%0 Journal Article
%9 ACL : Articles dans des revues avec comité de lecture répertoriées par l'AERES
%A Martin, R.
%A Monteiller, V.
%A Komatitsch, D.
%A Perrouty, S.
%A Jessell, Mark
%A Bonvalot, Sylvain
%A Lindsay, M.
%T Gravity inversion using wavelet-based compression on parallel hybrid CPU/GPU systems : application to southwest Ghana
%D 2013
%L PAR00011166
%G ENG
%J Geophysical Journal International
%@ 0956-540X
%K Wavelet transform ; Inverse theory ; Numerical approximations and analysis ; Satellite gravity ; Gravity anomalies and Earth structure
%K GHANA
%M ISI:000326964400014
%N 3
%P 1594-1619
%R 10.1093/gji/ggt334
%U https://www.documentation.ird.fr/hor/PAR00011166
%V 195
%W Horizon (IRD)
%X We solve the 3-D gravity inverse problem using a massively parallel voxel (or finite element) implementation on a hybrid multi-CPU/multi-GPU (graphics processing units/GPUs) cluster. This allows us to obtain information on density distributions in heterogeneous media with an efficient computational time. In a new software package called TOMOFAST3D, the inversion is solved with an iterative least-square or a gradient technique, which minimizes a hybrid L-1-/L-2-norm-based misfit function. It is drastically accelerated using either Haar or fourth-order Daubechies wavelet compression operators, which are applied to the sensitivity matrix kernels involved in the misfit minimization. The compression process behaves like a pre-conditioning of the huge linear system to be solved and a reduction of two or three orders of magnitude of the computational time can be obtained for a given number of CPU processor cores. The memory storage required is also significantly reduced by a similar factor. Finally, we show how this CPU parallel inversion code can be accelerated further by a factor between 3.5 and 10 using GPU computing. Performance levels are given for an application to Ghana, and physical information obtained after 3-D inversion using a sensitivity matrix with around 5.37 trillion elements is discussed. Using compression the whole inversion process can last from a few minutes to less than an hour for a given number of processor cores instead of tens of hours for a similar number of processor cores when compression is not used.
%$ 066 ; 020