Despite the common practice of oversubscription, power capacity is largely under-utilized in data centers. A significant factor driving this under-utilization is fluctuation of the aggregate power demand, resulting in unused “spot (power) capacity”. In this paper, we tap into spot capacity for improving power infrastructure utilization in multi-tenant data centers, an important but under-explored type of data center where multiple tenants house their own physical servers. We propose a novel market, called SpotDC, to allocate spot capacity to tenants on demand. Specifically, SpotDC extracts tenants’ racklevel spot capacity demand through an elastic demand function, based on which the operator sets the market price for spot capacity allocation. We evaluate SpotDC using both testbed experiments and simulations, demonstrating that SpotDC improves power infrastructure utilization and creates a “win-win” situation: the data center operator increases its profit (by nearly 10%), while tenants improve their performance (by 1.2–1.8x on average compared to the no spot capacity case, yet at a marginal cost).
{"title":"A Spot Capacity Market to Increase Power Infrastructure Utilization in Multi-tenant Data Centers","authors":"M. A. Islam, Xiaoqi Ren, Shaolei Ren, A. Wierman","doi":"10.1145/3078505.3078542","DOIUrl":"https://doi.org/10.1145/3078505.3078542","url":null,"abstract":"Despite the common practice of oversubscription, power capacity is largely under-utilized in data centers. A significant factor driving this under-utilization is fluctuation of the aggregate power demand, resulting in unused “spot (power) capacity”. In this paper, we tap into spot capacity for improving power infrastructure utilization in multi-tenant data centers, an important but under-explored type of data center where multiple tenants house their own physical servers. We propose a novel market, called SpotDC, to allocate spot capacity to tenants on demand. Specifically, SpotDC extracts tenants’ racklevel spot capacity demand through an elastic demand function, based on which the operator sets the market price for spot capacity allocation. We evaluate SpotDC using both testbed experiments and simulations, demonstrating that SpotDC improves power infrastructure utilization and creates a “win-win” situation: the data center operator increases its profit (by nearly 10%), while tenants improve their performance (by 1.2–1.8x on average compared to the no spot capacity case, yet at a marginal cost).","PeriodicalId":154694,"journal":{"name":"2018 IEEE International Symposium on High Performance Computer Architecture (HPCA)","volume":"735 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122704608","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Minsoo Rhu, Mike O'Connor, Niladrish Chatterjee, Jeff Pool, S. Keckler
Popular deep learning frameworks require users to fine-tune their memory usage so that the training data of a deep neural network (DNN) fits within the GPU physical memory. Prior work tries to address this restriction by virtualizing the memory usage of DNNs, enabling both CPU and GPU memory to be utilized for memory allocations. Despite its merits, virtualizing memory can incur significant performance overheads when the time needed to copy data back and forth from CPU memory is higher than the latency to perform DNN computations. We introduce a high-performance virtualization strategy based on a "compressing DMA engine" (cDMA) that drastically reduces the size of the data structures that are targeted for CPU-side allocations. The cDMA engine offers an average 2.6x (maximum 13.8x) compression ratio by exploiting the sparsity inherent in offloaded data, improving the performance of virtualized DNNs by an average 53% (maximum 79%) when evaluated on an NVIDIA Titan Xp.
{"title":"Compressing DMA Engine: Leveraging Activation Sparsity for Training Deep Neural Networks","authors":"Minsoo Rhu, Mike O'Connor, Niladrish Chatterjee, Jeff Pool, S. Keckler","doi":"10.1109/HPCA.2018.00017","DOIUrl":"https://doi.org/10.1109/HPCA.2018.00017","url":null,"abstract":"Popular deep learning frameworks require users to fine-tune their memory usage so that the training data of a deep neural network (DNN) fits within the GPU physical memory. Prior work tries to address this restriction by virtualizing the memory usage of DNNs, enabling both CPU and GPU memory to be utilized for memory allocations. Despite its merits, virtualizing memory can incur significant performance overheads when the time needed to copy data back and forth from CPU memory is higher than the latency to perform DNN computations. We introduce a high-performance virtualization strategy based on a \"compressing DMA engine\" (cDMA) that drastically reduces the size of the data structures that are targeted for CPU-side allocations. The cDMA engine offers an average 2.6x (maximum 13.8x) compression ratio by exploiting the sparsity inherent in offloaded data, improving the performance of virtualized DNNs by an average 53% (maximum 79%) when evaluated on an NVIDIA Titan Xp.","PeriodicalId":154694,"journal":{"name":"2018 IEEE International Symposium on High Performance Computer Architecture (HPCA)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132416853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：