The Gemini System Interconnect

Alverson, Robert; Roweth, Duncan; Kaplan, Larry

doi:10.1109/hoti.2010.23

Cited by 169 publications

(81 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have demonstrated Nessie services on the Cray XT and XE systems at Sandia National Laboratories (SNL) and Oak Ridge National , the Cray XT4/5 systems at Oak Ridge National Laboratory, and a large InfiniBand cluster at SNL. The Nessie RPC layer has direct support of Cray's SeaStar interconnect [7], through the Portals API [8]; Cray's Gemini interconnect [4]; and InfiniBand [5]. The server fetches bulk data through RDMA commands until it has satisfied the request.…”

Section: Nessie Architecturementioning

confidence: 99%

See 1 more Smart Citation

Data co-processing for extreme scale analysis level II ASC milestone (4745).

Rogers¹,

Moreland²,

Oldfield³

et al. 2013

View full text Add to dashboard Cite

Exascale supercomputing will embody many revolutionary changes in the hardware and software of high-performance computing. A particularly pressing issue is gaining insight into the science behind the exascale computations. Power and I/O speed constraints will fundamentally change current visualization and analysis workflows. A traditional post-processing workflow involves storing simulation results to disk and later retrieving them for visualization and data analysis. However, at exascale, scientists and analysts will need a range of options for moving data to persistent storage, as the current offline or post-processing pipelines will not be able to capture the data necessary for data analysis of these extreme scale simulations. This Milestone explores two alternate workflows, characterized as in situ and in transit, and compares them. We find each to have its own merits and faults, and we provide information to help pick the best option for a particular use.

show abstract

Section: Nessie Architecturementioning

confidence: 99%

“…We estimate the boundary as the contour at a threshold between high and low volume fraction. 4. Find the fragment connected components.…”

Section: Experiments Drivermentioning

confidence: 99%

Data co-processing for extreme scale analysis level II ASC milestone (4745).

Rogers¹,

Moreland²,

Oldfield³

et al. 2013

View full text Add to dashboard Cite

show abstract

“…We have demonstrated Nessie services on the Cray XT3 at Sandia National Laboratories, the Cray XT4/5 systems at ORNL, and a large InfiniBand cluster at SNL. The Nessie RPC layer has direct support of Cray's SeaStar interconnect [26], through the Portals API [27]; Cray's Gemini interconnect [6]; and InfiniBand [9]. Network protocol for a Nessie storage server executing a write request.…”

Section: Nessiementioning

confidence: 99%

“…Here we present performance results from three different HPC platforms: the Red Storm system at Sandia [30], a Cray XT3 that uses the Seastar interconnect [26] interfaced through the Portals API [27]; RedSky, a cluster of Oracle/Sun Blade Servers on a an InfiniBand network; and the Cielo supercomputer, a Cray XE6 system that uses the new Gemini interconnect [6]. Figure 7.6 shows a comparison of using the xfer-write-rdma and xfer-write-encode methods to transfer an array of data t data structures from Figure 7.3.…”

Section: Performance Of the Transfer Servicementioning

confidence: 99%

Report of experiments and evidence for ASC L2 milestone 4467 : demonstration of a legacy application's path to exascale.

Curry¹,

Ferreira²,

Pedretti³

et al. 2012

View full text Add to dashboard Cite

This report documents thirteen of Sandia's contributions to the Computational Systems and Software Environment (CSSE) within the Advanced Simulation and Computing (ASC) program between fiscal years 2009 and 2012. It describes their impact on ASC applications. Most contributions are implemented in lower software levels allowing for application improvement without source code changes. Improvements are identified in such areas as reduced run time, characterizing power usage, and Input/Output (I/O). Other experiments are more forward looking, demonstrating potential bottlenecks using mini-application versions of the legacy codes and simulating their network activity on Exascale-class hardware. 3 AcknowledgmentsThe authors would like to thank the Red Storm, Cielo, and Cielo Del Sur operations teams for their support during the dedicated experiments. These systems are valuable resources and the teams' prompt response ensured that we maximized our time on them.

show abstract

“…The usage of MPI DDTs greatly simplifies this example. If the used interconnect supports non-contiguous transfers (such as Cray Gemini [2]) the two copies can be avoided completely. Therefore the usage of DDTs not only simplifies the code but also can improve the performance due to the zero-copy formulation.…”

Section: Introductionmentioning

confidence: 99%

Micro-applications for Communication Data Access Patterns and MPI Datatypes

Schneider

Gerstenberger

Hoefler

2012

Recent Advances in the Message Passing Interface

View full text Add to dashboard Cite

Abstract. Data is often communicated from different locations in application memory and is commonly serialized (copied) to send buffers or from receive buffers. MPI datatypes are a way to avoid such intermediate copies and optimize communications, however, it is often unclear which implementation and optimization choices are most useful in practice. We extracted the send/recv-buffer access pattern of a representative set of scientific applications into micro-applications that isolate their data access patterns. We also observed that the buffer-access patterns in applications can be categorized into three different groups. Our microapplications show that up to 90% of the total communication time can be spent with local serialization and we found significant performance discrepancies between state-of-the-art MPI implementations. Our microapplications aim to provide a standard benchmark for MPI datatype implementations to guide optimizations similarly to SPEC CPU and the Livermore loops do for compiler optimizations.

show abstract

The Gemini System Interconnect

Cited by 169 publications

References 0 publications

Data co-processing for extreme scale analysis level II ASC milestone (4745).

Data co-processing for extreme scale analysis level II ASC milestone (4745).

Report of experiments and evidence for ASC L2 milestone 4467 : demonstration of a legacy application's path to exascale.

Micro-applications for Communication Data Access Patterns and MPI Datatypes

Contact Info

Product

Resources

About