BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Reflections on Paralelization of Gravity Inversion
DTSTART;VALUE=DATE-TIME:20121018T104500Z
DTEND;VALUE=DATE-TIME:20121018T111500Z
DTSTAMP;VALUE=DATE-TIME:20261010T220151Z
UID:indico-contribution-175@indico.ipb.ac.rs
DESCRIPTION:Speakers: Neki Frasheri (Polytechnic University of Tirana\, Fa
 culty of Information Technology)\nIn the paper there is presented a summar
 y of results obtained for the parallelization of 3D gravity inversion usin
 g the principle of algorithm CLEAN [Hogborn 1974]\, undertaken in framewor
 k o f FP7 project HP-SEE. The problem is “ill posed” with the definiti
 on of [Hadamard 1902].\nThe core of our algorithm consists in cross-calcul
 ation of the effect of 3D array of underground nodes to each of 2D array g
 round surface points. The process is iterative\, updating at each iteratio
 n the mass density of one node with a predefined step in order to decrease
  of the least squares error. The volume of calculations for each iteration
  was calculated of the order O (N^5). The number of iterations to obtain t
 he same anomalous body resulted of the order O (N^3)\, leading to an order
  of calculations O (N^8). \nExperiments were carried out in two HPC system
 s – the HPCG system of IICT-BAS in Sofia\, Bulgaria\, and in the SGE sys
 tem of NIIFI at University of Pécs\, Hungary\, using both OpenMP and MPI.
  Obtained results confirmed the order of calculations of O (N^8). The abso
 lute user-time and wall-time was obtained. For moderate sized models with 
 geosection 4000m*4000m*2000m using 3D arrays of 101*101*51 nodes\, using u
 p to 1\,000 parallel cores\, the run-time reached the level of 100\,000 se
 conds (27 hours). \nTests with models and field data resulted with clear m
 ass density contrasts between the anomalous bodies and the medium where th
 ey were situated\, in the same way as in real geological structures. Tests
  with multi-bodies geosections indicated the tendency of the algorithm to 
 converge towards single body solutions. \nMPI tests carried out in HPCG sy
 stem of IICT-BAS showed a systematic time overhead when the number of para
 llel processes increased from 8 to 32. Hypothesizing that this was caused 
 because of inter-process communication between computer nodes\, a predicti
 on of potential run-time in multi-cluster MPI grid platforms was undertake
 n\, taking into consideration the relative low bandwidth of campus and met
 ropolitan links (compared with the bandwidth of the BUS connecting cores o
 f a computer node). The extrapolation of data led to the hypothesis that t
 he increase in cores using multi-cluster grids may not result in reduction
  of the run-time.\n\nhttps://events.saifa.rs/event/291/contributions/175/
LOCATION:National Library of Serbia
URL:https://events.saifa.rs/event/291/contributions/175/
END:VEVENT
END:VCALENDAR