Site Tools


Sidebar

IPCC

Intel® supports R&D of Lustre in the community by offering IPCC (Parallel Computing Center) grants. The following list describes the grants in short.

André Brinkman from Johannes Gutenberg University

Topic: Lustre QoS (Quality of Service): Network Request Scheduler and Monitoring revisited

Anticipated results: allow the administrator to specify complex policies for prioritizing and balancing IO performance by user, node, MPI job and other factors combined with improved RPC ordering and merging on the server to optimize and balance performance under varying IO workload.

https://software.intel.com/en-us/articles/intel-parallel-computing-center-at-johannes-gutenberg-university-mainz

Thomas Ludwig from University of Hamburg

Topic: Enhanced Adaptive Compression in Lustre

Anticipated results: use client-side information and file data history to leverage the growing number cores on the client to adaptively compress file data to reduce network traffic, reduce required storage space, and improve aggregate server IO throughput. For workloads where client-side compression is sub-optimal, optimize data compression within ZFS on the server for files that benefit from it.

https://software.intel.com/en-us/articles/intel-parallel-computing-center-at-university-of-hamburg-scientific-computing

Walter Schön from GSI Helmholtz Centre for Heavy Ion Research

Topic: Developing a Copytool for TSM Storage Backends

Anticipated results: We propose to develop robust TSM copytool for TSM storage backends which seamlessly integrates into the Lustre HSM framework and TSM storage backends. Such a copytool is of great interests in academic environments and enterprises, and popularizes the Lustre file system for a broader field of applications.

https://software.intel.com/en-us/articles/intel-parallel-computing-center-at-gsi-and-fair

Last modified: 2016/02/17 17:35 by kunkel