اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدDesign and Evaluation of a Collective IO Model for Loosely Coupled Petascale Programming
144
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Loosely coupled programming is a powerful paradigm for rapidly creating higher-level applications from scientific programs on petascale systems, typically using scripting languages. This paradigm is a form of many-task computing (MTC) which focuses on the passing of data between programs as ordinary files rather than messages. While it has the significant benefits of decoupling producer and consumer and allowing existing application programs to be executed in parallel with no recoding, its typical implementation using shared file systems places a high performance burden on the overall system and on the user who will analyze and consume the downstream data. Previous efforts have achieved great speedups with loosely coupled programs, but have done so with careful manual tuning of all shared file system access. In this work, we evaluate a prototype collective IO model for file-based MTC. The model enables efficient and easy distribution of input data files to computing nodes and gathering of output results from them. It eliminates the need for such manual tuning and makes the programming of large-scale clusters using a loosely coupled model easier. Our approach, inspired by in-memory approaches to collective operations for parallel programming, builds on fast local file systems to provide high-speed local file caches for parallel scripts, uses a broadcast approach to handle distribution of common input data, and uses efficient scatter/gather and caching techniques for input and output. We describe the design of the prototype model, its implementation on the Blue Gene/P supercomputer, and present preliminary measurements of its performance on synthetic benchmarks and on a large-scale molecular dynamics application.
قيم البحث
اقرأ أيضاً
We have extended the Falkon lightweight task execution framework to make loosely coupled programming on petascale systems a practical and useful programming model. This work studies and measures the performance factors involved in applying this appro
ach to enable the use of petascale systems by a broader user community, and with greater ease. Our work enables the execution of highly parallel computations composed of loosely coupled serial jobs with no modifications to the respective applications. This approach allows a new-and potentially far larger-class of applications to leverage petascale systems, such as the IBM Blue Gene/P supercomputer. We present the challenges of I/O performance encountered in making this model practical, and show results using both microbenchmarks and real applications from two domains: economic energy modeling and molecular dynamics. Our benchmarks show that we can scale up to 160K processor-cores with high efficiency, and can achieve sustained execution rates of thousands of tasks per second.
In order to achieve near-time insights, scientific workflows tend to be organized in a flexible and dynamic way. Data-driven triggering of tasks has been explored as a way to support workflows that evolve based on the data. However, the overhead intr
oduced by such dynamic triggering of tasks is an under-studied topic. This paper discusses different facets of dynamic task triggers. Particularly, we explore different ways of constructing a data-driven dynamic workflow and then evaluate the overheads introduced by such design decisions. We evaluate workflows with varying data size, percentage of interesting data, temporal data distribution, and number of tasks triggered. Finally, we provide advice based upon analysis of the evaluation results for users looking to construct data-driven scientific workflows.
Programming models for building large-scale distributed applications assist the developer in reasoning about consistency and distribution. However, many of the programming models for weak consistency, which promise the largest scalability gains, have
little in the way of evaluation to demonstrate the promised scalability. We present an experience report on the implementation and large-scale evaluation of one of these models, Lasp, originally presented at PPDP `15, which provides a declarative, functional programming style for distributed applications. We demonstrate the scalability of Lasps prototype runtime implementation up to 1024 nodes in the Amazon cloud computing environment. It achieves high scalability by uniquely combining hybrid gossip with a programming model based on convergent computation. We report on the engineering challenges of this implementation and its evaluation, specifically related to operating research prototypes in a production cloud environment.
The ROOT I/O (RIO) subsystem is foundational to most HEP experiments - it provides a file format, a set of APIs/semantics, and a reference implementation in C++. It is often found at the base of an experiments framework and is used to serialize the e
xperiments data; in the case of an LHC experiment, this may be hundreds of petabytes of files! Individual physicists will further use RIO to perform their end-stage analysis, reading from intermediate files they generate from experiment data. RIO is thus incredibly flexible: it must serve as a file format for archival (optimized for space) and for working data (optimized for read speed). To date, most of the technical work has focused on improving the former use case. We present work designed to help improve RIO for analysis. We analyze the real-world impact of LZ4 to decrease decompression times (and the corresponding cost in disk space). We introduce new APIs that read RIO data in bulk, removing the per-event overhead of a C++ function call. We compare the performance with the existing RIO APIs for simple structure data and show how this can be complimentary with efforts to improve the parallelism of the RIO stack.
A major driver behind the success of modern machine learning algorithms has been their ability to process ever-larger amounts of data. As a result, the use of distributed systems in both research and production has become increasingly prevalent as a
means to scale to this growing data. At the same time, however, distributing the learning process can drastically complicate the implementation of even simple algorithms. This is especially problematic as many machine learning practitioners are not well-versed in the design of distributed systems, let alone those that have complicated communication topologies. In this work we introduce Launchpad, a programming model that simplifies the process of defining and launching distributed systems that is specifically tailored towards a machine learning audience. We describe our framework, its design philosophy and implementation, and give a number of examples of common learning algorithms whose designs are greatly simplified by this approach.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
