Capacity-Latency Tradeoffs in CXL Memory Expander at Hyperscale
IEEE Computer Architecture Letters · 2026-01 · 25 (1) · 122-125
DOI: 10.1109/lca.2026.3676988 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Computer Architecture Letters, 2026-01. Verfasst von Haneul Park und 4 weiteren Autor:innen. Die Arbeit nennt 19 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: PRISM and CogniSense, Semiconductor Research Corporation, DARPA.
Abstract (Originalsprache)
DRAM dominates the cost of hyperscale datacenter servers today and AI-driven memory supply shortages are pushing costs even higher. CXL memory expanders are a new technology to reduce server costs by reusing legacy DDR DRAM modules from older servers. Memory expanders further save cost through inline compression that increases the effective capacity. This paper investigates two critical design points of a memory expander: the compression algorithm and the compression block size. Both design points introduce tradeoffs between latency and capacity that involve complex interactions with the memory expander architecture, access patterns to the device, and the entropy of stored data. Current commercial devices use simple, fast compression and a 4 kB block as the default. Using a compressibility benchmark and the latest traces of cold memory accesses from a hyperscaler, we find that using a different compression algorithm and block size can simultaneously increase expander capacity and reduce access latency. The key insight is counter-intuitive: using a slower, more complex compression algorithm improves latency because the reduction in DRAM traffic more than offsets the additional decompression time.
Förderung
- PRISM and CogniSense
- Semiconductor Research Corporation
- DARPA
Zitieren
APA
Park, H., Ayers, G., Kim, N. S., Levis, P., & Morris, B. (2026). Capacity-Latency Tradeoffs in CXL Memory Expander at Hyperscale. IEEE Computer Architecture Letters, 25(1), 122-125. https://doi.org/10.1109/lca.2026.3676988
Vancouver
Park H, Ayers G, Kim NS, Levis P, Morris B. Capacity-Latency Tradeoffs in CXL Memory Expander at Hyperscale. IEEE Computer Architecture Letters. 2026;25(1):122-125. doi:10.1109/lca.2026.3676988
BibTeX
@article{Park2026,
author = {Park, Haneul and Ayers, Grant and Kim, Nam Sung and Levis, Philip and Morris, Brian},
title = {Capacity-Latency Tradeoffs in CXL Memory Expander at Hyperscale},
journal = {IEEE Computer Architecture Letters},
year = {2026},
volume = {25},
number = {1},
pages = {122--125},
doi = {10.1109/lca.2026.3676988}
}
EUVIMED