Date: Sat, 27 Nov 2010 21:13:47 +0900
--------------------------------------------------
2010年 11月 15日に SC10 (Supercomputing '10) 併設ワークショップとして、
PDSI (Petascale Data Storage Institute) 主催で開催された Petascale
Data Storage Workshopのペーパー・スライドが公開されています:
5th Petascale Data Storage Workshop, Supercomputing '10
November 15, 2010
http://www.pdsi-scidac.org/events/PDSW10/
Agenda (著者等一部省略)
Keynote Speaker: John Shalf (LBNL/NERSC)
"Exascale Computing Hardware Challenges"
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/PDSW_SC2010_ExascaleChallenges-Shalf.pdf
POSTER SESSION 1 - List of participants and links to posters
http://www.pdsi-scidac.org/events/PDSW10/postersession.html
SESSION 1: Keeping Data
"Self-Adjusting Two-Failure Tolerant Disk Arrays"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/corderi.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/corderi.pdf
"Using a Shared Storage Class Memory Device to Improve the Reliability of RAID Arrays"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/chaarawi.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/chaarawi.pdf
"Semantic Data Placement for Power Management in Archival Storage"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/wildani.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/wildani.pdf
SESSION 2: Accessing Data
"Workload Characterization of a Leadership Class Storage Cluster"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/kim.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/kim.pdf
"Performance Analysis of Commodity and Enterprise Class Flash Devices"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/kim.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/kim.pdf
"Virtualization-based Bandwidth Management for Parallel Storage Systems"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/xu.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/xu.pdf
SESSION 3: Moving Data
"Extracting Information ASAP!"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/abbasi.pdf
"Collective Prefetching for Parallel I/O Systems"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/chen.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/chen.pdf
"Towards Parallel Access of Multi-dimensional, Multi-resolution Scientific Data"
http://www.pdsi-scidac.org/events/PDSW10/resources/papers/kumar.pdf
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/kumar.pdf
Short Announcements
"A New Community Resource for Experiments at Scale: PRObE"
Speaker: Garth Gibson
http://www.pdsi-scidac.org/events/PDSW10/resources/slides/gibson-probe.pdf
※すごいプロジェクトが開始されています。
古い DOEスーパコンピュータを、NSF予算で大規模計算機研究用に Re-Use:
"NSF Funds Computer Systems Research Center At New Mexico Consortium in
Los Alamos", Sept. 30, 2010
http://www.cmu.edu/news/archive/2010/September/sept30_nsfcomputersystemsconsortium.shtml
"The National Science Foundation (NSF) has announced a $10 million
award to the New Mexico Consortium (NMC) at the Los Alamos National
Laboratory (LANL), Carnegie Mellon University and the University of
Utah to build and operate the Parallel Reconfigurable Observational
Environment (PRObE), a one-of-a-kind computer systems research center."
"This innovative concept utilizes decommissioned supercomputing systems
from the Department of Energy (DOE) facilities to provide a very
large-scale systems research capability."
...
"PRObE builds on an existing partnership between the LANL and the NMC
to support educational and research collaborations with universities.
Carnegie Mellon provides expertise as a leader in computer systems
research. The University of Utah will adapt software developed for its
network emulation testbed - Emulab - to PRObE."
"National Science Foundation funds Computer Systems Research Center at
the New Mexico Consortium in Los Alamos"
October 7, 2010
http://www.lanl.gov/news/stories/national_science_foundation_funds_computer_systems_research_center_at_the_new_mexico_consortium_in_l.html
http://institute.lanl.gov/isti/collaborations/probe/
PRObE (Parallel Reconfigurable Observational Environment)
http://newmexicoconsortium.org/probe
Los Alamosでの役割を終え、PRObEで再利用されるマシン例:
Machines
http://newmexicoconsortium.org/probe/machines
Fall 2011: Sitka (2048 cores)
1024 Nodes, Dual Socket, Single Core AMD Opteron
2 GB RAM per core
Full fat-tree Myrinet high-speed interconnect
Fall 2012: Kodiak (2048 cores)
1024 Nodes, Dual Socket, Single Core AMD Opteron
4 GB RAM per core
Fat-tree SDR Infiniband high-speed interconnect
以下は Many Core向けに新規購入するシステム
Summer 2011: Susitna (1728 cores)
36 Nodes, Quad Socket, 12 core AMD Magny-Cours (or similar)
1-2GB RAM / core
EDR Infiniband high-speed interconnect
Summer 2013: Matanuska (3456 cores)
36 Nodes, Quad Socket, 24 core AMD Magny-Cours (or similar)
1-2GB RAM / core
100 GigaBit Ethernet (or similar)
※2011年夏で EDR InfiniBand、2013年夏で 100 GbE・・・
Emulab - Network Emulation Testbed
http://www.emulab.org/
0 件のコメント:
コメントを投稿