Zobrazeno 91 - 100
of 240
pro vyhledávání: '"Hari Subramoni"'
Autor:
Jerome Vienne, Kapil Arya, Gene Cooperman, Shawn Matott, Jiajun Cao, Dhabaleswar K. Panda, Rohan Garg, Hari Subramoni
Publikováno v:
ICPADS
Fault tolerance for the upcoming exascale generation has long been an area of active research. One of the components of a fault tolerance strategy is checkpointing. Petascale-level checkpointing is demonstrated through a new mechanism for virtualizat
Publikováno v:
CloudCom
The performance of Hadoop components can be significantly improved by leveraging advanced features such as Remote Direct Memory Access (RDMA) on modern HPC clusters, where high-performance networks like InfiniBand (IB) and RoCE have been deployed wid
Publikováno v:
SC16: International Conference for High Performance Computing, Networking, Storage and Analysis.
Autor:
Ching-Hsiang Chu, Dhabaleswar K. Panda, Akshay Venkatesh, Hari Subramoni, Bracy Elton, Khaled Hamidouche
Publikováno v:
SBAC-PAD
High-performance streaming applications are beginning to leverage the compute power offered by graphics processing units (GPUs) and high network throughput offered by high performance interconnects such as InfiniBand (IB) to boost their performance a
Publikováno v:
CLUSTER
The Message Passing Interface (MPI) standard specifies the use of (source, tag, communicator) tuple to identify whether an incoming message is what the receiver process is expecting. The cost associated with this process, commonly known as "tag match
Publikováno v:
CCGrid
Dense systems with large number of cores per node are becoming increasingly popular. Existing designs of the Process Management Interface (PMI) show poor scalability in terms of performance and memory consumption on such systems with large number of
Exploiting Maximal Overlap for Non-Contiguous Data Movement Processing on Modern GPU-Enabled Systems
Autor:
Hari Subramoni, Dip Sankar Banerjee, Khaled Hamidouche, Dhabaleswar K. Panda, Akshay Venkatesh, C-H. Chu
Publikováno v:
IPDPS
GPU accelerators are widely used in HPC clusters due to their massive parallelism and high throughput-per-watt. Data movement continues to be the major bottleneck on GPU clusters, more so when data is non-contiguous, which is common in scientific app
Autor:
Dhabaleswar K. Panda, Albert Mathews Augustine, Mark Arnold, Hari Subramoni, Khaled Hamidouche, Xiaoyi Lu, Jonathan Perkins
Publikováno v:
Lecture Notes in Computer Science ISBN: 9783319413204
Modern high-end computing is being driven by the tight integration of several hardware and software components. On the hardware front, there are the multi-/many-core architectures (including accelerators and co-processors) and high-end interconnects
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::d5aae3204cbf4490c876b8838bc0aac0
https://doi.org/10.1007/978-3-319-41321-1_16
https://doi.org/10.1007/978-3-319-41321-1_16
Autor:
Hari Subramoni, Jonathan Perkins, Dhabaleswar K. Panda, Khaled Hamidouche, Ammar Ahmad Awan, Akshay Venkatesh
Publikováno v:
EuroMPI
As we move towards efficient exascale systems, heterogeneous accelerators like NVIDIA GPUs are becoming a significant compute component of modern HPC clusters. It has become important to utilize every single cycle of every compute device available in
Publikováno v:
CLUSTER
Noncontiguous data communication has been heavily adopted in scientific applications, especially for those written with MPI. Common strategies to handle noncontiguous data, like packing/unpacking, incur significant performance overhead during communi