Home > Storage > PowerScale (Isilon) > Industry Solutions and Verticals > Electronic Design Automation > Synthetic EDA Workloads for At-Scale Storage Benchmarking > Overview
Electronic Design Automation (EDA) workloads have a huge amount of concurrent I/O that accesses the same project workspaces from many simulation farms or HPC clusters. Project workspaces have a huge number of small files that storage must handle accordingly.
In an EDA design flow, a front-end design shows very high concurrency, a large number of small files, metadata intensive I/O, mostly random but with mixed characteristics. The back-end design shows large file sizes with low/medium concurrency and long runtime jobs.
Android is an open-source operating system for mobile devices and a corresponding open-source project led by Google. Similar to the EDA workspace, Android source code build will handle a large number of small files, including source code, libraries, and a pattern of metadata intensive I/O, as shown in the following figure. During a build, over 90% of I/O is metadata, and roughly 10% is for R/W I/O.
Looking at the following figure, Android source code build is a good candidate to synthesize EDA workloads for an at-Scale storage benchmark.
To test performance in a realistic manner, this new benchmark will submit a large number of Android code build jobs to an EDA simulation farm or HPC cluster using a job scheduler – much like EDA jobs are submitted in a production environment. This will create a high concurrency I/O pattern and achieve benchmark results that better reflect real-world conditions – something that traditional benchmarks fail to do.