按年被提到
20222022 · 1 次2026
系统、数据库与网络›计算机体系结构与硬件·arXiv:2204.03290 · cs.AR

Memory Performance of AMD EPYC Rome and Intel Cascade Lake SP Server Processors

Markus Velten, Robert Schöne, Thomas Ilsche, Daniel Hackenberg · 2022
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2022-05-10 · 最近 2024-10-19
2不同的人
2活跃月份
2讨论串
36被引(OpenAlex)
arXiv 摘要(候选层不写导读)

Modern processors, in particular within the server segment, integrate more cores with each generation. This increases their complexity in general, and that of the memory hierarchy in particular. Software executed on such processors can suffer from performance degradation when data is distributed disadvantageously over the available resources. To optimize data placement and access patterns, an in-depth analysis of the processor design and its implications for performance is necessary. This paper describes and experimentally evaluates the memory hierarchy of AMD EPYC Rome and Intel Xeon Cascade Lake SP server processors in detail. Their distinct microarchitectures cause different performance patterns for memory latencies, in particular for remote cache accesses. Our findings illustrate the complex NUMA properties and how data placement and cache coherence states impact access latencies to local and remote locations. This paper also compares theoretical and effective bandwidths for accessing data at the different memory levels and main memory bandwidth saturation at reduced core counts. The presented insight is a foundation for modeling performance of the given microarchitectures, which enables practical performance engineering of complex applications. Moreover, security research on side-channel attacks can also leverage the presented findings.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论magicalhippo · 2024-10

“Curious about the latency of this. This[1] page, which is from 2022, seems to indicate that the CLX memory controllers adds about 200ns, and you got some retimers which add a few tens of ns. Then there's the cabling, assuming 2/3 the speed of light that's 5ns per meter. With the memory latency itself it seems this adds up about 350ns, while a local memor…”

Inventec 96 DIMM CXL Expansion Box at OCP Summit 2024 for TBs of Memory原评论 ↗
被提交为 story · 1 次