w4kek/WSJT-X
1
0
Fork 0
mirror of https://github.com/saitohirga/WSJT-X.git synced 2024-11-26 14:18:38 -05:00
Code Issues Packages Projects Releases Wiki Activity
b47b4b077b
WSJT-X/lib/jt9.f90

315 lines
11 KiB
Fortran
Raw Normal View History

Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
program jt9
! Decoder for JT9. Can run stand-alone, reading data from *.wav files;
! or as the back end of wsjt-x, with data placed in a shared memory region.
use options
use prog_args
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
use, intrinsic :: iso_c_binding
use FFTW3
Make Fortran profiling timer function a callback with a default null implementation Groundwork for calling the decoders directly from C/C++ threads. To access the timer module timer_module must now be used. Instrumented code need only use the module function 'timer' which is now a procedure pointer that is guaranteed to be associated (unless null() is assigned to it, which should not be done). The default behaviour of 'timer' is to do nothing. If a Fortran program wishes to profile code it should now use the timer_impl module which contains a default timer implementation. The main program should call 'init_timer([filename])' before using 'timer' or calling routines that are instrumented. If 'init_timer([filename])'. If it is called then an optional file name may be provided with 'timer.out' being used as a default. The procedure 'fini_timer()' may be called to close the file. The default timer implementation is thread safe if used with OpenMP multi-threaded code so long as the OpenMP thread team is given the copyin(/timer_private/) attribute for correct operation. The common block /timer_private/ should be included for OpenMP use by including the file 'timer_common.inc'. The module 'lib/timer_C_wrapper.f90' provides a Fortran wrapper along with 'init' and 'fini' subroutines which allow a C/C++ application to call timer instrumented Fortran code and for it to receive callbacks of 'timer()' subroutine invocations. No C/C++ timer implementation is provided at this stage. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6320 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-27 10:40:57 -05:00
use timer_module, only: timer
use timer_impl, only: init_timer, fini_timer
Make WAV file reading correct so that optional chunks can be tolerated The WAV file format allows for optional header content, allow for this in preparation for adding some metadata to WSJT-X recorded WAV files. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6335 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-01-02 09:29:02 -05:00
use readwav
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
include 'jt9com.f90'
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
integer(C_INT) iret
Make WAV file reading correct so that optional chunks can be tolerated The WAV file format allows for optional header content, allow for this in preparation for adding some metadata to WSJT-X recorded WAV files. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6335 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-01-02 09:29:02 -05:00
type(wav_header) wav
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
real*4 s(NSMAX)
character c
character(len=500) optarg, infile
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
character wisfile*80
First working code that uses wider detection bandwidths for submodes JT65B,C. Still needs work! git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6507 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-03-07 15:00:23 -05:00
!### ndepth was defined as 60001. Why???
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
integer :: arglen,stat,offset,remain,mode=0,flow=200,fsplit=2700, &
First working code that uses wider detection bandwidths for submodes JT65B,C. Still needs work! git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6507 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-03-07 15:00:23 -05:00
fhigh=4000,nrxfreq=1500,ntrperiod=1,ndepth=1,nexp_decode=0
Fix the jt9 command line acceptance constraints git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6295 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-18 12:57:18 -05:00
logical :: read_files = .true., tx9 = .false., display_help = .false.
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
type (option) :: long_options(24) = [ &
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
option ('help', .false., 'h', 'Display this help message', ''), &
Tidy up command line processing in Fortran tools Build script now builds jt65. The tools jt65 and jt65sim usilize the options module for all comamnd line parsing. The tool jt65sim now has conventinal command line options and switchs. Default values have been added where appropriate and options documentation updated. An option to disable PRNG seeding has been added. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6276 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-15 10:35:04 -05:00
option ('shmem',.true.,'s','Use shared memory for sample data','KEY'), &
option ('tr-period', .true., 'p', 'Tx/Rx period, default MINUTES=1', &
'MINUTES'), &
option ('executable-path', .true., 'e', &
'Location of subordinate executables (KVASD) default PATH="."', &
'PATH'), &
option ('data-path', .true., 'a', &
'Location of writeable data files, default PATH="."', 'PATH'), &
option ('temp-path', .true., 't', &
'Temporary files path, default PATH="."', 'PATH'), &
option ('lowest', .true., 'L', &
'Lowest frequency decoded (JT65), default HERTZ=200', 'HERTZ'), &
option ('highest', .true., 'H', &
'Highest frequency decoded, default HERTZ=4007', 'HERTZ'), &
option ('split', .true., 'S', &
'Lowest JT9 frequency decoded, default HERTZ=2700', 'HERTZ'), &
option ('rx-frequency', .true., 'f', &
'Receive frequency offset, default HERTZ=1500', 'HERTZ'), &
option ('patience', .true., 'w', &
'FFTW3 planing patience (0-4), default PATIENCE=1', 'PATIENCE'), &
option ('fft-threads', .true., 'm', &
'Number of threads to process large FFTs, default THREADS=1', &
'THREADS'), &
option ('jt65', .false., '6', 'JT65 mode', ''), &
option ('jt9', .false., '9', 'JT9 mode', ''), &
option ('jt4', .false., '4', 'JT4 mode', ''), &
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
option ('qra64', .false., 'q', 'QRA64 mode', ''), &
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
option ('sub-mode', .true., 'b', 'Sub mode, default SUBMODE=A', 'A'), &
Tidy up command line processing in Fortran tools Build script now builds jt65. The tools jt65 and jt65sim usilize the options module for all comamnd line parsing. The tool jt65sim now has conventinal command line options and switchs. Default values have been added where appropriate and options documentation updated. An option to disable PRNG seeding has been added. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6276 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-15 10:35:04 -05:00
option ('depth', .true., 'd', &
'JT9 decoding depth (1-3), default DEPTH=1', 'DEPTH'), &
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
option ('tx-jt9', .false., 'T', 'Tx mode is JT9', ''), &
option ('my-call', .true., 'c', 'my callsign', 'CALL'), &
option ('my-grid', .true., 'G', 'my grid locator', 'GRID'), &
option ('his-call', .true., 'x', 'his callsign', 'CALL'), &
option ('his-grid', .true., 'g', 'his grid locator', 'GRID'), &
option ('experience-decode', .true., 'X', &
'experience based decoding flags (1..n), default FLAGS=0', &
'FLAGS') ]
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
type(dec_data), allocatable :: shared_data
character(len=12) :: mycall, hiscall
character(len=6) :: mygrid, hisgrid
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
common/patience/npatience,nthreads
Reintegrate merge from wsjtx_exp branch. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6122 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-11-17 20:28:12 -05:00
common/decstats/ntry65a,ntry65b,n65a,n65b,num9,numfano
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
data npatience/1/,nthreads/1/
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
nsubmode = 0
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
do
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
call getopt('hs:e:a:b:r:m:p:d:f:w:t:964qTL:S:H:c:G:x:g:X:',long_options,c, &
Tidy up command line processing in Fortran tools Build script now builds jt65. The tools jt65 and jt65sim usilize the options module for all comamnd line parsing. The tool jt65sim now has conventinal command line options and switchs. Default values have been added where appropriate and options documentation updated. An option to disable PRNG seeding has been added. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6276 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-15 10:35:04 -05:00
optarg,arglen,stat,offset,remain,.true.)
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
if (stat .ne. 0) then
exit
end if
select case (c)
Add more decoding parameters to jt9 command line program Options have been added to control frequency cut off values, mode and, Tx mode. The comand line parser now has optional long option names and usage help. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4949 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-10 09:04:54 -05:00
case ('h')
display_help = .true.
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
case ('s')
Fix the jt9 command line acceptance constraints git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6295 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-18 12:57:18 -05:00
read_files = .false.
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
shm_key = optarg(:arglen)
case ('e')
exe_dir = optarg(:arglen)
case ('a')
data_dir = optarg(:arglen)
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
case ('b')
nsubmode = ichar (optarg(:1)) - ichar ('A')
Take advantage of the new KVASD v1.12 Pass the temporary directory to jt9 and use it to give the correct paths to temporary files. Also jt9 passes the absolute path to kvasd.dat in the temporary directory to kvasd. Clear out all the annoying cruft that has accumulated due to having to run with $CWD as the temporary directory. Use QStandardPaths to find the writable data directory where needed rather than passing it around between objects. This now works because the $CWD hasn't been changed. Do away with the CMake option WSJT_STANDARD_FILE_LOCATIONS as it is no longer needed. Fix astro status file azel.dat formatting. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4732 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-12-02 19:06:54 -05:00
case ('t')
temp_dir = optarg(:arglen)
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
case ('m')
read (optarg(:arglen), *) nthreads
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
case ('p')
Fix defect in jt9 standalone mode causing randon file read Made default values for decoding parameters better in standalone jt65 tool. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6291 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 18:21:38 -05:00
read (optarg(:arglen), *) ntrperiod
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
case ('d')
read (optarg(:arglen), *) ndepth
case ('f')
read (optarg(:arglen), *) nrxfreq
Add more decoding parameters to jt9 command line program Options have been added to control frequency cut off values, mode and, Tx mode. The comand line parser now has optional long option names and usage help. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4949 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-10 09:04:54 -05:00
case ('L')
read (optarg(:arglen), *) flow
case ('S')
read (optarg(:arglen), *) fsplit
case ('H')
read (optarg(:arglen), *) fhigh
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
case ('q')
mode = 164
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
case ('4')
mode = 4
Add more decoding parameters to jt9 command line program Options have been added to control frequency cut off values, mode and, Tx mode. The comand line parser now has optional long option names and usage help. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4949 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-10 09:04:54 -05:00
case ('6')
if (mode.lt.65) mode = mode + 65
case ('9')
if (mode.lt.9.or.mode.eq.65) mode = mode + 9
case ('T')
tx9 = .true.
FFTW wisdom is now built into jt9[.exe]. New optional argument to jt9: -w patience Default is patience = 1 Example timing measurements for 130610_2343.wav: patience plan execute (s) (s) ----------------------------------------------- 0 0.01 1.25 FFTW_ESTIMATE 1 0.69 1.25 FFTW_ESTIMATE_PATIENT 2 16.97 1.15 FFTW_MEASURE 3 390.88 1.15 FFTW_PATIENT Conclusions, consistent with expectation based on past experience with similar FFTs: - First decode (in each mode) with patience = 2 is slow. - Speed advantage of patience = 2 is small but measurable. - No measurable advantage in using patience > 2. Present mainwindow.cpp has "-w 1" hard-wired. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4610 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-11-20 13:48:53 -05:00
case ('w')
read (optarg(:arglen), *) npatience
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
case ('c')
read (optarg(:arglen), *) mycall
case ('G')
read (optarg(:arglen), *) mygrid
case ('x')
read (optarg(:arglen), *) hiscall
case ('g')
read (optarg(:arglen), *) hisgrid
case ('X')
read (optarg(:arglen), *) nexp_decode
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
end select
end do
Tidy up command line processing in Fortran tools Build script now builds jt65. The tools jt65 and jt65sim usilize the options module for all comamnd line parsing. The tool jt65sim now has conventinal command line options and switchs. Default values have been added where appropriate and options documentation updated. An option to disable PRNG seeding has been added. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6276 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-15 10:35:04 -05:00
if (display_help .or. stat .lt. 0 &
Fix the jt9 command line acceptance constraints git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6295 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-18 12:57:18 -05:00
.or. (.not. read_files .and. remain .gt. 0) &
.or. (read_files .and. remain .lt. 1)) then
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
Tidy up command line processing in Fortran tools Build script now builds jt65. The tools jt65 and jt65sim usilize the options module for all comamnd line parsing. The tool jt65sim now has conventinal command line options and switchs. Default values have been added where appropriate and options documentation updated. An option to disable PRNG seeding has been added. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6276 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-15 10:35:04 -05:00
print *, 'Usage: jt9 [OPTIONS] file1 [file2 ...]'
print *, ' Reads data from *.wav files.'
print *, ''
print *, ' jt9 -s <key> [-w patience] [-m threads] [-e path] [-a path] [-t path]'
print *, ' Gets data from shared memory region with key==<key>'
print *, ''
print *, 'OPTIONS:'
print *, ''
Add more decoding parameters to jt9 command line program Options have been added to control frequency cut off values, mode and, Tx mode. The comand line parser now has optional long option names and usage help. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4949 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-10 09:04:54 -05:00
do i = 1, size (long_options)
call long_options(i) % print (6)
end do
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
go to 999
endif
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
iret=fftwf_init_threads() !Initialize FFTW threading
! Default to 1 thread, but use nthreads for the big ones
call fftwf_plan_with_nthreads(1)
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
! Import FFTW wisdom, if available
wisfile=trim(data_dir)//'/jt9_wisdom.dat'// C_NULL_CHAR
iret=fftwf_import_wisdom_from_filename(wisfile)
FFTW wisdom is now built into jt9[.exe]. New optional argument to jt9: -w patience Default is patience = 1 Example timing measurements for 130610_2343.wav: patience plan execute (s) (s) ----------------------------------------------- 0 0.01 1.25 FFTW_ESTIMATE 1 0.69 1.25 FFTW_ESTIMATE_PATIENT 2 16.97 1.15 FFTW_MEASURE 3 390.88 1.15 FFTW_PATIENT Conclusions, consistent with expectation based on past experience with similar FFTs: - First decode (in each mode) with patience = 2 is slow. - Speed advantage of patience = 2 is small but measurable. - No measurable advantage in using patience > 2. Present mainwindow.cpp has "-w 1" hard-wired. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4610 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-11-20 13:48:53 -05:00
Reintegrate merge from wsjtx_exp branch. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6122 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-11-17 20:28:12 -05:00
ntry65a=0
ntry65b=0
n65a=0
n65b=0
Add decoder statistics to timer.out. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4958 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-13 09:22:54 -05:00
num9=0
numfano=0
Fix the jt9 command line acceptance constraints git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6295 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-18 12:57:18 -05:00
if (.not. read_files) then
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
call jt9a() !We're running under control of WSJT-X
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
go to 999
endif
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
allocate(shared_data)
Better to set nflatten=0 here, for command-line speed tests. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4935 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-04 16:42:49 -05:00
nflatten=0
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
do iarg = offset + 1, offset + remain
call get_command_argument (iarg, optarg, arglen)
infile = optarg(:arglen)
Make WAV file reading correct so that optional chunks can be tolerated The WAV file format allows for optional header content, allow for this in preparation for adding some metadata to WSJT-X recorded WAV files. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6335 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-01-02 09:29:02 -05:00
call wav%read (infile)
nfsample=wav%audio_format%sample_rate
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
i1=index(infile,'.wav')
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
if(i1.lt.1) i1=index(infile,'.WAV')
if(infile(i1-5:i1-5).eq.'_') then
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
read(infile(i1-4:i1-1),*,err=1) nutc
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
else
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
read(infile(i1-6:i1-3),*,err=1) nutc
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
endif
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
go to 2
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
1 nutc=0
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
2 nsps=0
Fix defect in jt9 standalone mode causing randon file read Made default values for decoding parameters better in standalone jt65 tool. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6291 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 18:21:38 -05:00
if(ntrperiod.eq.1) then
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
nsps=6912
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%nzhsym=181
Fix defect in jt9 standalone mode causing randon file read Made default values for decoding parameters better in standalone jt65 tool. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6291 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 18:21:38 -05:00
else if(ntrperiod.eq.2) then
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
nsps=15360
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%nzhsym=178
Fix defect in jt9 standalone mode causing randon file read Made default values for decoding parameters better in standalone jt65 tool. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6291 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 18:21:38 -05:00
else if(ntrperiod.eq.5) then
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
nsps=40960
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%nzhsym=172
Fix defect in jt9 standalone mode causing randon file read Made default values for decoding parameters better in standalone jt65 tool. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6291 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 18:21:38 -05:00
else if(ntrperiod.eq.10) then
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
nsps=82944
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%nzhsym=171
Fix defect in jt9 standalone mode causing randon file read Made default values for decoding parameters better in standalone jt65 tool. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6291 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 18:21:38 -05:00
else if(ntrperiod.eq.30) then
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
nsps=252000
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%nzhsym=167
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
endif
if(nsps.eq.0) stop 'Error: bad TRperiod'
kstep=nsps/2
k=0
nhsym0=-999
npts=(60*ntrperiod-6)*12000
if(iarg .eq. offset + 1) then
Make Fortran profiling timer function a callback with a default null implementation Groundwork for calling the decoders directly from C/C++ threads. To access the timer module timer_module must now be used. Instrumented code need only use the module function 'timer' which is now a procedure pointer that is guaranteed to be associated (unless null() is assigned to it, which should not be done). The default behaviour of 'timer' is to do nothing. If a Fortran program wishes to profile code it should now use the timer_impl module which contains a default timer implementation. The main program should call 'init_timer([filename])' before using 'timer' or calling routines that are instrumented. If 'init_timer([filename])'. If it is called then an optional file name may be provided with 'timer.out' being used as a default. The procedure 'fini_timer()' may be called to close the file. The default timer implementation is thread safe if used with OpenMP multi-threaded code so long as the OpenMP thread team is given the copyin(/timer_private/) attribute for correct operation. The common block /timer_private/ should be included for OpenMP use by including the file 'timer_common.inc'. The module 'lib/timer_C_wrapper.f90' provides a Fortran wrapper along with 'init' and 'fini' subroutines which allow a C/C++ application to call timer instrumented Fortran code and for it to receive callbacks of 'timer()' subroutine invocations. No C/C++ timer implementation is provided at this stage. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6320 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-27 10:40:57 -05:00
call init_timer (trim(data_dir)//'/timer.out')
FFTW wisdom is now built into jt9[.exe]. New optional argument to jt9: -w patience Default is patience = 1 Example timing measurements for 130610_2343.wav: patience plan execute (s) (s) ----------------------------------------------- 0 0.01 1.25 FFTW_ESTIMATE 1 0.69 1.25 FFTW_ESTIMATE_PATIENT 2 16.97 1.15 FFTW_MEASURE 3 390.88 1.15 FFTW_PATIENT Conclusions, consistent with expectation based on past experience with similar FFTs: - First decode (in each mode) with patience = 2 is slow. - Speed advantage of patience = 2 is small but measurable. - No measurable advantage in using patience > 2. Present mainwindow.cpp has "-w 1" hard-wired. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4610 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-11-20 13:48:53 -05:00
call timer('jt9 ',0)
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
endif
Fix defect in jt9 standalone mode causing randon file read Made default values for decoding parameters better in standalone jt65 tool. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6291 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 18:21:38 -05:00
shared_data%id2=0 !??? Why is this necessary ???
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
do iblk=1,npts/kstep
k=iblk*kstep
call timer('read_wav',0)
Make WAV file reading correct so that optional chunks can be tolerated The WAV file format allows for optional header content, allow for this in preparation for adding some metadata to WSJT-X recorded WAV files. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6335 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-01-02 09:29:02 -05:00
read(unit=wav%lun,end=3) shared_data%id2(k-kstep+1:k)
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
go to 4
3 call timer('read_wav',1)
print*,'EOF on input file ',infile
exit
4 call timer('read_wav',1)
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
nhsym=(k-2048)/kstep
if(nhsym.ge.1 .and. nhsym.ne.nhsym0) then
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
if(mode.eq.9 .or. mode.eq.74) then
! Compute rough symbol spectra for the JT9 decoder
ingain=0
call timer('symspec ',0)
Reintegrate the wsjtx_exp branch into the trunk This merge brings the WSPR feature development into the main line ready for release in a future v1.6 release. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5424 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-05-27 09:08:28 -04:00
nminw=1
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
call symspec(shared_data,k,ntrperiod,nsps,ingain,nminw,pxdb,s,df3, &
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
ihsym,npts8)
call timer('symspec ',1)
endif
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
nhsym0=nhsym
Reintegrate Joe's experimental VHF & up features As at ^/branches/wsjtx_exp@5271 git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5272 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-04-22 13:48:03 -04:00
if(nhsym.ge.181) exit
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
endif
enddo
Make WAV file reading correct so that optional chunks can be tolerated The WAV file format allows for optional header content, allow for this in preparation for adding some metadata to WSJT-X recorded WAV files. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6335 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-01-02 09:29:02 -05:00
close(unit=wav%lun)
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%nutc=nutc
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
shared_data%params%ndiskdat=.true.
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%ntr=60
shared_data%params%nfqso=nrxfreq
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
shared_data%params%newdat=.true.
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%npts8=74736
shared_data%params%nfa=flow
shared_data%params%nfsplit=fsplit
shared_data%params%nfb=fhigh
shared_data%params%ntol=20
shared_data%params%kin=64800
shared_data%params%nzhsym=181
shared_data%params%ndepth=ndepth
shared_data%params%dttol=3.
Merge improvements to the FT algorithm into Bill's new structure for the Fortran routines. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6328 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-30 20:30:31 -05:00
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
shared_data%params%minsync=0 !### TEST ONLY
Display sine-wave frequency for QRA64 "@nnnn" messages. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7326 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-21 15:30:11 -05:00
! shared_data%params%nfqso=1500 !### TEST ONLY
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
mycall="G3WDG " !### TEST ONLY
hiscall="VK7MO " !### TEST ONLY
Display sine-wave frequency for QRA64 "@nnnn" messages. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7326 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-21 15:30:11 -05:00
hisgrid="QE37 " !### TEST ONLY
Update jt9 to include decoding of QRA64. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@7322 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-11-16 13:25:03 -05:00
if(mode.eq.164 .and. nsubmode.lt.100) nsubmode=nsubmode+100
Merge improvements to the FT algorithm into Bill's new structure for the Fortran routines. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6328 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-30 20:30:31 -05:00
Tweak some details used for collecting decode statistics. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6300 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-21 09:48:51 -05:00
shared_data%params%naggressive=10
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%n2pass=1
First (very basic!) functionality for message averaging in JT65. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6511 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-03-08 16:04:05 -05:00
! shared_data%params%nranera=8 !### ntrials=10000
shared_data%params%nranera=7 !### ntrials=3000
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
shared_data%params%nrobust=.false.
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%nexp_decode=nexp_decode
shared_data%params%mycall=mycall
shared_data%params%mygrid=mygrid
shared_data%params%hiscall=hiscall
shared_data%params%hisgrid=hisgrid
if (shared_data%params%mycall == '') shared_data%params%mycall='K1ABC'
if (shared_data%params%hiscall == '') shared_data%params%hiscall='W9XYZ'
if (shared_data%params%hisgrid == '') shared_data%params%hiscall='EN37'
if (tx9) then
shared_data%params%ntxmode=9
else
shared_data%params%ntxmode=65
end if
if (mode.eq.0) then
shared_data%params%nmode=65+9
else
shared_data%params%nmode=mode
end if
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
shared_data%params%nsubmode=nsubmode
Get rid of jt9com and npar common blocks Also use correct C binding and have compilers determine sizes and offsets. The wsjtx.exe program now owns the decoder shared data that is shared with symspec. It is now in struct dec_data, still a global variable for now but hopefully a MainWindow member variable soon. The struct dec_data (in both C/C++ and Fortran) has a sub structure with the decoder parameters which enables copying and manipulating a lot cleaner. New of changed types of shared data must still be modified in commons.h and a new file lib/jt9com.f90, they must stay in sync as a pointer to the structure is passed between C and Fortran. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6290 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-17 15:29:55 -05:00
shared_data%params%datetime="2013-Apr-16 15:13" !### Temp
if(mode.eq.9 .and. fsplit.ne.2700) shared_data%params%nfa=fsplit
Restructuring in preparation for direct decoder invocation from wsjtx Re-factor the JT4, JT65 and JT9 decoders as Fortran modules using type bound procedures, the decoder types implement a callback procedure such that he client of the decoder can interpret the decode results as they need. The JT4 decoder has a second callback that delivers message averaging status. Also the previously separate source files lib/jt4a.f90 and lib/avg4.f90 have been merged into lib/jt4_decode.f90 as private type bound procedures of the new jt4_decoder type. Re-factored the lib/decoder.f90 subroutine to utilize the new decoder types. Added local procedures to process decodes and averaging results including the necessary OpenMP synchronization directives for parallel JT9+JT65 decoding. Added the jt65_test module which is a basic test harness for JT65 decoding. Re-factored the jt65 utility to utilize the new jt65_test module. Changed a few integers to logical variables where their meaning is clearly binary. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6324 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-29 18:52:55 -05:00
call multimode_decoder(shared_data%ss,shared_data%id2,shared_data%params,nfsample)
Work around what appears to be a gcc 4.9.0 defect. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4036 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-04-16 12:43:52 -04:00
enddo
call timer('jt9 ',1)
call timer('jt9 ',101)
FFTW wisdom is now built into jt9[.exe]. New optional argument to jt9: -w patience Default is patience = 1 Example timing measurements for 130610_2343.wav: patience plan execute (s) (s) ----------------------------------------------- 0 0.01 1.25 FFTW_ESTIMATE 1 0.69 1.25 FFTW_ESTIMATE_PATIENT 2 16.97 1.15 FFTW_MEASURE 3 390.88 1.15 FFTW_PATIENT Conclusions, consistent with expectation based on past experience with similar FFTs: - First decode (in each mode) with patience = 2 is slow. - Speed advantage of patience = 2 is small but measurable. - No measurable advantage in using patience > 2. Present mainwindow.cpp has "-w 1" hard-wired. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4610 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-11-20 13:48:53 -05:00
999 continue
Make Fortran profiling timer function a callback with a default null implementation Groundwork for calling the decoders directly from C/C++ threads. To access the timer module timer_module must now be used. Instrumented code need only use the module function 'timer' which is now a procedure pointer that is guaranteed to be associated (unless null() is assigned to it, which should not be done). The default behaviour of 'timer' is to do nothing. If a Fortran program wishes to profile code it should now use the timer_impl module which contains a default timer implementation. The main program should call 'init_timer([filename])' before using 'timer' or calling routines that are instrumented. If 'init_timer([filename])'. If it is called then an optional file name may be provided with 'timer.out' being used as a default. The procedure 'fini_timer()' may be called to close the file. The default timer implementation is thread safe if used with OpenMP multi-threaded code so long as the OpenMP thread team is given the copyin(/timer_private/) attribute for correct operation. The common block /timer_private/ should be included for OpenMP use by including the file 'timer_common.inc'. The module 'lib/timer_C_wrapper.f90' provides a Fortran wrapper along with 'init' and 'fini' subroutines which allow a C/C++ application to call timer instrumented Fortran code and for it to receive callbacks of 'timer()' subroutine invocations. No C/C++ timer implementation is provided at this stage. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6320 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-12-27 10:40:57 -05:00
! Output decoder statistics
call fini_timer ()
open (unit=12, file=trim(data_dir)//'/timer.out', status='unknown', position='append')
Reintegrate merge from wsjtx_exp branch. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6122 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-11-17 20:28:12 -05:00
write(12,1100) n65a,ntry65a,n65b,ntry65b,numfano,num9
1100 format(58('-')/' JT65_1 Tries_1 JT65_2 Tries_2 JT9 Tries'/ &
Reformat last few lines of timer.out. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@5006 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-03-05 10:27:13 -05:00
58('-')/6i8)
Add decoder statistics to timer.out. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4958 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-13 09:22:54 -05:00
! Save wisdom and free memory
Made a start at implementing an option to use multi-threaded FFTs. New command-line option for jt9: [-m nthreads]. Default is nthreads=1. Also refactored a loop in filbig.f90 that was taking far too much time. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4916 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-01-30 16:28:10 -05:00
iret=fftwf_export_wisdom_to_filename(wisfile)
Thread safety enhancements to jt9 Serialize thread unsafe FFTW3 calls Serialize FFTW3 plan initializations in four2a Serialize decoder results output to file/console git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4928 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-03 20:41:26 -05:00
call four2a(a,-1,1,1,1)
Many changes in program jt9[.exe] aimed at speeding up the decoders. The long FFTs can now use the multi-threaded FFTW routines. Subroutine decode9.f90 was renamed jt9fano.f90. The JT9 decoder's top-level functions were removed from decoder.f90 and put into a separate subroutine decjt90.f90. Subroutine decoder.f90 is now configured for possible use of OpenMP SECTIONS, with the JT9 and JT65 decoders running concurrently on a multi-core machine. Note, however, that this concurrent processing is not yet fully implemented. Probably calls to timer need to be removed; some variables used in calls to jt65a and decjt9 may need to be declared PRIVATE in decoder; some sections probably need to be declared CRITICAL; probably some SAVE statements in downstream routines have made them not thread-safe; etc., etc. I'm a neophyte at using OpenMP. Comments, suggestions, and/or tests by others will be welcome! git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4919 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2015-02-01 11:23:36 -05:00
call filbig(a,-1,1,0.0,0,0,0,0,0) !used for FFT plans
call fftwf_cleanup_threads()
call fftwf_cleanup()
FFTW wisdom is now built into jt9[.exe]. New optional argument to jt9: -w patience Default is patience = 1 Example timing measurements for 130610_2343.wav: patience plan execute (s) (s) ----------------------------------------------- 0 0.01 1.25 FFTW_ESTIMATE 1 0.69 1.25 FFTW_ESTIMATE_PATIENT 2 16.97 1.15 FFTW_MEASURE 3 390.88 1.15 FFTW_PATIENT Conclusions, consistent with expectation based on past experience with similar FFTs: - First decode (in each mode) with patience = 2 is slow. - Speed advantage of patience = 2 is small but measurable. - No measurable advantage in using patience > 2. Present mainwindow.cpp has "-w 1" hard-wired. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4610 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2014-11-20 13:48:53 -05:00
end program jt9
Reference in New Issue Copy Permalink