### Workunits of the current experiment

В текущем эксперименте сгенерированы все вокрюниты. Скоро мы запустим новый эксперимент.

Acoustics@home is a BOINC-based volunteer computing project aimed at solving inverse problems in underwater acoustics.
You can contribute to our research by running a free program on your computer.

In underwater acoustics the notion of geoacoustic inversion refers
to a collection of techniques that can be used for the reconstruction of medium parameters. The medium parameters are usually
understood as the sound speed profile in water (i.e., the dependence of the sound speed on the depth) and the sound speed and density
in the sea bottom sediment layers. Medium parameters reconstruction is of great importance for problems of underwater communication and
for the development of underwater navigation systems.

While normally measurements for the geoacoustic inversion are performed using expensive receiver arrays, recently it was
shown that single-hydrophone recording of a broadband pulse signal can be also
successfully used for estimating the medium parameters.

The method of geoacoustic inversion developed in our study is based on using modal dispersion data. Waveguide dispersion is usually
induced by the difference in group velocities of propagation of normal waves of different mode numbers at different frequencies.
By using special algorithms of the frequency-time analysis of signals, it is possible to filter out the modal components of a pulse
signal from its time series recorded by a single hydrophone. The implementation of this method in practice can be thought of as a
solution of an optimization problem in a (very large) discrete
search space, and every evaluation of the cost function requires numerous solutions of an acoustic spectral problem. Thus, the whole
computational burden can be easily divided into a large number of relatively simple independent tasks, which can be solved using
volunteer computing.

27 Nov 2017, 15:09:06 UTC

All workunits of the current experiment have been generated. We will start a new experiment soon.

В текущем эксперименте сгенерированы все вокрюниты. Скоро мы запустим новый эксперимент.

В текущем эксперименте сгенерированы все вокрюниты. Скоро мы запустим новый эксперимент.

8 Oct 2017, 14:45:26 UTC

The report 'Vadim Bulavintsev, Pavel Petrov, Mikhail Posypkin, Oleg Zaikin. A GPU-enabled Black-box Optimization in Application to Dispersion-based Geoacoustic Inversion' was made at the conference OPTIMA'2017. In this report, we discuss a prototype of Acoustics@home's GPU application.

На конференции OPTIMA'2017 был сделан доклад 'Вадим Булавинцев, Павел Петров, Михаил Посыпкин, Олег Заикин. GPU-релизация black-box оптимизации в применении к решению обратных задач геоакустики, основывающихся на дисперии'.

На конференции OPTIMA'2017 был сделан доклад 'Вадим Булавинцев, Павел Петров, Михаил Посыпкин, Олег Заикин. GPU-релизация black-box оптимизации в применении к решению обратных задач геоакустики, основывающихся на дисперии'.

6 Oct 2017, 20:28:50 UTC

In the course of 'Ocean of fantasy' challenge, 26.4 % of scenario 906 workunits were calculated. Before the challenge, only 10 % of them were calculated in about two weeks. So, 36.4 % of the search space is recently processed. The project's performance has reached 9 teraflops. We want to thank all volunteers, who took part in the challenge!

В рамках соревнования 'Ocean of fantasy' было обработано 26.4 % воркюнитов сценария 906. Перед соревнованием было обработано только 10 % за 2 недели. Итак, на данный момент пройдено 36.4 % пространства поиска. Производительность проекта достигла 9 терафлопс. Мы благодарим всех кранчеров, которые приняли участие в этом соревновании!

В рамках соревнования 'Ocean of fantasy' было обработано 26.4 % воркюнитов сценария 906. Перед соревнованием было обработано только 10 % за 2 недели. Итак, на данный момент пройдено 36.4 % пространства поиска. Производительность проекта достигла 9 терафлопс. Мы благодарим всех кранчеров, которые приняли участие в этом соревновании!

News is available as an RSS feed