@Article{IJNAM-2-177,
author = {Y. Yang, T. Dai, Z. Han, J. Shu and Z. Pan},
title = {The Parallel Strategy of a Large Scale Simulation About Ten Million Nodes to Reservoir with Multiple Layers},
journal = {International Journal of Numerical Analysis and Modeling},
year = {2005},
volume = {2},
number = {0},
pages = {177--182},
abstract = {<p style="text-align: justify;">Aim at large scale fine reservoir numerical simulation
application research on Shenwei computer, the multilayer two dimension
two phase parallel software transplanted successfully and a large scale
integral simulation about ten million nodes were realized in the
environment of Shenwei parallel computer. The whole preconditioning
alternating Schward and another many improved algorithm, the parallel
optimal methods about coefficient matrix and saturation calculation made
the parallel efficiency increased effectively about multilayer two
dimension two phase parallel software. Especially the deep research
about the communication and load-balanced technology fitting for Shenwei
computer make the parallel function of the software to large scale
increase. The multilayer two dimension two phase parallel software
transplanted and the parallel computer resource of homegrown Shenwei
high behavior parallel computer with 112 CPUs was to simulate the
production history of 12 sandgroups of the second Shahejian in second
block of Shengtuo. The simulation scale is 10 million nodes and the
time exhausted is about 5 hours which satisfies the application
requisition of reservoir simulation. This verifies the reliability and
stability of the software and makes the whole parallel efficiency to
79%. It is first time to bring out the independent copyright reservoir
simulation parallel software with satisfactory back and forth processing
function in homegrown Shenwei computer. Especially the application of
the whole preconditioning alternating Schward region decomposition
algorithm, the deep research of load-balanced technology and the large
scale application etc. are all innovative.</p>},
issn = {2617-8710},
doi = {https://doi.org/},
url = {http://global-sci.org/intro/article_detail/ijnam/960.html}
}