache on the service of a single core. In this paper we propose a Page Walk Cache shared among the cores of a CMP and analyze its behavior and how it improves performance for parallel applications, comparing it to a private Page Walk Cache scheme. Evaluation results show that shared PTWC increases up to a 19.8% the hit-rate, leading to a TLB miss latency reduction of up to 12.2%, when compared to a private PTWC. |
Year | DOI | Venue |
---|---|---|
2013 | 10.1007/978-3-642-54420-0_43 | Lecture Notes in Computer Science |
Keywords | Field | DocType |
virtual memory,page walk,cache organization | Physical address,Computer science,Virtual memory,Latency (engineering),Cache,Virtual address space,Page table,Parallel computing,Cache algorithms,Translation lookaside buffer,Distributed computing | Conference |
Volume | ISSN | Citations |
8374 | 0302-9743 | 1 |
PageRank | References | Authors |
0.34 | 5 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Albert Esteve | 1 | 12 | 2.17 |
María Engracia Gómez | 2 | 149 | 17.48 |
Antonio Robles | 3 | 481 | 30.40 |