| The high resolution solar image reconstruction takes an important place in the solar physics research,in order to treat the huge volume of quasi real-time solar observation data and cope with the computing burden of the same magnitude for high resolution reconstruction of solar image,a number of advanced ground-based solar telescopes at home and abroad have adopted the Triple-Spectral,a method of reconstruction algorithm that can be parallely realized,to reconstruct the high resolution image.Good treatment results are obtained from this method.However,it is still hard to meet the demand for solar observation data treatment at the current efficiency,since the volume of solar telescope observation data is increasing,the solar high resolution observation has been hindered severely for a long time due to huge observation data volume,slow reconstruction speed and other factors.In order to improve the efficiency of solar image reconstruction,the solar image reconstruction of parallel Triple-Spectral method based on the data of different dimension solar telescope is studied in this paper.And obtained the following research results:(1)This study has analyzed the bottleneck of Triple-Spectral traditional parallel implementation method.In order to deeply analyze the characteristics of Triple-Spectral traditional parallel implementation method,this study has applied observation data of different parameters to perform detailed test and analysis for reconstruction process of traditional parallel implementation method based on existing experimental cluster conditions.The test results show that the improvement of image reconstruction efficiency is seriously affected due to the significant increasing of communication time in processing once the process is increased in certain degree.(2)This study has proposed global-shared low-exchange parallel solar image reconstruction method.In view of shortcomings of traditional parallel implementation methods,this study has proposed the full shared memory parallel solar image reconstruction method from the perspective of optimizing parallel processing framework,reducing communication process and simplifying task distribution.The method uses one-time batch distribution task and shared memory mechanism,which greatly simplifies the parallel processing of sub-block reconstruction task and significantly eliminates the communication process during reconstruction.The test results show that this method could greatly reduce the communication time and cost in reconstruction process,effectively solve substantial increase of communication time caused by the increasing of process,and effectively improve the efficiency of image reconstruction.(3)This study has designed and fulfilled a high resolution solar image reconstruction system.In view of data processing requirements of the ONSET(Optical and Near-Infrared Solar Eruption Tracer)and NVST(The 1m New Vacuum Solar Telescope)solar telescope,this study uses the combination way of MPI and shared memory mechanism in the process on the basis of algorithm design of(2),and have designed and fulfilled the image reconstruction system which could process the solar telescope data in batches.Actual applications show that under the cluster or single service,this parallel method can achieve better reconstruction efficiency.In particular,in the absence of large-scale clustering conditions,by using this parallel reconstruction method,better reconstruction efficiency can be obtained under limited computing resources.With the aid of the research outcome of the paper,it is expected to tackle the puzzles remained in the high resolution reconstruction of ONSET and NVST,and bring down the data storage burden.Effectively improve the operational efficiency and scientific output of the two solar telescopes,and provide a good technical reference for the efficient processing of other solar telescope observation data. |
