SEARCH · Engineering Papers

Results for “INFINIBAND”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4

All sources NASA NTRS DOE OSTI

DOE OSTI2019

Graph Algorithms in PGAS: Chapel and UPC++

The Partitioned Global Address Space (PGAS) programming model can be implemented «««< HEAD either with programming language features or with runtime library APIs, each implementation favoring different aspects (e.g., productivity, abstraction, flexibility, or performance). Certain language and runtime features, such as collectives, explicit and asynchronous communication primitives, and constructs facilitating overlap of communication and computation (such as futures and conjoined futures) can enable better performance and scaling for irregular applications, in particular for distributed graph analytics. We compare graph algorithms in one of each of these environments: the Chapel PGAS programming language and the UPC++ PGAS runtime library. We implement algorithms for breadth-first search and triangle counting graph kernels in both environments. We discuss the code in each of the environments, and compile performance data on a Cray Aries and an Infiniband platform. Our results show that the library-based approach of UPC++ currently provides strong performance while Chapel provides a high-level abstraction that, harder to optimize, still provides comparable performance. We argue that various language primitives and runtime library features should be considered in conjunction with each other to design future graph analytics libraries so as to incorporate the best of both worlds.

Distributed Algorithms, Graph algorithms, HPC↗

DOE OSTI2010

Scaling Linear Algebra Kernels using Remote Memory Access

This paper describes the scalability of linear algebra kernels based on remote memory access approach. The current approach differs from the other linear algebra algorithms by the explicit use of shared memory and remote memory access (RMA) communication rather than message passing. It is suitable for clusters and scalable shared memory systems. The experimental results on large scale systems (Linux-Infiniband cluster, Cray XT) demonstrate consistent performance advantages over ScaLAPACK suite, the leading implementation of parallel linear algebra algorithms used today. For example, on a Cray XT4 for a matrix size of 102400, our RMA-based matrix multiplication achieved over 55 teraflops while ScaLAPACK’s pdgemm measured close to 42 teraflops on 10000 processes.

99 GENERAL AND MISCELLANEOUS↗

DOE OSTI2018

Tartan: Evaluating Modern GPU Interconnect via a Multi-GPU Benchmark Suite

In this paper, we fill the gap by proposing a multi-GPU benchmark suite named Tartan, which contains microbenchmarks, scale-up and scale-out applications. We then apply Tartan to evaluate the four latest types of modern GPU interconnects, i.e., PCI- e, NVLink-V1, NVLink-V2 and InfiniBand with GPUDirect- RDMA from two recently released NVIDIA super AI platforms as well as ORNL’s exascale prototype system. Based on empirical evaluation, we observe four new types of NUMA effects: three types are triggered by NVLink’s topology, connectivity and routing, while one type is caused by PCI-e (i.e., anti-locality). They are very important for performance tuning in multi-GPU environment. Our evaluation results show that, unless the current CPU-GPU master-slave programming model can be replaced, it is difficult for scale-up multi-GPU applications to really benefit from faster intra-node interconnects such as NVLinks; while for inter-node scale-out applications, although interconnect is more crucial to the overall performance, GPUDirect-RDMA appears to be not always the optimal choice.

HPC↗

DOE OSTI2017

Exploring DataVortex Systems for Irregular Applications

Emerging applications for data analytics and knowledge discovery typically have irregular or unpredictable communication patterns that do not scale well on parallel systems designed for traditional bulk-synchronous HPC applications. New network architectures that focus on minimizing (short) message latencies, rather than maximizing (large) transfer bandwidths, are emerging as possible alternatives to better support those applications with irregular communication patterns. We explore a system based upon one such novel network architecture, the Data Vortex interconnection network, and examine how this system performs by running benchmark code written for the Data Vortex network, as well as a reference MPI-over- Infiniband implementation, on the same cluster.

Data Vortex, High-Performance Computing, Networkin↗

DOE OSTI2010

Parallelizing the QUDA Library for Multi-GPU Calculations in Lattice Quantum Chromodynamics

Graphics Processing Units (GPUs) are having a transformational effect on numerical lattice quantum chromodynamics (LQCD) calculations of importance in nuclear and particle physics. The QUDA library provides a package of mixed precision sparse matrix linear solvers for LQCD applications, supporting single GPUs based on NVIDIA's Compute Unified Device Architecture (CUDA). This library, interfaced to the QDP++/Chroma framework for LQCD calculations, is currently in production use on the "9g" cluster at the Jefferson Laboratory, enabling unprecedented price/performance for a range of problems in LQCD. Nevertheless, memory constraints on current GPU devices limit the problem sizes that can be tackled. In this contribution we describe the parallelization of the QUDA library onto multiple GPUs using MPI, including strategies for the overlapping of communication and computation. We report on both weak and strong scaling for up to 32 GPUs interconnected by InfiniBand, on which we sustain in excess of 4 Tflops.

43 PARTICLE ACCELERATORS↗

DOE OSTI2012

A High Performance Computing Network and System Simulator for the Power Grid: NGNS^2

Designing and planing next generation power grid sys- tems composed of large power distribution networks, monitoring and control networks, autonomous generators and consumers of power requires advanced simulation infrastructures. The objective is to predict and analyze in time the behavior of networks of systems for unexpected events such as loss of connectivity, malicious attacks and power loss scenarios. This ultimately allows one to answer questions such as: “What could happen to the power grid if ...”. We want to be able to answer as many questions as possible in the shortest possible time for the largest possible systems. In this paper we present a new High Performance Computing (HPC) oriented simulation infrastructure named Next Generation Network and System Simulator (NGNS2 ). NGNS2 allows for the distribution of a single simulation among multiple computing elements by using MPI and OpenMP threads. NGNS2 provides extensive configuration, fault tolerant and load balancing capabilities needed to simulate large and dynamic systems for long periods of time. We show the preliminary results of the simulator running approximately two million simulated entities both on a 64-node commodity Infiniband cluster and a 48-core SMP workstation.

Simulation, power grid, netowrks↗

DOE OSTI2018

Wide-area lustre file system using LNet routers

Scientific and big data computations are increasingly being distributed across wide-area networks, and they often require access to remote files. The file systems that are directly mounted over wide-area networks transparently support such computations, and also obviate the need for special purpose file transfer tools. In typical distributed file systems, the access is limited to local sites, and in particular, the reach of Lustre file system implemented over InfiniBand (IB) is limited to at most tens of miles due to 2.5ms latency bound. We describe LNet router methods that connect IB Lustre file system to remote Ethernet clients over wide-area networks. We collect extensive Lustre throughput measurements over 10Gbps connections with 0-366ms round-trip times. They demonstrate that Gbps throughput can be sustained over connections spanning the globe. We present Lustre throughput profiles over local and wide-area connections, which show the effects of various buffers and credits; in particular, they highlight the throughput limits for large transfers over wide-area connections. Furthermore, the measurements show the positive effects of pipelining in achieving higher throughput for successively file transfers compared to rates indicated by IOzone benchmark rates.

Rao, Nageswara↗

DOE OSTI2017

Unraveling Network-induced Memory Contention: Deeper Insights with Machine Learning

Remote Direct Memory Access (RDMA) is expected to be an integral communication mechanism for future exascale systems enabling asynchronous data transfers, so that applications may fully utilize CPU resources while simultaneously sharing data amongst remote nodes. We examine Network-induced Memory Contention (NiMC) on Infiniband networks. We expose the interactions between RDMA, main-memory and cache, when applications and out-of-band services compete for memory resources. We then explore NiMCs resulting impact on application-level performance. For a range of hardware technologies and HPC workloads, we quantify NiMC and show that NiMCs impact grows with scale resulting in up to 3X performance degradation at scales as small as 8K processes even in applications that previously have been shown to be performance resilient in the presence of noise. In addition, this work examines the problem of predicting NiMC's impact on applications by leveraging machine learning and easily accessible performance counters. This approach provides additional insights about the root cause of NiMC and facilitates dynamic selection of potential solutions. Finally, we evaluated three potential techniques to reduce NiMCs impact, namely hardware offloading, core reservation and network throttling.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2014

OpenSHMEM Reference Implementation using UCCS-uGNI Transport Layer. In: PGAS '14 Proceedings of the 8th International Conference on Partitioned Global Address Space Programming Models, Article No. 18

OpenSHMEM is a library interface implementation and specification that enables the implementation of the Partitioned Global Address Space (PGAS) model. It exports modern RDMA network functionality and communication semantics to applications very efficiently. There are many closed source implementations of OpenSHMEM for modern RDMA interconnects such as InfiniBand and Cray's Gemini and Aries. Given the important role that Cray systems play in HPC, in this paper, we present an open source implementation of OpenSHMEM for Cray XE/XK/XC systems. To implement OpenSHMEM, we use the uGNI interface. uGNI is a generic interface that is designed for multiple programming models. The interface fits well the goal of UCCS. Having OpenSHMEM with UCCS-uGNI allows usage of the same implementation over multiple interconnects. This also translates into many advantages that come with common code such as resource sharing, increasing productivity because of less code maintenance, etc. Preliminary results show that OpenSHMEM-UCCS performs comparable to state-of-the-art Cray SHMEM for Put, Get, and AMO operations.

Janjusic, Tomislav↗

DOE OSTI2015

Diagnosing the Causes and Severity of One-sided Message Contention

Two trends suggest network contention for one-sided messages is poised to become a performance problem that concerns application developers: an increased interest in one-sided programming models and a rising ratio of hardware threads to network injection bandwidth. Unfortunately, it is difficult to reason about network contention and one-sided messages because one-sided tasks can either decrease or increase contention. We present effective and portable techniques for diagnosing the causes and severity of one-sided message contention. To detect that a message is affected by contention, we maintain statistics representing instantaneous (non-local) network resource demand. Using lightweight measurement and modeling, we identify the portion of a message's latency that is due to contention and whether contention occurs at the initiator or target. We attribute these metrics to program statements in their full static and dynamic context. We characterize contention for an important computational chemistry benchmark on InfiniBand, Cray Aries, and IBM Blue Gene/Q interconnects. We pinpoint the sources of contention, estimate their severity, and show that when message delivery time deviates from an ideal model, there are other messages contending for the same network links. With a small change to the benchmark, we reduce contention up to 50% and improve total runtime as much as 20%.

Network Contention, Performance Analysis, Dynamic ↗

DOE OSTI2015

Exploiting communication concurrency on high performance computing systems

Although logically available, applications may not exploit enough instantaneous communication concurrency to maximize hardware utilization on HPC systems. This is exacerbated in hybrid programming models such as SPMD+OpenMP. We present the design of a "multi-threaded" runtime able to transparently increase the instantaneous network concurrency and to provide near saturation bandwidth, independent of the application configuration and dynamic behavior. The runtime forwards communication requests from application level tasks to multiple communication servers. Our techniques alleviate the need for spatial and temporal application level message concurrency optimizations. Experimental results show improved message throughput and bandwidth by as much as 150% for 4KB bytes messages on InfiniBand and by as much as 120% for 4KB byte messages on Cray Aries. For more complex operations such as all-to-all collectives, we observe as much as 30% speedup. This translates into 23% speedup on 12,288 cores for a NAS FT implemented using FFTW. We also observe as much as 76% speedup on 1,500 cores for an already optimized UPC+OpenMP geometric multigrid application using hybrid parallelism.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2015

A Case for Application Oblivious Energy-Efficient MPI Runtime

Power has become the major impediment in designing large scale high-end systems. Message Passing Interface (MPI) is the {\em de facto} communication interface used as the back-end for designing applications, programming models and runtime for these systems. Slack --- the time spent by an MPI process in a single MPI call --- provides a potential for energy and power savings, if an appropriate power reduction technique such as core-idling/Dynamic Voltage and Frequency Scaling (DVFS) can be applied without perturbing application's execution time. Existing techniques that exploit slack for power savings assume that application behavior repeats across iterations/executions. However, an increasing use of adaptive, data-dependent workloads combined with system factors (OS noise, congestion) makes this assumption invalid. This paper proposes and implements Energy Aware MPI (EAM) --- an application-oblivious energy-efficient MPI runtime. EAM uses a combination of communication models of common MPI primitives (point-to-point, collective, progress, blocking/non-blocking) and an online observation of slack for maximizing energy efficiency. Each power lever incurs time overhead, which must be amortized over slack to minimize degradation. When predicted communication time exceeds a lever overhead, the lever is used {\em as soon as possible} --- to maximize energy efficiency. When mis-prediction occurs, the lever(s) are used automatically at specific intervals for amortization. We implement EAM using MVAPICH2 and evaluate it on ten applications using up to 4096 processes. Our performance evaluation on an InfiniBand cluster indicates that EAM can reduce energy consumption by 5--41\% in comparison to the default approach, with negligible (less than 4\% in all cases) performance loss.

Venkatesh, Akshay↗

DOE OSTI2019

An evaluation of the CORAL interconnects

The US Department of Energy deployed the Summit and Sierra supercomputers with the latest state-of-the-art network interconnect technology in 2018 and both systems entered production in 2019. In this paper, we provide an in-depth assessment of the systems' network interconnects that are based on Enhanced Data Rate (EDR) 100 Gb/s Mellanox InfiniBand. Both systems use second-generation EDR Host Channel Adapters (HCAs) and switches with several new features such as Adaptive Routing (AR), switch-based collectives, and HCA-based tag matching. Although based on the same components, Summit's network is "non-blocking" (i.e., a fully provisioned Clos network) and Sierra's network has a 2:1 taper between the racks and aggregation switches. We evaluate the two systems' interconnects using traditional communication benchmarks as well as production applications. We find that the new Adaptive Routing dramatically improves performance but the other new features still need improvement.

Zimmer, Christopher↗

DOE OSTI2023

Application Experiences on a GPU-Accelerated Arm-based HPC Testbed

This paper assesses and reports the experience of ten teams working to port, validate, and benchmark several High Performance Computing applications on a novel GPU-accelerated Arm testbed system. The testbed consists of eight NVIDIA Arm HPC Developer Kit systems, each one equipped with a server-class Arm CPU from Ampere Computing and two data center GPUs from NVIDIA Corp. The systems are connected together using InfiniBand interconnect. The selected applications and mini-apps are written using several programming languages and use multiple accelerator-based programming models for GPUs such as CUDA, OpenACC, and OpenMP offloading. Working on application porting requires a robust and easy-to-access programming environment, including a variety of compilers and optimized scientific libraries. The goal of this work is to evaluate platform readiness and assess the effort required from developers to deploy well-established scientific workloads on current and future generation Arm-based GPU-accelerated HPC systems. The reported case studies demonstrate that the current level of maturity and diversity of software and tools is already adequate for large-scale production deployments.

Elwasif, Wael↗

DOE OSTI2014

IBTopo

IBTopo is a set of perl modules and scripts that enable one to determine the network topology of an Infiniband network.

Moody, A.T.↗

DOE OSTI2015

OpenSM Monitoring System

The OpenSM Monitoring System includes a collection of diagnostic and monitoring tools for use on Infiniband networks. The information this system gathers is obtained from a service, which in turn is obtained directly from the OpenSM subnet manager.

Meier, T.A.↗

DOE OSTI2015

Single Sided Messaging v. 0.6.6

Single-Sided Messaging (SSM) is a portable, multitransport networking library that enables applications to leverage potential one-sided capabilities of underlying network transports. It also provides desirable semantics that services for highperformance, massively parallel computers can leverage, such as an explicit cancel operation for pending transmissions, as well as enhanced matching semantics favoring large numbers of buffers attached to a single match entry. This release supports TCP/IP, shared memory, and Infiniband.

Curry, MatthewLeon↗

DOE OSTI2016

Portals Reference Implementation v. 1.0

The Portals reference implementation is based on the Portals 4.X API, published by Sandia National Laboratories as a freely available public document. It is designed to be an implementation of the Portals Networking Application Programming Interface and is used by several other upper layer protocols like SHMEM, GASNet and MPI. It is implemented over existing networks, specifically Ethernet and InfiniBand networks. This implementation provides Portals networks functionality and serves as a software emulation of Portals compliant networking hardware. It can be used to develop software using the Portals API prior to the debut of Portals networking hardware, such as Bull’s BXI interconnect, as well as a substitute for portals hardware on development platforms that do not have Portals compliant hardware. The reference implementation provides new capabilities beyond that of a typical network, namely the ability to have messages matched in hardware in a way compatible with upper layer software such as MPI or SHMEM. It also offers methods of offloading network operations via triggered operations, which can be used to create offloaded collective operations. Specific details on the Portals API can be found at http://portals4.org.

Grant, RyanE.↗