我们需要多大和多快的计算机
|
摘要
分析了石油勘探开发中油藏描述和建模的精度要求及地震勘探技术发展的主要特征,介绍了当前地震勘探的数据采集技术水平和典型数据规模。在详细分析和测试的基础上,定量预测了未来10-15年内地震勘探处理的数据规模和计算量需求发展趋势,并在分析高性能计算技术发展趋势的基础上,对地震数据处理中实际应用的高性能计算机系统的规模进行了预测。地震勘探对高性能计算具有巨大而持续的需求,综合考虑需求、技术、成本等因素,预计石油工业高性能计算的发展速度不会低于每5年提高10倍,分别于2005年、2010年和2015年达到10×1012/s,100×1012/s和1000×1012/s的规模,以满足1TB,10TB和100TB数量级规模地震数据处理的性能要求。
The author analyzes the precision requirement of reservoir characterization and modeling as well as the features of development of seismic prospecting technology for petroleum exploration and production, and describes the technical status of seismic data acquisition and typical size of seismic data volume. Based on detailed a-nalysis and test, I present a quantitative prediction of the data size and computation requirement of seismic data processing in the next 10 to15 years. With the consideration of precision requirement, technology, cost, and other factors, I believe that seismic prospecting has a great and continuous demand for high performance computing (HPC). HPC will grow at a speed of ten-fold every five years, and will reach the performances of lOTflops, 100Tflops, and 1000Tflops in the year of 2005, 2010, and 2010 respectively to meet the needs of seismic data processing for the data size of 1TB, 10TB, and 100TB.
The author analyzes the precision requirement of reservoir characterization and modeling as well as the features of development of seismic prospecting technology for petroleum exploration and production, and describes the technical status of seismic data acquisition and typical size of seismic data volume. Based on detailed a-nalysis and test, I present a quantitative prediction of the data size and computation requirement of seismic data processing in the next 10 to15 years. With the consideration of precision requirement, technology, cost, and other factors, I believe that seismic prospecting has a great and continuous demand for high performance computing (HPC). HPC will grow at a speed of ten-fold every five years, and will reach the performances of lOTflops, 100Tflops, and 1000Tflops in the year of 2005, 2010, and 2010 respectively to meet the needs of seismic data processing for the data size of 1TB, 10TB, and 100TB.
引文
1 张德忠.地震勘探数据整理[A].见:皮声洪编.地震勘探数据整理与石油数据银行文集[C].北京:石油工业出版社,2001． 2-14
2 Christie Phil. Raising the standards of seismic data quality [J]. Oilfield Review,2001,Summer: 16-31
3 Pickering Stephen. Q-Reservoir: Advanced seismic technology for reservoir solutions [J]. Oil & Gas North Africa Magazine, 2001, Janrary-February: 26-30
4赵改善,包红林.集群技术及其在石油工业中的应用[J].石油物探,2001,40(3) :118-126
5 Russell Brian. Zero distance and infinite resources [J]. The Leading Edge, 2003, 22(2) : 150-151
6 Beasley Craig J. Beyond the "more data, faster computers" syndrome [J]. The Leading Edge, 2003, 22(2) . 152-154
7 Klepacki David J. High performance computing at IBM: The Bluegene/L supercomputer [J]. The Leading Edge, 2003, 22(2) :155-157
8 Paul Donald L. Creating the more integrated energy company [J]. The Leading Edge, 2003, 22(2) . 161-162
2 Christie Phil. Raising the standards of seismic data quality [J]. Oilfield Review,2001,Summer: 16-31
3 Pickering Stephen. Q-Reservoir: Advanced seismic technology for reservoir solutions [J]. Oil & Gas North Africa Magazine, 2001, Janrary-February: 26-30
4赵改善,包红林.集群技术及其在石油工业中的应用[J].石油物探,2001,40(3) :118-126
5 Russell Brian. Zero distance and infinite resources [J]. The Leading Edge, 2003, 22(2) : 150-151
6 Beasley Craig J. Beyond the "more data, faster computers" syndrome [J]. The Leading Edge, 2003, 22(2) . 152-154
7 Klepacki David J. High performance computing at IBM: The Bluegene/L supercomputer [J]. The Leading Edge, 2003, 22(2) :155-157
8 Paul Donald L. Creating the more integrated energy company [J]. The Leading Edge, 2003, 22(2) . 161-162
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心