NOTE: Please type in one of the program names in your browser to view it. sar_exactlogdeteigvals_2DN4_2500.f is the latest version of SAR program. The SAR programs here use OpenMP shared-memory programming. The fortran programs form the sythetic data x and y using NAG subroutines. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The files in this directory are: ================================ OriginsVersion/ - The SAR software in fortran with OpenMP that run on SGI Origins machines dlw3gx-m.outs/ - output from golden section search and fitting step of SAR epsilon_2500_0mean_1std_2DN004.data - vector epsilon data form_yx_nag_2500.f - the fortran program to form the synthetic data x and y isaku_wada4_formepsilon.c - the C file that forms the high quality normally distributed random numbers - thanks to Prof. Isaku Wada iw_run4_formepsilon_2500_4 - the run script to run in the queue of IBM SP readme - this file results.prllSARmodel.4PE.2DN4.2500.Notcomplete.ibmsp-1 - 1st result file from runs on IBM SP on 4 processors results.prllSARmodel.4PE.2DN4.2500.Notcomplete.ibmsp-2 - 2nd result file from runs on IBM SP on 4 processors results.prllSARmodel.4PE.2DN4.2500.Notcomplete.ibmsp-3 - 3rd result file from runs on IBM SP on 4 processors We can go down to 5 seconds on 16 processors from serial execution time of 48 seconds sar2DN4_V2_2500_rtcfunc.old.f - Please use this file to run SAR in parallel - tested :-) sar_exactlogdeteigvals_2DN4_2500.f - latest SAR file (still on construction for calculating the minimum of log-likelihood fucntion which was also not calculated in matlab files.) sp_2D4N_V7_26_4PE - run scripts to run parallel SAR software on IBM SP on 4 processors sp_2D4N_V7_26_4PE_scrpt sp_2DN4.2500_nag_form_yx - run scripts to form the synthetic data x and y sp_2DN4.2500_nag_form_yx_scrpt ss_mat_2500 - run script to run matlab files in queue on IBM SP example4_SAR_model_solns_via_ML.m - basically the same matlab file as shown in matlab programs train_test_SAR.m x_2500_2DN004.data - vector x data y_training_2500_2DN004.data - vector y data =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=