Adaptive-Latency DRAM: Reducing DRAM Latency by Exploiting Timing\n Margins

作者
Donghyuk Lee,Yoongu Kim,Gennady Pekhimenko,Samira Khan,Vivek Seshadri,Kevin Chen–Chuan Chang,Onur Mutlu
出处
期刊:Cornell University - arXiv [Cornell University]
标识
DOI:10.48550/arxiv.1805.03047
摘要

This paper summarizes the idea of Adaptive-Latency DRAM (AL-DRAM), which was\npublished in HPCA 2015, and examines the work's significance and future\npotential. AL-DRAM is a mechanism that optimizes DRAM latency based on the DRAM\nmodule and the operating temperature, by exploiting the extra margin that is\nbuilt into the DRAM timing parameters. DRAM manufacturers provide a large\nmargin for the timing parameters as a provision against two worst-case\nscenarios. First, due to process variation, some outlier DRAM chips are much\nslower than others. Second, chips become slower at higher temperatures. The\ntiming parameter margin ensures that the slow outlier chips operate reliably at\nthe worst-case temperature, and hence leads to a high access latency.\n Using an FPGA-based DRAM testing platform, our work first characterizes the\nextra margin for 115 DRAM modules from three major manufacturers. The\nexperimental results demonstrate that it is possible to reduce four of the most\ncritical timing parameters by a minimum/maximum of 17.3%/54.8% at 55C while\nmaintaining reliable operation. AL-DRAM uses these observations to adaptively\nselect reliable DRAM timing parameters for each DRAM module based on the\nmodule's current operating conditions. AL-DRAM does not require any changes to\nthe DRAM chip or its interface; it only requires multiple different timing\nparameters to be specified and supported by the memory controller. Our real\nsystem evaluations show that AL-DRAM improves the performance of\nmemory-intensive workloads by an average of 14% without introducing any errors.\nOur characterization and proposed techniques have inspired several other works\non analyzing and/or exploiting different sources of latency and performance\nvariation within DRAM chips.\n

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
温柔的吐司完成签到,获得积分10
刚刚
花花发布了新的文献求助10
1秒前
butterflycat完成签到,获得积分10
1秒前
正直小懒猪完成签到,获得积分10
1秒前
薛小飞完成签到 ,获得积分10
1秒前
可靠的安珊完成签到 ,获得积分10
2秒前
2秒前
RYYYYYYY233完成签到 ,获得积分0
2秒前
炙热灵枫发布了新的文献求助10
2秒前
3秒前
6秒前
彭于晏的应助被缺粥采纳,获得10
7秒前
7秒前
7秒前
cici完成签到,获得积分10
8秒前
8秒前
羊腿完成签到 ,获得积分10
9秒前
TGJ发布了新的文献求助10
9秒前
慕青的应助被ssswww采纳,获得10
9秒前
1210xi完成签到,获得积分10
10秒前
世间安得双全法完成签到,获得积分0
11秒前
Maestro_S完成签到,获得积分0
11秒前
11秒前
11秒前
weqhdgjfk完成签到,获得积分10
11秒前
无头苍蝇完成签到,获得积分10
11秒前
躺平青年阿俊完成签到,获得积分10
11秒前
12秒前
st完成签到,获得积分10
12秒前
jzhou65完成签到,获得积分10
13秒前
13秒前
sense发布了新的文献求助10
14秒前
花花完成签到,获得积分10
14秒前
14秒前
14秒前
CC66发布了新的文献求助10
15秒前
15秒前
捞鱼完成签到,获得积分10
16秒前
16秒前
16秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 2000
Sex, Aging, and Death in a Medieval Medical Compendium: Trinity College Cambridge MS R.14.52, Its Texts, Language, and Scribe 500
高分子基功能界面水蒸发材料与应用研究 500
外场增强的水汽发电研究 500
基于结构调控的功能化水诱导发电材料与器件研究 500
基于界面调控的高效湿气发电材料与器件研究 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8190478
求助须知:如何正确求助?哪些是违规求助? 9577686
关于积分的说明 21853808
捐赠科研通 7789496
什么是DOI,文献DOI怎么找? 3436407
关于科研通互助平台的介绍 2560235
邀请新用户注册赠送积分活动 2483054