Wyniki wyszukiwania - BazTech

1

Improving energy efficiency of supercomputer systems through software-aided liquid cooling management

Januszewski R., Różycki R., Waligóra G.

Foundations of Computing and Decision Sciences

|

2018

|

Vol. 43, No. 2

89--103

EN

Many fields of modern science rely more and more on the immense computing power of supercomputers. Modern, multi-thousand node systems can consume megawatts of electrical energy in highly uneven manner, challenging the data center infrastructure, both power and cooling coils. The traditional way of managing the infrastructure makes each subsystem of a data center (e.g. cooling) independent from all other in the way it relies only on local sensors to manage the infrastructure. The erratic nature of computing in a large data center makes this approach suboptimal. In the paper we show that by challenging the traditional split between the infrastructure and the computing equipment, one can gain significant boost in energy efficiency of the entire ecosystem. A solution that predicts cooling power demand basing on the information from a supercomputer resource manager, and then sets up the parameters of the cooling loop, is presented along with potential benefits in terms of reduction of the power draw.

2

Creating Acoustic Maps Employing Supercomputing Cluster

Czyżewski A., Kotus J., Szczodrak M.

Archives of Acoustics

|

2011

|

Vol. 36, No. 2

395-418

EN

The implemented online urban noise pollution monitoring system is presented with regard to its conceptual assumptions and technical realization. A concept of the noise source parameters dynamic assessment is introduced. The idea of noise modeling, based on noise emission characteristics and emission simulations, was developed and practically utilized in the system. Furthermore, the working system architecture and the data acquisition scheme are described. The method for increasing the speed of noise map calculation employing a supercomputer is explained. The practical implementation of noise maps generation and visualization system is presented, together with introduced improvements in the domain of continuous noise monitoring and acoustic maps creation. Some results of tests performed using the system prototype are shown. The main focus is put on assessing the efficiency of the acoustic maps created with the discussed system, in comparison to results obtained with traditional methods.

3

Potokowa realizacja operacji pomnóż i dodaj dla argumentów zmiennoprzecinkowych podwójnej precyzji

Russek P., Wiatr K.

Pomiary Automatyka Kontrola

|

2007

|

R. 53, nr 7

36-38

PL

Operacja pomnóż i dodaj to fundament realizacji obliczeń numerycznych we współczesnej nauce i technice. Możliwość szybkiej realizacji tej opera-cji ma zasadnicze znaczenie dla efektywności systemu obliczeniowego. Obok techniki przyśpieszania obliczeń polegającej na równoległej ich realizacji duże znaczenie i zastosowanie ma również technika przetwarzania potokowego. Zwiększa ona przepustowość modułów obliczeniowych wydłużając opóźnienie. W przypadku operatora pomnóż i dodaj zastosowanie techniki potokowej ze względu na pętle sprzężenia zwrotnego w ścieżce danych napotyka pewne problemy. W pracy zaprezentowano sposób potokowej realizacji operacji pomnóż i dodaj oraz wyniki jej implementacji w FPGA dla argumentów zmiennoprzecinkowych podwójnej precyzji.

EN

Multiply and accumulate operation is a foundation of contemporary numerical computation in science and technology. Ability for its fast execution is crucial for performance of computing system. In computing acceleration beside parallel processing technique also pipelining has an important role as a way to increase system throughput. In a case of multiply-and-accumulate (MAC) operation there is a problematic issue that comes from the feedback loop necessary in MAC architecture. In this paper double precision MAC pipeline architecture is proposed and FPGA implementation results presented.