Networking, Grids, Middleware and Distributed Platforms
Some Optimization Techniques of the Matrix Multiplication Algorithm
Nenad Anchev, Marjan Gusev, Sasko Ristov, Blagoj Atanasovski
Abstract
Dense matrix-matrix multiplication algorithm is widely used in large scientific applications, and often it is an important factor of the overall performance of the application. Therefore, optimizing this algorithm, both for parallel and serial execution would give an overall performance boost. In this paper we overview the most used dense matrix multiplication optimization techniques applicable for multicore processors. These methods can speedup the multicore parallel execution focusing on reducing the number of memory accesses and improving the algorithm according to hardware architecture and organization.
Keywords
HPC, CPU, Cache, Performance
Full text is available at IEEE Xplore digital library.
Benchmarking Cloud Databases for the Requirements of the Internet of Things
Adrian Copie, Teodor-Florin Fortis, Victor Ion Munteanu
Abstract
The Internet, since its inception, has been continuously evolving, creating both problems and solutions. Recent trends show that an increasing number of uniquely identifiable things (objects, sensors and devices) are making their way to this medium. A unique opportunity for integration within the already existing landscape of cloud services has been created. Further challenges related to the management and governance of these devices have been identified.
The work presented in this paper addresses the issue of storing information received from different things in cloud databases in order to facilitate further exploitation in the context of cloud services.
Keywords
Cloud computing, Internet of Things, Cloud databases, IoT Governance, Data store Benchmarking
Full text is available at IEEE Xplore digital library.
Does the Performance Scale the Same as the Cost in the Cloud
Monika Simjanoska, Goran Velkoski, Sasko Ristov and Marjan Gusev
Abstract
Cloud computing is a paradigm that offers on-demand scalable resources with the "pay-per-usage" model. Cloud service providers’ price rises linearly as the resources scale. However, the main challenge for the cloud customers is "Does the performance scale as the price for the rented resources in the cloud"? Also, how does the performance scales for different server load? In this paper we analyze the performance and the cost of a web service that utilize both memory and CPU varying the server load with different message size and number of concurrent messages in order to determine the real cost of rented CPU resources. The results show that the web service’s cost rises linearly with the resources, i.e. the lowest cost is determined while hosted on one CPU.
Keywords
Cloud Computing, Web Services, Performance, Resources, Cost
Full text is available at IEEE Xplore digital library.
Towards New Energy Efficiency Limits of High Performance Clusters
Draško Tomić, Emir Imamagić, Luko Gjenero
Abstract
In recent years performance of High Performance Computing Clusters took precedence over their power consumption. However, costs of energy and demand for ecologically acceptable IT solutions are higher than ever before, therefore a need for HPC clusters with acceptable power consumption becomes increasingly important. Consequently, the Green500 list, which takes into account both performance and power consumption of HPC clusters, almost reached the popularity of the Top500 list. Interestingly, the Green500 list is not an opponent to Top500 list; its core idea is to complement the Top500. Therefore, the Top500 list still serves as the basis for the Green500 list, and its numbers regarding measured HPL performance, are a basis for calculating the Green500 list. Indeed, the Green500 is the Top500 list ordered by HPL measured performance per Watt. Rmax numbers gained from High Performance Linpack benchmarks serve as performance input parameters, and total power consumed during execution of HPL on a certain HPC clusters is a power consumption parameter. The critical question remains: how to measure the consumed power correctly? This paper proposes that if it is not possible to measure the consumed power, one can still use maximum power consumption numbers rated from hardware vendors to find at least the lower bound green efficiency of HPC clusters. The main idea behind this approach is that Rmax values found on Top500 list never achieve Rpeak theoretical values, and that even most efficient HPL benchmark can never utilize computing nodes at their maximum. Furthermore by comparing MFLOPS/W results we gained with those found on Green500 list, we noted the excellent efficiency of the new HPC Isabella cluster recently powered on at University Computing Centre in Zagreb, ranking in just behind University of North Carolina KillDevil Top500 super cluster.
Keywords
HPC, green computing, energy efficiency
Full text is available at IEEE Xplore digital library.
Affinity-aware HPC Applications in Multichip and Multicore Multiprocessor
Goran Velkoski, Sasko Ristov, Marjan Gusev
Abstract
Introducing multi level cache memory reduces the gap between the CPU and main memory and speeds up the program execution. The speedup in modern multiprocessors can scale up to linear speedup according to Gustafson's law. Each CPU core usually possesses private L1 and L2 cache memory and shares L3 cache memory in multi-core processor architectures. Furthermore, private or shared cache memory could have significant impact to the algorithm performance in parallel implementation. Private cache increases the overall cache size used during the execution. On the other hand, shared cache reduces cache misses if all CPU cores use the same data. In this paper we analyze the matrix vector multiplication algorithm performance for sequential and parallel implementation in multi-chip multi-core multiprocessor in order to determine the CPU affinity that provides the best performance. We also realize theoretical analysis to determine the problem size regions where selecting appropriate CPU affinity can produce the best performance using the same resources.
Keywords
CPU Cache, Matrix - Vector Multiplication, Performance, Speed
Full text is available at IEEE Xplore digital library.
The Effects of Maturation of the Cloud Computing Market on Open Source Cloud Computing Infrastructure Projects
Ivan Voras, Marin Orlic, Branko Mihaljevic
Abstract
Since its rise in around year 2008, “Cloud computing” has become a staple buzzword for large business users and scholarly institutions, and has even succeeded in achieving mindshare and penetration into the consumer culture. The technologies and ideas behind it have become applied, though with varying rates of success, and the initial enthusiasm has given way to practical considerations. This trend can be directly observed from comparing the current state of Open source cloud computing projects with their past state. In this paper we present an overview of how Open source cloud computing landscape has changed in the last three years.
Keywords
cloud computing, trends, open source, OpenStack, Nebula, Eucalyptus, OpenQRM, Nebula, mOSAIC, Abiquo, provisioning
Full text is available at IEEE Xplore digital library.