An I/O kernel for a Big Data Clustering alogirthm used in SC15 paper [1].
A newer version of the I/O kernel with more access patterns has been integrated in h5bench (https://github.com/hpc-io/h5bench/). See h5bench_read.c in h5bench/h5bench_patterns/.
Citation:
Md. Mostofa Ali Patwary, Suren Byna, Nadathur Rajagopalan Satish, Narayanan Sundaram, Zarija Lukic, Vadim Roytershteyn, Michael J. Anderson, Yushu Yao, Prabhat, and Pradeep Dubey, "BD-CATS: Big Data Clustering at Trillion Particle Scale", Supercomputing 2015 (SC15), November 2015 PDF: https://sdm.lbl.gov/~sbyna/research/papers/201511-SC15-BD-CATS.pdf