Hybridisation strategies and data structures for the NEMO ocean model

2017 ◽  
Vol 32 (6) ◽  
pp. 864-881 ◽  
Author(s):  
Italo Epicoco ◽  
Silvia Mocavero ◽  
Andrew R Porter ◽  
Stephen M Pickles ◽  
Mike Ashworth ◽  
...  
Keyword(s):  
High Performance Computing ◽  
Shared Memory ◽  
Data Structures ◽  
High Performance ◽  
Ocean Model ◽  
Data Layout ◽  
European Climate ◽  
Ocean Models ◽  
Performance Computing

This work describes the introduction of a second level of parallelism based on the OpenMP shared memory paradigm to NEMO, one of the most widely used ocean models in the European climate community. Although the existing parallelisation scheme in NEMO, based on the MPI paradigm, has served it well for many years, it is becoming unsuited to current high-performance computing architectures due to their increasing tendency to have fat nodes containing tens of compute cores. Three different parallel approaches for introducing OpenMP are presented, discussed and compared on several platforms. Finally we have also considered the effect on performance of the data layout employed in NEMO.

scholarly journals Data structures access model for remote shared memory

2021 ◽  
Vol 244 ◽  
pp. 07001
Author(s):  
Anatoliy Nyrkov ◽  
Konstantin Ianiushkin ◽  
Andrey Nyrkov ◽  
Yulia Romanova ◽  
Vagiz Gaskarov
Keyword(s):  
Shared Memory ◽  
Data Structures ◽  
Data Model ◽  
High Performance ◽  
Direct Memory Access ◽  
Performance Comparison ◽  
Memory Access ◽  
Memory Storage ◽  
Race Conditions ◽  
Performance Computing

Recent achievements in high-performance computing significantly narrow the performance gap between single and multi-node computing, and open up opportunities for systems with remote shared memory. The combination of in-memory storage, remote direct memory access and remote calls requires rethinking how data organized, protected and queried in distributed systems. Reviewed models let us implement new interpretations of distributed algorithms allowing us to validate different approaches to avoid race conditions, decrease resource acquisition or synchronization time. In this paper, we describe the data model for mixed memory access with analysis of optimized data structures. We also provide the result of experiments, which contain a performance comparison of data structures, operating with different approaches, evaluate the limitations of these models, and show that the model does not always meet expectations. The purpose of this paper to assist developers in designing data structures that will help to achieve architectural benefits or improve the design of existing distributed system.

Abstractions and Middleware for Petascale Computing and Beyond

2012 ◽  
pp. 161-178
Author(s):  
Ivo F. Sbalzarini
Keyword(s):  
Numerical Simulations ◽  
High Performance Computing ◽  
Data Structures ◽  
High Performance ◽  
Source Code ◽  
Limiting Factors ◽  
Communication Overhead ◽  
Parallel Data ◽  
Petascale Computing ◽  
Performance Computing

As high-performance computing moves to the petascale and beyond, a number of algorithmic and software challenges need to be addressed. This paper reviews the main performance-limiting factors in today’s high-performance computing software and outlines a possible new programming paradigm to address them. The proposed paradigm is based on abstract parallel data structures and operations that encapsulate much of the complexity of an application, but still make communication overhead explicit. The authors argue that all numerical simulations can be formulated in terms of the presented abstractions, which thus define an abstract semantic specification language for parallel numerical simulations. Simulations defined in this language can automatically be translated to source code containing the appropriate calls to a middleware that implements the underlying abstractions. Finally, the structure and functionality of such a middleware are outlined while demonstrating its feasibility on the example of the parallel particle-mesh library (PPM).

Testing and Performance Analysis of FCA Algorithm Accelerating IBM Platform MPI

2014 ◽  
Vol 539 ◽  
pp. 429-433
Author(s):  
Lin Feng Zhang ◽  
Lei Chen
Keyword(s):  
Performance Analysis ◽  
High Performance Computing ◽  
Shared Memory ◽  
High Performance ◽  
New Method ◽  
Computing Environment ◽  
Collective Communications ◽  
And Performance ◽  
Performance Computing

MPI is one of the important standards in high performance computing. MPI performance is generally focused on collective communications. And FCA (Fabric Collective Accelerator) is a new method accelerating collective communications. Through high performance computing environment testing, this paper mainly analyses the result of FCA with shared memory and without share memory accelerating IBM Platform MPI, FCA's principle and integration between IBM Platform MPI and FCA. At the same time, this paper may be a good reference for high performance computing using FCA.

Abstractions and Middleware for Petascale Computing and Beyond

2012 ◽  
pp. 1998-2015
Author(s):  
Ivo F. Sbalzarini
Keyword(s):  
Numerical Simulations ◽  
High Performance Computing ◽  
Data Structures ◽  
High Performance ◽  
Source Code ◽  
Limiting Factors ◽  
Communication Overhead ◽  
Parallel Data ◽  
Petascale Computing ◽  
Performance Computing

As high-performance computing moves to the petascale and beyond, a number of algorithmic and software challenges need to be addressed. This paper reviews the main performance-limiting factors in today’s high-performance computing software and outlines a possible new programming paradigm to address them. The proposed paradigm is based on abstract parallel data structures and operations that encapsulate much of the complexity of an application, but still make communication overhead explicit. The authors argue that all numerical simulations can be formulated in terms of the presented abstractions, which thus define an abstract semantic specification language for parallel numerical simulations. Simulations defined in this language can automatically be translated to source code containing the appropriate calls to a middleware that implements the underlying abstractions. Finally, the structure and functionality of such a middleware are outlined while demonstrating its feasibility on the example of the parallel particle-mesh library (PPM).

The Recent Revolution in High Performance Computing

MRS Bulletin ◽  
1997 ◽  
Vol 22 (10) ◽  
pp. 5-6
Author(s):  
Horst D. Simon
Keyword(s):  
High Performance Computing ◽  
High Performance ◽  
New Technologies ◽  
New Technology ◽  
Parallel Architecture ◽  
Time Frame ◽  
Good News ◽  
Computing Industry ◽  
Time And Energy ◽  
Performance Computing

Recent events in the high-performance computing industry have concerned scientists and the general public regarding a crisis or a lack of leadership in the field. That concern is understandable considering the industry's history from 1993 to 1996. Cray Research, the historic leader in supercomputing technology, was unable to survive financially as an independent company and was acquired by Silicon Graphics. Two ambitious new companies that introduced new technologies in the late 1980s and early 1990s—Thinking Machines and Kendall Square Research—were commercial failures and went out of business. And Intel, which introduced its Paragon supercomputer in 1994, discontinued production only two years later.During the same time frame, scientists who had finished the laborious task of writing scientific codes to run on vector parallel supercomputers learned that those codes would have to be rewritten if they were to run on the next-generation, highly parallel architecture. Scientists who are not yet involved in high-performance computing are understandably hesitant about committing their time and energy to such an apparently unstable enterprise.However, beneath the commercial chaos of the last several years, a technological revolution has been occurring. The good news is that the revolution is over, leading to five to ten years of predictable stability, steady improvements in system performance, and increased productivity for scientific applications. It is time for scientists who were sitting on the fence to jump in and reap the benefits of the new technology.

Sign in / Sign up

Export Citation Format

Share Document