Search NASA⌕ Search

Engineering topics

Ladd, Joshua S

Publications and source records attributed to Ladd, Joshua S.

Optimizing Blocking and Nonblocking Reduction Operations for Multicore Systems: Hierarchical Design and Implementation

Many scientific simulations, using the Message Passing Interface (MPI) programming model, are sensitive to the performance and scalability of reduction collective operations such as MPI Allreduce and MPI Reduce. These operations are the most widely used abstractions to perform mathematical operations over all processes that are part of the simulation. In this work, we propose a hierarchical design to implement the reduction operations on multicore systems. This design aims to improve the efficiency of reductions by 1) tailoring the algorithms and customizing the implementations for various communication mechanisms in the system 2) providing the ability to configure the depth of hierarchy to match the system architecture, and 3) providing the ability to independently progress each of this hierarchy. Using this design, we implement MPI Allreduce and MPI Reduce operations (and its nonblocking variants MPI Iallreduce and MPI Ireduce) for all message sizes, and evaluate on multiple architectures including InfiniBand and Cray XT5. We leverage and enhance our existing infrastructure, Cheetah, which is a framework for implementing hierarchical collective operations to implement these reductions. The experimental results show that the Cheetah reduction operations outperform the production-grade MPI implementations such as Open MPI default, Cray MPI, and MVAPICH2, demonstrating its efficiency, flexibility and portability. On Infini- Band systems, with a microbenchmark, a 512-process Cheetah nonblocking Allreduce and Reduce achieves a speedup of 23x and 10x, respectively, compared to the default Open MPI reductions. The blocking variants of the reduction operations also show similar performance benefits. A 512-process nonblocking Cheetah Allreduce achieves a speedup of 3x, compared to the default MVAPICH2 Allreduce implementation. On a Cray XT5 system, a 6144-process Cheetah Allreduce outperforms the Cray MPI by 145%. The evaluation with an application kernel, Conjugate Gradient solver, shows that the Cheetah reductions speeds up total time to solution by 195%, demonstrating the potential benefits for scientific simulations.

Gorentla Venkata, Manjunath↗

Assessing the Performance and Scalability of a Novel Multilevel K-Nomial Allgather on CORE-Direct Systems

In this paper, we propose a novel allgather algorithm, Reindexed Recursive K-ing (RRK), which leverages flexibility in the algorithm's tree topology and ability to make asynchronous progress coupled with Core-Direct communication offload capability to optimize the MPI-Allgather for Core-Direct enabled systems. In particular, the RRK introduces a reindexing scheme which ensures contiguous data transfers while adding only a single additional send and receive operation for any radix, k, or communicator size, N. This allows us to improve algorithm scalability by avoiding the use of a scatter/gather elements (SGE) list on InfiniBand networks. The implementations of the RRK algorithm and its evaluation shows that it performs and scales well on Core-Direct systems for a wide range of message sizes and various communicator configurations.

99 GENERAL AND MISCELLANEOUS↗

ConnectX-2 CORE-Direct Enabled Asynchronous Broadcast Collective Communications

This paper describes the design and implementation of InfiniBand (IB) CORE-Direct based blocking and nonblocking broadcast operations within the Cheetah collective operation framework. It describes a novel approach that fully ofFLoads collective operations and employs only user-supplied buffers. For a 64 rank communicator, the latency of CORE-Direct based hierarchical algorithm is better than production-grade Message Passing Interface (MPI) implementations, 150% better than the default Open MPI algorithm and 115% better than the shared memory optimized MVAPICH implementation for a one kilobyte (KB) message, and for eight mega-bytes (MB) it is 48% and 64% better, respectively. Flat-topology broadcast achieves 99.9% overlap in a polling based communication-computation test, and 95.1% overlap for a wait based test, compared with 92.4% and 17.0%, respectively, for a similar Central Processing Unit (CPU) based implementation.

99 GENERAL AND MISCELLANEOUS↗

Collective Framework and Performance Optimizations to Open MPI for Cray XT Platforms

The performance and scalability of collective operations plays a key role in the performance and scalability of many scientific applications. Within the Open MPI code base we have developed a general purpose hierarchical collective operations framework called Cheetah, and applied it at large scale on the Oak Ridge Leadership Computing Facility's Jaguar (OLCF) platform, obtaining better performance and scalability than the native MPI implementation. This paper discuss Cheetah's design and implementation, and optimizations to the framework for Cray XT 5 platforms. Our results show that the Cheetah's Broadcast and Barrier perform better than the native MPI implementation. For medium data, the Cheetah's Broadcast outperforms the native MPI implementation by 93% for 49,152 processes problem size. For small and large data, it out performs the native MPI implementation by 10% and 9%, respectively, at 24,576 processes problem size. The Cheetah's Barrier performs 10% better than the native MPI implementation for 12,288 processes problem size.

99 GENERAL AND MISCELLANEOUS↗

Cheetah: A Framework for Scalable Hierarchical Collective Operations

Collective communication operations, used by many scientific applications, tend to limit overall parallel application performance and scalability. Computer systems are becoming more heterogeneous with increasing node and core-per-node counts. Also, a growing number of data-access mechanisms, of varying characteristics, are supported within a single computer system. We describe a new hierarchical collective communication framework that takes advantage of hardware-specific data-access mechanisms. It is flexible, with run-time hierarchy specification, and sharing of collective communication primitives between collective algorithms. Data buffers are shared between levels in the hierarchy reducing collective communication management overhead. We have implemented several versions of the Message Passing Interface (MPI) collective operations, MPI Barrier() and MPI Bcast(), and run experiments using up to 49, 152 processes on a Cray XT5, and a small InfiniBand based cluster. At 49, 152 processes our barrier implementation outperforms the optimized native implementation by 75%. 32 Byte and one Mega-Byte broadcasts outperform it by 62% and 11%, respectively, with better scalability characteristics. Improvements relative to the default Open MPI implementation are much larger.

99 GENERAL AND MISCELLANEOUS↗

Accurate Fault Prediction of BlueGene/P RAS Logs Via Geometric Reduction

The authors are building two algorithms for fault prediction using raw system-log data. This work is preliminary, and has only been applied to a limited dataset, however the results seem promising. The conclusions are that: (1) obtaining useful data from RAS-logs is challenging; (2) extracting concentrated information improves efficiency and accuracy; and (3) function evaluation algorithms are fast and lend well to scaling.

97 MATHEMATICS AND COMPUTING↗