… the workloads change very fast (so-called workload churn) [… a new benchmark or proxy for
every possible workload [18]. On … We consider each big data, AI and Internet service workload …

… In this paper, the authors aim to bridge the gap between data motifs and proxy benchmarks,
… Big Data and AI Workloads [6]. The proposed solution was a data motif-based benchmark …

… range from simple computational kernels, mini-apps, and proxy-… of BD and AI workloads. So
instead of applying a separate … suggest using data motif-based workloads, which is why they …

… In this paper, we use big data benchmarks as samples to illustrate the validity of our …
Data motif-based proxy benchmarks for big data and AI workloads. In: Proceedings of the …

… We analyze the performance profiles of multiple variants of our proxy application, developed
using … Data motif-based proxy benchmarks for big data and AI workloads. In 2018 IEEE …

… the impact of data traffic on workload performance by collectively investigating a large number
of … [18], [50] establish a proxy benchmark for AI workloads by identifying eight data motifs. …

… The proxies in this section as well as section 4.2 both compress the original … Data motifbased
proxy benchmarks for big data and AI workloads. “In 2018 IEEE International Symposium …

… We evaluate UniRedl with representative deep learning and big data workloads. The
experimental results demonstrate that on average UniRedl improves application performance by …

… workload generator to simulate the users’ operations. In the process of load data and query
data, we use the performance … Q3 workloads respectively belong to load data workloads and …

… The following sections will introduce chosen data sets, workloads, the workload generator, …
Data motifs: a lens towards fully understanding big data and AI workloads. In: 2018 27th …