Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Parallel applications”

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.

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

At least 199 records · Page 11

ConnectX-2 InfiniBand Management Queues: First Investigation of the New Support for Network Offloaded Collective Operations

This paper introduces the newly developed Infini-Band (IB) Management Queue capability, used by the Host Channel Adapter (HCA) to manage network task data flow dependancies, and progress the communications associated with such flows. These tasks include sends, receives, and the newly supported wait task, and are scheduled by the HCA based on a data dependency description provided by the user. This functionality is supported by the ConnectX-2 HCA, and provides the means for delegating collective communication management and progress to the HCA, also known as collective communication offload. This provides a means for overlapping collective communications managed by the HCA and computation on the Central Processing Unit (CPU), thus making it possible to reduce the impact of system noise on parallel applications using collective operations. This paper further describes how this new capability can be used to implement scalable Message Passing Interface (MPI) collective operations, describing the high level details of how this new capability is used to implement the MPI Barrier collective operation, focusing on the latency sensitive performance aspects of this new capability. This paper concludes with small scale benchmark experiments comparing implementations of the barrier collective operation, using the new network offload capabilities, with established point-to-point based implementations of these same algorithms, which manage the data flow using the central processing unit. These early results demonstrate the promise this new capability provides to improve the scalability of high performance applications using collective communications. The latency of the HCA based implementation of the barrier is similar to that of the best performing point-to-point based implementation managed by the central processing unit, starting to outperform these as the number of processes involved in the collective operation increases.

Graham, Richard L↗

Endpoint-based parallel data processing in a parallel active messaging interface of a parallel computer

Endpoint-based parallel data processing in a parallel active messaging interface (`PAMI`) of a parallel computer, the PAMI composed of data communications endpoints, each endpoint including a specification of data communications parameters for a thread of execution on a compute node, including specifications of a client, a context, and a task, the compute nodes coupled for data communications through the PAMI, including establishing a data communications geometry, the geometry specifying, for tasks representing processes of execution of the parallel application, a set of endpoints that are used in collective operations of the PAMI including a plurality of endpoints for one of the tasks; receiving in endpoints of the geometry an instruction for a collective operation; and executing the instruction for a collective operation through the endpoints in dependence upon the geometry, including dividing data communications operations among the plurality of endpoints for one of the tasks.

97 MATHEMATICS AND COMPUTING↗

Endpoint-based parallel data processing with non-blocking collective instructions in a parallel active messaging interface of a parallel computer

Endpoint-based parallel data processing with non-blocking collective instructions in a PAMI of a parallel computer is disclosed. The PAMI is composed of data communications endpoints, each including a specification of data communications parameters for a thread of execution on a compute node, including specifications of a client, a context, and a task. The compute nodes are coupled for data communications through the PAMI. The parallel application establishes a data communications geometry specifying a set of endpoints that are used in collective operations of the PAMI by associating with the geometry a list of collective algorithms valid for use with the endpoints of the geometry; registering in each endpoint in the geometry a dispatch callback function for a collective operation; and executing without blocking, through a single one of the endpoints in the geometry, an instruction for the collective operation.

97 MATHEMATICS AND COMPUTING↗

Endpoint-based parallel data processing with non-blocking collective instructions in a parallel active messaging interface of a parallel computer

Methods, apparatuses, and computer program products for endpoint-based parallel data processing with non-blocking collective instructions in a parallel active messaging interface (`PAMI`) of a parallel computer are provided. Embodiments include establishing by a parallel application a data communications geometry, the geometry specifying a set of endpoints that are used in collective operations of the PAMI, including associating with the geometry a list of collective algorithms valid for use with the endpoints of the geometry. Embodiments also include registering in each endpoint in the geometry a dispatch callback function for a collective operation and executing without blocking, through a single one of the endpoints in the geometry, an instruction for the collective operation.

97 MATHEMATICS AND COMPUTING↗

Boxfish

Boxfish is a tool for visualizing performance data. It allows users to quickly project performance data gathered from parallel applications onto intuitive domains. It also includes a projection mechanism that automatically finds relations in performance data so the the data can easily be mapped onto non-obviously related domains. For this release, Boxfish allows performance data to be projected onto the nodes and links of a computer network, as well as onto grids or patch-structured applications. This allows a user of the visualization tool to quickly see correlations between performance data and network structure, application data, or communication patterns.

Source record↗

Hybrid MPI

Enables parallel applications wrriten to use the Message Passing Interface Standard to communicate across different processors to more efficiently leverage shared memory hardware devices.

Source record↗

Processing data communications events by awakening threads in parallel active messaging interface of a parallel computer

Processing data communications events in a parallel active messaging interface (`PAMI`) of a parallel computer that includes compute nodes that execute a parallel application, with the PAMI including data communications endpoints, and the endpoints are coupled for data communications through the PAMI and through other data communications resources, including determining by an advance function that there are no actionable data communications events pending for its context, placing by the advance function its thread of execution into a wait state, waiting for a subsequent data communications event for the context; responsive to occurrence of a subsequent data communications event for the context, awakening by the thread from the wait state; and processing by the advance function the subsequent data communications event now pending for the context.

97 MATHEMATICS AND COMPUTING↗

Stack Trace Ahalysis Tool

STAT is a light weight debugging tool that gathers and merges stack traces from all of the processes in a parallel application. STAT uses the MRNet tree based overlay network to broadcast commands from the tool front-end to the STAT daemons and for the front-end to gather the traces from the STAT daemons. As the traces propagate through the MRNet network tree, they are merged across all tasks to form a similar function call patterns and to delineate a small set of equivalence classes. A representative task from each of these classes can then be fed into a full feature debugger like TolalView for root cause analysis.

Lee, Gregory L.↗

symPACK: A GPU-Capable Fan-Out Sparse Cholesky Solver

Sparse symmetric positive definite systems of equations are ubiquitous in scientific workloads and applications. Parallel sparse Cholesky factorization is the method of choice for solving such linear systems. Therefore, the development of parallel sparse Cholesky codes that can efficiently run on today’s large-scale heterogeneous distributed-memory platforms is of vital importance. Modern supercomputers offer nodes that contain a mix of CPUs and GPUs. To fully utilize the computing power of these nodes, scientific codes must be adapted to offload expensive computations to GPUs. We present symPACK, a GPU-capable parallel sparse Cholesky solver that uses one-sided communication primitives and remote procedure calls provided by the UPC++ library. We also utilize the UPC++ "memory kinds" feature to enable efficient communication of GPU-resident data. We show that on a number of large problems, symPACK outperforms comparable state-of-the-art GPU-capable Cholesky factorization codes by up to 14x on the NERSC Perlmutter supercomputer.

Bellavita, Julian↗

ConnectX2 In niBand Management Queues: New support for Network Of oaded

This paper introduces the newly developed InfiniBand (IB) Management Queue capability, used by the Host Channel Adapter (HCA) to manage network task data flow dependancies, and progress the communications associated with such flows. These tasks include sends, receives, and the newly supported wait task, and are scheduled by the HCA based on a data dependency description provided by the user. This functionality is supported by the ConnectX-2 HCA, and provides the means for delegating collective communication management and progress to the HCA, also known as collective communication offload. This provides a means for overlapping collective communications managed by the HCA and computation on the Central Processing Unit (CPU), thus making it possible to reduce the impact of system noise on parallel applications using collective operations. This paper further describes how this new capability can be used to implement scalable Message Passing Interface (MPI) collective operations, describing the high level details of how this new capability is used to implement the MPI Barrier collective operation, focusing on the latency sensitive performance aspects of this new capability. This paper concludes with small scale benchmark experiments comparing implementations of the barrier collective operation, using the new network offload capabilities, with established point-to-point based implementations of these same algorithms, which manage the data flow using the central processing unit. These early results demonstrate the promise this new capability provides to improve the scalability of high-performance applications using collective communications. The latency of the HCA based implementation of the barrier is similar to that of the best performing point-to-point based implementation managed by the central processing unit, starting to outperform these as the number of processes involved in the collective operation increases.

99 GENERAL AND MISCELLANEOUS↗

Massively-parallel Lagrangian particle code and applications

Massively-parallel, distributed-memory algorithms for the Lagrangian particle hydrodynamic method (Samulyak et al., 2018) have been developed, verified, and implemented. The key component of parallel algorithms is a particle management module that includes a parallel construction of octree databases, dynamic adaptation and refinement of octrees, and particle migration between parallel subdomains. The particle management module is based on the p4est (parallel forest of k-trees) library. The massively-parallel Lagrangian particle code has been applied to a variety of fundamental science and applied problems. A summary of Lagrangian particle code applications to the injection of impurities into thermonuclear fusion devices and to the simulation of supersonic hydrogen jets in support of laser-plasma wakefield acceleration research has also been presented.

97 MATHEMATICS AND COMPUTING↗

Semantic-Aware Automatic Parallelization of Modern Applications Using High-Level Abstractions

Automatic introduction of OpenMP for sequential applications has attracted significant attention recently because of the proliferation of multicore processors and the simplicity of using OpenMP to express parallelism for shared-memory systems. However, most previous research has only focused on C and Fortran applications operating on primitive data types. Modern applications using high-level abstractions, such as C++ STL containers and complex user-defined class types, are largely ignored due to the lack of research compilers that are readily able to recognize high-level object-oriented abstractions and leverage their associated semantics. In this paper, we use a source-to-source compiler infrastructure, ROSE, to explore compiler techniques to recognize high-level abstractions and to exploit their semantics for automatic parallelization. Several representative parallelization candidate kernels are used to study semantic-aware parallelization strategies for high-level abstractions, combined with extended compiler analyses. Preliminary results have shown that semantics of abstractions can help extend the applicability of automatic parallelization to modern applications and expose more opportunities to take advantage of multicore processors.

99 GENERAL AND MISCELLANEOUS↗

Runtime optimization of an application executing on a parallel computer

Identifying a collective operation within an application executing on a parallel computer; identifying a call site of the collective operation; determining whether the collective operation is root-based; if the collective operation is not root-based: establishing a tuning session and executing the collective operation in the tuning session; if the collective operation is root-based, determining whether all compute nodes executing the application identified the collective operation at the same call site; if all compute nodes identified the collective operation at the same call site, establishing a tuning session and executing the collective operation in the tuning session; and if all compute nodes executing the application did not identify the collective operation at the same call site, executing the collective operation without establishing a tuning session.

97 MATHEMATICS AND COMPUTING↗

Runtime optimization of an application executing on a parallel computer

Identifying a collective operation within an application executing on a parallel computer; identifying a call site of the collective operation; determining whether the collective operation is root-based; if the collective operation is not root-based: establishing a tuning session and executing the collective operation in the tuning session; if the collective operation is root-based, determining whether all compute nodes executing the application identified the collective operation at the same call site; if all compute nodes identified the collective operation at the same call site, establishing a tuning session and executing the collective operation in the tuning session; and if all compute nodes executing the application did not identify the collective operation at the same call site, executing the collective operation without establishing a tuning session.

97 MATHEMATICS AND COMPUTING↗

Runtime optimization of an application executing on a parallel computer

Identifying a collective operation within an application executing on a parallel computer; identifying a call site of the collective operation; determining whether the collective operation is root-based; if the collective operation is not root-based: establishing a tuning session and executing the collective operation in the tuning session; if the collective operation is root-based, determining whether all compute nodes executing the application identified the collective operation at the same call site; if all compute nodes identified the collective operation at the same call site, establishing a tuning session and executing the collective operation in the tuning session; and if all compute nodes executing the application did not identify the collective operation at the same call site, executing the collective operation without establishing a tuning session.

97 MATHEMATICS AND COMPUTING↗

Skel: Generative Software for Producing Skeletal I/O Applications

Massively parallel computations consist of a mixture of computation, communication, and I/O. As part of the co-design for the inevitable progress towards exascale computing, we must apply lessons learned from past work to succeed in this new age of computing. Of the three components listed above, implementing an effective parallel I/O solution has often been overlooked by application scientists and was usually added to large scale simulations only when existing serial techniques had failed. As scientists teams scaled their codes to run on hundreds of processors, it was common to call on an I/O expert to implement a set of more scalable I/O routines. These routines were easily separated from the calculations and communication, and in many cases, an I/O kernel was derived from the application which could be used for testing I/O performance independent of the application. These I/O kernels developed a life of their own used as a broad measure for comparing different I/O techniques. Unfortunately, as years passed and computation and communication changes required changes to the I/O, the separate I/O kernel used for benchmarking remained static no longer providing an accurate indicator of the I/O performance of the simulation making I/O research less relevant for the application scientists. In this paper we describe a new approach to this problem where I/O kernels are replaced with skeletal I/O applications automatically generated from an abstract set of simulation I/O parameters. We realize this abstraction by leveraging the ADIOS middleware's XML I/O specification with additional runtime parameters. Skeletal applications offer all of the benefits of I/O kernels including allowing I/O optimizations to focus on useful I/O patterns. Moreover, since they are automatically generated, it is easy to produce an updated I/O skeleton whenever the simulation's I/O changes. In this paper we analyze the performance of automatically generated I/O skeletal applications for the S3D and GTS codes. We show that these skeletal applications achieve performance comparable to that of the production applications. We wrap up the paper with a discussion of future changes to make the skeletal application better approximate the actual I/O performed in the simulation.

99 GENERAL AND MISCELLANEOUS↗