Proceedings of the 46th International Symposium on Computer Architecture 2019
DOI: 10.1145/3307650.3322223
|View full text |Cite
|
Sign up to set email alerts
|

Translation ranger

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

2
24
0

Year Published

2020
2020
2023
2023

Publication Types

Select...
4
2
1

Relationship

0
7

Authors

Journals

citations
Cited by 46 publications
(26 citation statements)
references
References 29 publications
2
24
0
Order By: Relevance
“…We quantitatively compare our proposed TLB compression mechanism with a recently published range TLB optimization [69] to demonstrate that TLB compression is more effective and suitable in GPU executions. Our scheme can capture non-continuous page accesses and does not rely on the OS to generate continuous physical pages as in range TLB.…”
Section: Comparison To An Alternative Tlb Optimizationmentioning
confidence: 99%
See 3 more Smart Citations
“…We quantitatively compare our proposed TLB compression mechanism with a recently published range TLB optimization [69] to demonstrate that TLB compression is more effective and suitable in GPU executions. Our scheme can capture non-continuous page accesses and does not rely on the OS to generate continuous physical pages as in range TLB.…”
Section: Comparison To An Alternative Tlb Optimizationmentioning
confidence: 99%
“…Bharadwaj et al [10] studied distributed TLB slicing and corresponding network topologies, to accelerate address translation. Yan et al [69] proposed translation ranger, an OS support, to enable continuous page accesses such that the TLB can store fewer translations. Pham et al [43] proposed a Bloom filter-based hardware mechanism that can be used to reduce the overheads imposed by cache flushes due to virtual page remappings.…”
Section: Related Workmentioning
confidence: 99%
See 2 more Smart Citations
“…Address translation overheads due to data accesses is a major performance bottleneck in workloads featuring large data working sets [55,58,66,71,82,84,99,117,175,202,237,298]. These workloads exacerbate TLB pressure, causing frequent data TLB misses that incur high performance and energy costs due to the page walks required for fetching the corresponding address translation entries.…”
Section: Agile Prefetching For the Data Tlb Miss Streammentioning
confidence: 99%