#!/bin/sh # This is a shell archive, meaning: # 1. Remove everything above the #!/bin/sh line. # 2. Save the resulting text in a file. # 3. Execute the file with /bin/sh (not csh) to create the files: # README # cputm.c # d_and_c.f # data # example.f # ftsubs.f # ftsubs.graph.f # maindp.f # make1 # makefile # newevdp0.f # oldtest # putq.c # second.f # stateig.f # statses.f # stseswait.f # stuffspawn.f # testrun # This archive created: Thu Apr 28 17:28:13 1988 export PATH; PATH=/bin:$PATH if test -f 'README' then echo shar: over-writing existing file "'README'" fi cat << \SHAR_EOF > 'README' To build the library needed to use schedule type: "make sched" or "make graph" make sched will produce the schedule library in file sched.a . This is the standard libarary for schedule. All examples begining with 'x' use this library. make graph will produce the trace version of schedule in file graph.a . This version produces output which can be viewed using the Schedule Trace Facility. All examples begining with 'g' use this library. To run the examples type: "make xtest; xtest" or "make gtest; gtest" This will make and run a test program which solves a triangular system of equations using the spawning capabilities of schedule. "make xdandc; xdandc" or "make gdandc; gdandc" This will make and run a test program which implements a recursive divide and conquer technique using the spawning capabilities of schedule. "make xeig; xeig" or "make geig; geig" This will make and run a test program which computes the eigenvalues of a symmetric tridiagonal eigenvalue problem using a divide and conquer technique. "make xwait; xwait" or "make gwait; gwait" This will make and run a test program which solves the same problem as above except that it uses the schedule wait command. "make xexample; xexample" or "make gexample; gexample" This will make and run a test program which computes the inner product of two vectors. A complete make and test can be carried out by doing: testrun >& testout diff oldtest testout (The only differences should be the times.) SHAR_EOF if test -f 'cputm.c' then echo shar: over-writing existing file "'cputm.c'" fi cat << \SHAR_EOF > 'cputm.c' #include #include #include long cputm_() { long seconds, microseconds, milliseconds; struct rusage buffer; getrusage(RUSAGE_SELF, &buffer); seconds = buffer.ru_utime.tv_sec; microseconds = buffer.ru_utime.tv_usec; milliseconds = 1000*seconds + microseconds/1000; return(milliseconds); } SHAR_EOF if test -f 'd_and_c.f' then echo shar: over-writing existing file "'d_and_c.f'" fi cat << \SHAR_EOF > 'd_and_c.f' $STDUNIT integer a(256),klevl(8),myid(256) external top write(6,*) ' input nprocs nlevls ' read (5,*) nprocs,nlevls do 50 j = 1,8 klevl(j) = j 50 continue call sched(nprocs,top,a,nlevls,klevl,myid) do 100 j = 1,2**nlevls-1 write(6,*) a(j) 100 continue stop end subroutine top(a,nlevls,klevl,myid) integer a(*),klevl(*),myid(*) external split c write(6,*) ' from top ' , a(1) jobtag = 1 icango =0 nchks = 0 myid(1) = 1 a(1) = 1 c call dep(jobtag,icango,nchks,mychkn) call putq(jobtag,split,myid,a,nlevls,klevl) c return end subroutine split(myid,a,nlevls,klevl) integer a(*),klevl(*),myid(*),rnode external clone c write(6,*) ' from split ',a(1) c if (klevl(1) .ge. nlevls) return c lnode = 2*a(1) rnode = lnode + 1 indx = lnode - a(1) + 1 a(indx) = lnode a(indx+1) = rnode mytag = myid(a(1)) c call nxtag(mytag,jobtag) myid(lnode) = jobtag call spawn(mytag,jobtag,clone,myid,a(indx),nlevls,klevl(2)) c call nxtag(mytag,jobtag) myid(rnode) = jobtag call spawn(mytag,jobtag,clone,myid,a(indx+1),nlevls,klevl(2)) c return end SHAR_EOF if test -f 'data' then echo shar: over-writing existing file "'data'" fi cat << \SHAR_EOF > 'data' 12 4 SHAR_EOF if test -f 'example.f' then echo shar: over-writing existing file "'example.f'" fi cat << \SHAR_EOF > 'example.f' $STDUNIT program main integer n, k c external parprd c real a(1000), b(1000), temp(50), sigma read (5,*) nprocs n = 100 k = 10 c do 100 j = 1, n a(j) = j b(j) = 1 100 continue c call sched(nprocs, parprd, n, k, a, b, temp, sigma) c write (6,*) ' sigma = ', sigma stop end c c subroutine parprd(n, k, a, b, temp, sigma) c integer n, k real a(*), b(*), temp(*), sigma c integer m1, m2, index, j, jobtag, icango, ncheks, mychkn(2) c external inprod, addup save m1, m2 c m1 = n/k index = 1 do 200 j = 1, k - 1 jobtag = j icango = 0 ncheks = 1 mychkn(1) = k + 1 call dep(jobtag, icango, ncheks, mychkn) call putq(jobtag, inprod, m1, a(index), b(index), temp(j)) index = index + m1 200 continue c m2 = n - index + 1 jobtag = k icango = 0 ncheks = 1 mychkn(1) = k + 1 call dep(jobtag, icango, ncheks, mychkn) call putq(jobtag, inprod, m2, a(index), b(index), temp(k)) index = index + m1 c jobtag = k + 1 icango = k ncheks = 0 call dep(jobtag, icango, ncheks, mychkn) call putq(jobtag, addup, k, sigma, temp) c return end c c subroutine inprod(m, a, b, sigma) integer m real a(*), b(*), sigma sigma = 0.0 do 100 j = 1, m sigma = sigma + a(j)*b(j) 100 continue return end c c subroutine addup(k, sigma, temp) integer k real sigma, temp(*) sigma = 0.0 do 100 j = 1, k sigma = sigma + temp(j) 100 continue return end SHAR_EOF if test -f 'ftsubs.f' then echo shar: over-writing existing file "'ftsubs.f'" fi cat << \SHAR_EOF > 'ftsubs.f' $STDUNIT $ALIGNWARN subroutine chekin(jobtag) integer jobtag c*********************************************************************** c c this subroutine reports unit of computation labeled by c jobtag has completed to all dependent nodes. these nodes are c recorded in parmq(i,jobtag) where 6 .le. i .le. nchks+5 c checkin consists of decrementing the value in each of these c locations by 1. each of these is done in a critical section c protected by qlock(ichek) c c if the value in parmq(2,ichek) is 0 where ichek is a process c dependent upon this one then ichek is placed on the readyq c by entering the critical section protected by rtlock. the c pointer rtail to the tail of the readyq is incremented c unless task done is to be recorded. task done is placed on c the ready q and the pointer rtail left in place if nchks .eq. 0 c is found. c c see the common block description in libopn for more detail. c c*********************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c c c common block description: c c a complete common block description is given in the routine libopn c c**************************************************************************** c c check to see if this process has completed. if not do not checkin c c first ask if any kids spawned by jobtag c if (parmq(4,jobtag) .ne. 0 .or. parmq(5,jobtag) .ne. 0 ) then c c either kids have been spawned or ientry has been referenced c indicating reentry is required c find out how many are waiting to complete c if (parmq(4,jobtag) .ne. 0) then call s_lock(qlock(jobtag)) parmq(2,jobtag) = parmq(2,jobtag) - parmq(4,jobtag) call s_unlock(qlock(jobtag)) endif c c reset number of kids c parmq(4,jobtag) = 0 c c update the number of times this procedure has been c entered c parmq(1,jobtag) = parmq(1,jobtag) + 1 c c return without checkin if all the kids have not c checked in to jobtag yet or if a reentry is required. c process jobtag is not done in either case. c if (parmq(2,jobtag) .ne. 0) return c c if ientry has been called but jobtag is not waiting c on any kids then jobtag is placed back on the readyq c if (parmq(5,jobtag) .ne. 0) then iwrkr = mod(jobtag,mxces) + 1 call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = jobtag rtail(iwrkr) = rtail(iwrkr) + 1 call s_unlock(trlock(iwrkr)) return endif endif c c the process has completed so chekin proceeds c nchks = parmq(3,jobtag) c c if this is the final process (indicated by nchks .eq. 0) then c record task done. do not advance the tail so task done sequence c is set. all subsequent gtprb queries will leave rhead .eq. rtail c with readyq(rhead) .eq. done. c if (nchks .eq. 0) then do 20 iwrkr = 1,mxces call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = done call s_unlock(trlock(iwrkr)) 20 continue return endif do 50 j = 6,nchks+5 mychek = parmq(j,jobtag) c c get unique access to the checkin node mychek c and checkin by decrementing the appropriate counter c mchkgo = 1 call s_lock(qlock(mychek)) parmq(2,mychek) = parmq(2,mychek) - 1 mchkgo = parmq(2,mychek) call s_unlock(qlock(mychek)) c c place mychek on readyq if parmq(2,mychek) is 0 c if (mchkgo .eq. 0 ) then iwrkr = mod(mychek,mxces) + 1 call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = mychek rtail(iwrkr) = rtail(iwrkr) + 1 call s_unlock(trlock(iwrkr)) endif 50 continue return c c last card of chekin c end subroutine dep(jobtag,icango,nchks,mychkn) integer jobtag,icango,nchks,mychkn(*) c************************************************************************* c c warning - this routine may only be used by driver in a static definition c of the data dependencies in the task. c c c usage c subroutine xxx() c external yyy c . c . c . c call dep(jobtag,icango,nchks,mychkn) c call putq(jobtag,yyy,) c . c . c . c c this subroutine puts data dependencies for problem on the queue. c no synchronization is necessary because each index of a column of c parmq is associated with a jobtag specified by the user and c associated with a unique unit of computation. the arguments of c dep specify a the data dependencies associated with the unit of c computation labeled by jobtag and are placed in a column of parmq c to the menue specified below. c c jobtag is an integer specifying a unique schedulable process c c c icango is a positive integer specifying how many processes must check in c to this process before it can be placed on the readyq. c c nchks is the number of processes that depend upon the completion of c this process. c c mychkn is an integer array specifying the jobtags of the processes c which depend upon completion of this process. c c************************************************************************* c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c c c common block description: c c for a complete common block description see the subroutine libopn c c c place process jobtag on the problem queue c no synchronization required to update qtail since c only one program work executes this code. c if ((jobtag .le. 0 .or. jobtag .gt. mxprcs) .or. & icango .lt. 0 .or. nchks .lt. 0) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' incorrect specification of dependencies ' write(6,*) ' all DEP parameters must be non-negative' write(6,*) ' input was ' write(6,*) ' jobtag ',jobtag ,'.... must be postitive ' write(6,*) ' but less than ',mxprcs write(6,*) ' icango ',icango write(6,*) ' nchks ',nchks write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif c qtail = qtail + 1 next = jobtag parmq(1,next) = 1 parmq(2,next) = icango parmq(3,next) = nchks parmq(4,next) = 0 c c check to see that exactly one node has ncheks set to 0 c if (nchks .eq. 0 .and. done .eq. 0) then done = -2 else if (nchks .eq. 0) done = 0 endif c c specify identifiers of processes which depend on this one c if there are too many abort c if (nchks .gt. nslots - 5) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' attempt to place too many dependencies ' write(6,*) ' on chekin list during call to dep ' write(6,*) ' with jobtag ',jobtag write(6,*) ' ' write(6,*) ' user tried to place ',nchks ,' dependencies ' write(6,*) ' the maximum number is ',nslots - 5 write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif c do 50 j = 1,nchks parmq(j+5,next) = mychkn(j) if (mychkn(j) .le. 0) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' incorrect specification of dependencies ' write(6,*) ' all mychkn entries must be positive' write(6,*) ' input was ' write(6,*) ' mychkn(',j,') = ',mychkn(j) write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif 50 continue c return c c last card of dep c end integer function gtprb(id,jobtag) c************************************************************************** c c this function gets unique access to the head of the readyq c pointed to by id and then claims the pointer to the next c schedulable process if there is one and returns with a nonzero c value when there is a process to schedule. if there are no entries c in the readyq indexed by id then the remaning ready ques are c polled in a round robin manner until schedulable process is found c or task done is recorded. if task done has been recorded the value c zero is returned in gtprb. if a nonzero value is returned in gtprb, c the integer jobtag will contain the identifier of the unit of c computation that is to be executed. c c input parameter c c id an integer specifying which readyq to access first c for work to do. c c output parameters c c jobtag an integer containing the next process to be executed c c gtprb the return value of this integer function is: c c 0 if task done has been posted c c nonzero if a schedulable process has been claimed. c c c*************************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c c common block description: c c for a complete common block description see the routine libopn c c nspins = 0 iwrkr = id 10 continue mhead = -1 c c gain access to head of readyq. if task done has not been recorded c then increment the head of the readyq. otherwise the head pointer c is left fixed so the next active process will receive task done. c call s_lock(hrlock(iwrkr)) if (rhead(iwrkr) .lt. rtail(iwrkr)) then mhead = rhead(iwrkr) rhead(iwrkr) = rhead(iwrkr) + 1 endif call s_unlock(hrlock(iwrkr)) c if (mhead .gt. 0) then c c there was a work unit on the readyq c jobtag = readyq(mhead,iwrkr) c if (jobtag .ne. done) then c c record the subroutine call identifier in gtprb and return c the process identifier in jobtag. c gtprb = parmq(1,jobtag) if (gtprb .gt. 1 .and. parmq(5,jobtag) .eq. 0) gtprb = -1 c else c c task done has been indicated. request a return from subroutine work c by returning the value 0 in gtprb. c gtprb = 0 c endif else c jobtag = readyq(rhead(iwrkr),iwrkr) if (jobtag .eq. done) then c c task done has been posted c gtprb = 0 c else c c there was not any work on the readyq c iwrkr = mod((iwrkr+1),mxces) if (iwrkr .eq. 0) iwrkr = mxces nspins = nspins + 1 if (mod(nspins,mxces) .eq. 0) call nops go to 10 c endif endif return c c last card of gtprb c end subroutine libopn(nproc) integer nproc c************************************************************************ c c this routine sets locks and initializes variables c and then spawns nproc generic work routines. c c nproc is a positive integer. care should be taken to c match nproc to the number of physical processors c available. c c************************************************************************ parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c integer ispace(mxces), itag(mxces) c c common block description: c c common/qdata/ c c parmq is a two dimensional integer array. each column of c this array represents a schedulable process. a process is c identified by its jobtag which corresponds to a unique c column of parmq. a column of parmq has the following c entries c c parmq(1,jobtag) = nentries c a nonzero integer. if process jobtag c is on the readyq then this integer c is equal to the one plus number of times c process jobtag has been entered. c thus when work executes this process c the integer is equal to the number c of times the process has been entered. c c parmq(2,jobtag) = icango c an integer specifying the number c of processes that must check in c before this process may scheduled c (ie. be placed on the ready queue) c c parmq(3,jobtag) = nchks c an integer specifying the number c of processes that this process c must checkin to. identifiers of c these processes are recorded below. c if nchks .eq. 0 then completion of c this process signifies completion of c task. c c parmq(4,jobtag) = the number of kids spawned by this c process. if this value is zero c then this process has not spawned c any subprocesses. c c parmq(5,jobtag) = entry_flag c has the value 1 if ientry was called c has the value 0 otherwise c c parmq(6:6+nchks,jobtag) is reserved for identifiers of the c nchks c processes that must wait for completion c of this process before they can execute. c c c phead pointer to head of parmq c c intspn pointer to first spawned process. all jobtags c with values greater than or equal to intspn will c be spawned processes. c c readyq a two dimensional integer array that holds the jobtags of those c processes that are ready to execute. the k-th column of c this array serves as a readyq for the k-th work routine. c on executing gtprb, the k-th work routine will look for work c in the k-th readyq first and then the others (round robin). c if readyq(j,id) .eq. done has been set then a return from c subroutine work(*,*,id) is indicated. c c rhead an integer array. the i-th entry of rhead is a pointer to the c head of the i-th column of readyq c c rtail an integer array. the i-th entry of rtail is a pointer to the c tail of the i-th column of readyq c c c common/sync/ c c qlock is an integer array of locks. there is one lock for each c column of parmq. the purpose of this lock is to ensure c unique access to a column of parmq during the checkin operation. c c hrlock is an integer lock. the purpose of this lock is to ensure c unique access to the pointer rhead to the head of the readyq. c c trlock is an integer lock. the purpose of this lock is to ensure c unique access to the pointer rtail to the tail of the readyq. c c done is a unique non positive integer set in libopn to indicate c task done. c if (nproc .gt. mxces-1) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' user asking for more physical processors' write(6,*) ' than are currently available on this system ' write(6,*) ' the maximum avilable is ', mxces write(6,*) ' ' stop endif c ierr = m_set_procs(nproc) c c if all goes well in the calls to dep (one and only one process c has ncheks equal to zero), done will be set to zero before the c processes begin execution. c done = -1 c c set qlocks off c initialize readyq(*) = -1 to set done sequence c initialize reentry indicator in parmq(5,*) c do 50 j = 1,mxces call s_init_lock(hrlock(j)) call s_init_lock(trlock(j)) rhead(j) = 1 rtail(j) = 1 do 20 i = 1,iprcs readyq(i,j) = -1 20 continue 50 continue c do 100 j = 1,mxprcs call s_init_lock(qlock(j)) parmq(5,j) = 0 100 continue c c set readyq locks off c call s_init_lock(qtlock) c c initialize queue pointers c phead = 1 intspn = 1 qtail = 2 next = 1 c c set lock on pointer to head of readyq so c no process may start until all process and data dependencies c have been specified by the user supplied routine driver. c do 150 j = 1,mxces call s_lock(hrlock(j)) 150 continue c c now spawn virtual processors. these generic work routines will c assume the identity of any schedulable process specified by driver. c C$DOACROSS share(ispace) do 200 j = 1,nproc call work(j,ispace(j)) 200 continue return c c last card of libopn c end subroutine nxtag(mypar,jobtag) integer mypar,jobtag c************************************************************************* c c c this subroutine puts data dependencies for problem on the queue. c no synchronization c is necessary because each index of a column of parmq is associated c with a jobtag specified by the user and associated with a unique c schedulable process. the arguments of putq specify a process and are c placed in a column of jobq according to the menue specified in the c common block description given below. c c jobtag is an integer specifying a unique schedulable process c c c icango is a positive integer specifying how many processes must check in c to this process before it can be placed on the readyq. c c nchks is the number of processes that depend upon the completion of c this process. c c mchkin is an integer array specifying the jobtags of the processes c which depend upon completion of this process. c c************************************************************************* c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c c c common block description: c c for a complete common block description see the subroutine libopn c c c c place this process on the next slot in the problem queue c call s_lock(qtlock) next = qtail qtail = qtail + 1 call s_unlock(qtlock) c if ( next .gt. mxprcs) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' user attempt to create too many processes' write(6,*) ' through dynamic spawning ' write(6,*) ' the maximum allowed is ',mxprcs write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif c jobtag = next parmq(1,next) = 1 parmq(2,next) = 0 parmq(3,next) = 1 parmq(6,next) = mypar c c update the icango counter of the parent process c by adding 2 to parmq(2,mypar)... prevents race condition. c add 1 to the number of kids spawned by parent mypar c call s_lock(qlock(mypar)) parmq(2,mypar) = parmq(2,mypar) + 2 parmq(4,mypar) = parmq(4,mypar) + 1 call s_unlock(qlock(mypar)) c c set number of kids spawned by next to zero c parmq(4,next) = 0 c return c c last card of nxtag c end subroutine start2 c c this routine allows parallel processing to start after user supplied c driver has completed by unlocking the head of the readyq c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock logical nostrt c c c for common block description see subroutine libopn. c if (done .ne. 0) then write(6,*) '*************SCHED USER ERROR********************' if (done .eq. -1 ) then write(6,*) ' no process has set nchks equal to 0 ' else write(6,*) ' more than one process has set nchks to 0 ' endif write(6,*) ' SCHEDULE will not be able to terminate job' write(6,*) ' correctly ' write(6,*) ' ' write(6,*) ' check subroutine passed to initial call to' write(6,*) ' to see that at exactly one call to DEP has ' write(6,*) ' set nchks = 0 ' write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif c nostrt = .true. do 100 iwrkr = 1,mxces if (rhead(iwrkr) .ne. rtail(iwrkr)) nostrt = .false. 100 continue if (nostrt) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' no process had an intitial icango of 0 ' write(6,*) ' SCHEDULE could not begin ' write(6,*) ' ' write(6,*) ' check subroutine passed to initial call to' write(6,*) ' to see that at least one call to DEP has ' write(6,*) ' set icango = 0 ' write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif c intspn = qtail do 200 iwrkr = 1,mxces call s_unlock(hrlock(iwrkr)) 200 continue c return c c last card of start2 c end subroutine place(jobtag) integer jobtag c************************************************************************* c c c this subroutine places a problem on the readyq c c jobtag is an integer specifying a unique schedulable process c c c icango is a positive integer specifying how many processes must check in c to this process before it can be placed on the readyq. c c nchks is the number of processes that depend upon the completion of c this process. c c mchkin is an integer array specifying the jobtags of the processes c which depend upon completion of this process. c c************************************************************************* c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c c c common block description: c c for a complete common block description see the subroutine libopn c c c c place this process on readyq if icango is 0 c when icango .eq. 0 this process does not depend on any c others. c icango = parmq(2,jobtag) if (icango .eq. 0 ) then iwrkr = mod(jobtag,mxces) + 1 call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = jobtag rtail(iwrkr) = rtail(iwrkr) + 1 call s_unlock(trlock(iwrkr)) endif c c last card of place c return end integer function ientry(jobtag,nentrs) c integer jobtag c***************************************************************************** c c this routine will allow process jobtag to continue after c spawned processes have all checked in. it should only be called if c processes have been spawned by jobtag through the use of c the subroutine spawn. c c go to (1000,2000,...,N000), ientry(jobtag,N) c 1000 continue c . c . c . c do 10 j = 1,nproc c . c . (set parameters to define spawned process) c . c call nxtag(myid,jobtag) c call spawn(myid,jobtag,subname,) c 10 continue c return c 2000 continue c . c . c . c return c N000 continue c c return c end c c this subroutine returns the number of times process jobtag c has been entered. if that number is equal to the total c number nentrs of expected reentries then parmq(5,jobtag) c is set to 0 indicating no more reentries required. c c***************************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c c report the entry point where process jobtag should resume c computation c if (nentrs .lt. 2) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' user call to IENTRY with number of ' write(6,*) ' labels in nentrs set less than 2 ' write(6,*) ' from process ',jobtag write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif c ientry = parmq(1,jobtag) if (ientry .lt. nentrs) then parmq(5,jobtag) = nentrs else parmq(5,jobtag) = 0 endif c return c c last card of ientry c end logical function wait(jobtag,ienter) c integer jobtag,ienter c***************************************************************************** c c this routine will allow process jobtag to continue after c spawned processes have all checked in. it should only be called if c processes have been spawned by jobtag through the use of c the subroutine spawn. this routine must be used in conjunction with c subroutine prtspn. the required syntax is c c go to (1000,...,L000,...,N000), ientry(mytag,N) c 1000 continue c . c . c . c do 100 j = 1,nproc c . c . (set parameters to define spawned process) c . c call nxtag(mytag,jobtag) c call spawn(mytag,jobtag,subname,) c 100 continue c label = L c if (wait(jobtag,label)) return c L000 continue c . c . c . c c if this subroutine returns a value of .true. then the calling process c jobtag should issue a return. if a value of .false. is returned then c the calling process jobtag should resume execution at the c statement immediately following the reference to wait (ie. at L000 in c the example above. a return value .true. indicates that some spawned c process has not yet completed and checked in. a return value .false. c indicates all spawned processes have checked in. c c***************************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock c c c check the icango counter to see if all spawned processes (kids) c have checked in. c icango = 1 call s_lock(qlock(jobtag)) icango = parmq(2,jobtag) - parmq(4,jobtag) call s_unlock(qlock(jobtag)) c if (icango .eq. 0) then c c all kids are done ... dont wait (ie return false) c wait = .false. c c record re_entry label where computation is to c resume after wait is complete c parmq(1,jobtag) = ienter c if (ienter .gt. parmq(5,jobtag)) then write(6,*) '*************SCHED USER ERROR*****************' write(6,*) ' executing SCHEDULE function WAIT ' write(6,*) ' return label larger than the maximum ' write(6,*) ' specified by user in call to ientry ' write(6,*) ' from process jobtag ',jobtag write(6,*) ' ' write(6,*) ' the maximum reentry number is ' write(6,*) ' ', parmq(5,jobtag) write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' stop endif c c set last re_entry indication (parmq(5,jobtag) = 0) c if this reentry point corresponds to last one c (recorded in parmq(5,jobtag) during call to ientry) c if (ienter .eq. parmq(5,jobtag)) parmq(5,jobtag) = 0 c else c c kids are not done c wait = .true. c c a checkin will be made so set the number of c entries to return label ienter - 1 to get c correct entry point after checkin c parmq(1,jobtag) = ienter - 1 c endif c return c c last card of wait c end c subroutine lckasn(ilock) integer*1 ilock call s_init_lock(ilock) return end c subroutine lockon(ilock) integer*1 ilock call s_lock(ilock) return end c subroutine lockoff(ilock) integer*1 ilock call s_unlock(ilock) return end c subroutine nops j = 1 return end SHAR_EOF if test -f 'ftsubs.graph.f' then echo shar: over-writing existing file "'ftsubs.graph.f'" fi cat << \SHAR_EOF > 'ftsubs.graph.f' $STDUNIT $ALIGNWARN subroutine chekin(jobtag) integer jobtag c*********************************************************************** c c this subroutine reports unit of computation labeled by c jobtag has completed to all dependent nodes. these nodes are c recorded in parmq(i,jobtag) where 6 .le. i .le. nchks+5 c checkin consists of decrementing the value in each of these c locations by 1. each of these is done in a critical section c protected by qlock(ichek) c c if the value in parmq(2,ichek) is 0 where ichek is a process c dependent upon this one then ichek is placed on the readyq c by entering the critical section protected by rtlock. the c pointer rtail to the tail of the readyq is incremented c unless task done is to be recorded. task done is placed on c the ready q and the pointer rtail left in place if nchks .eq. 0 c is found. c c see the common block description in libopn for more detail. c c*********************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c c common block description: c c a complete common block description is given in the routine libopn c c**************************************************************************** c c check to see if this process has completed. if not do not checkin c c first ask if any kids spawned by jobtag c if (parmq(4,jobtag) .ne. 0 .or. parmq(5,jobtag) .ne. 0 ) then c c either kids have been spawned or ientry has been referenced c indicating reentry is required c find out how many are waiting to complete c if (parmq(4,jobtag) .ne. 0) then call s_lock(qlock(jobtag)) parmq(2,jobtag) = parmq(2,jobtag) - parmq(4,jobtag) call s_unlock(qlock(jobtag)) endif c c reset number of kids c parmq(4,jobtag) = 0 c c update the number of times this procedure has been c entered c parmq(1,jobtag) = parmq(1,jobtag) + 1 c c return without checkin if all the kids have not c checked in to jobtag yet or if a reentry is required. c process jobtag is not done in either case. c if (parmq(2,jobtag) .ne. 0) return c c if ientry has been called but jobtag is not waiting c on any kids then jobtag is placed back on the readyq c if (parmq(5,jobtag) .ne. 0) then iwrkr = mod(jobtag,mxces) + 1 call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = jobtag rtail(iwrkr) = rtail(iwrkr) + 1 call s_unlock(trlock(iwrkr)) return endif endif c c the process has completed so chekin proceeds c call s_lock(qlock(mxprcs)) if (endgrf .gt. nbuffr) call dump(endgrf,igraph) insrt = endgrf endgrf = endgrf + 1 call s_unlock(qlock(mxprcs)) if (jobtag .ge. intspn) then igraph(1,insrt) = 5 igraph(2,insrt) = parmq(6,jobtag) igraph(3,insrt) = jobtag igraph(4,insrt) = second(foo) else igraph(1,insrt) = 2 igraph(2,insrt) = jobtag igraph(3,insrt) = second(foo) endif c nchks = parmq(3,jobtag) c c if this is the final process (indicated by nchks .eq. 0) then c record task done. do not advance the tail so task done sequence c is set. all subsequent gtprb queries will leave rhead .eq. rtail c with readyq(rhead) .eq. done. c if (nchks .eq. 0) then do 20 iwrkr = 1,mxces call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = done call s_unlock(trlock(iwrkr)) 20 continue return endif do 50 j = 6,nchks+5 mychek = parmq(j,jobtag) c c get unique access to the checkin node mychek c and checkin by decrementing the appropriate counter c mchkgo = 1 call s_lock(qlock(mychek)) parmq(2,mychek) = parmq(2,mychek) - 1 mchkgo = parmq(2,mychek) call s_unlock(qlock(mychek)) c c place mychek on readyq if parmq(2,mychek) is 0 c if (mchkgo .eq. 0 ) then iwrkr = mod(mychek,mxces) + 1 call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = mychek rtail(iwrkr) = rtail(iwrkr) + 1 call s_unlock(trlock(iwrkr)) endif 50 continue return c c last card of chekin c end subroutine dep(jobtag,icango,nchks,mychkn) integer jobtag,icango,nchks,mychkn(*) c************************************************************************* c c warning - this routine may only be used by driver in a static definition c of the data dependencies in the task. c c c usage c subroutine xxx() c external yyy c . c . c . c call dep(jobtag,icango,nchks,mychkn) c call putq(jobtag,yyy,) c . c . c . c c this subroutine puts data dependencies for problem on the queue. c no synchronization is necessary because each index of a column of c parmq is associated with a jobtag specified by the user and c associated with a unique unit of computation. the arguments of c dep specify a the data dependencies associated with the unit of c computation labeled by jobtag and are placed in a column of parmq c to the menue specified below. c c jobtag is an integer specifying a unique schedulable process c c c icango is a positive integer specifying how many processes must check in c to this process before it can be placed on the readyq. c c nchks is the number of processes that depend upon the completion of c this process. c c mychkn is an integer array specifying the jobtags of the processes c which depend upon completion of this process. c c************************************************************************* c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c c common block description: c c for a complete common block description see the subroutine libopn c c c place process jobtag on the problem queue c no synchronization required to update qtail since c only one program work executes this code. c if ((jobtag .le. 0 .or. jobtag .gt. mxprcs) .or. & icango .lt. 0 .or. nchks .lt. 0) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' incorrect specification of dependencies ' write(6,*) ' all DEP parameters must be non-negative' write(6,*) ' input was ' write(6,*) ' jobtag ',jobtag ,'.... must be postitive ' write(6,*) ' but less than ',mxprcs write(6,*) ' icango ',icango write(6,*) ' nchks ',nchks write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif c qtail = qtail + 1 next = jobtag parmq(1,next) = 1 parmq(2,next) = icango parmq(3,next) = nchks parmq(4,next) = 0 c c check to see that exactly one node has ncheks set to 0 c if (nchks .eq. 0 .and. done .eq. 0) then done = -2 else if (nchks .eq. 0) done = 0 endif c c specify identifiers of processes which depend on this one c if there are too many abort c if (nchks .gt. nslots - 5) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' attempt to place too many dependencies ' write(6,*) ' on chekin list during call to dep ' write(6,*) ' with jobtag ',jobtag write(6,*) ' ' write(6,*) ' user tried to place ',nchks ,' dependencies ' write(6,*) ' the maximum number is ',nslots - 5 write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif c do 50 j = 1,nchks parmq(j+5,next) = mychkn(j) if (mychkn(j) .le. 0) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' incorrect specification of dependencies ' write(6,*) ' all mychkn entries must be positive' write(6,*) ' input was ' write(6,*) ' mychkn(',j,') = ',mychkn(j) write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif 50 continue c call s_lock(qlock(mxprcs)) if (endgrf .gt. nbuffr) call dump(endgrf,igraph) insrt = endgrf endgrf = endgrf + 1 call s_unlock(qlock(mxprcs)) igraph(1,insrt) = 0 igraph(2,insrt) = jobtag igraph(3,insrt) = icango igraph(4,insrt) = nchks do 9001 jnsrt = 6,nchks + 5 igraph(jnsrt-1,insrt) = parmq(jnsrt,next) 9001 continue c return c c last card of dep c end integer function gtprb(id,jobtag) c************************************************************************** c c this function gets unique access to the head of the readyq c pointed to by id and then claims the pointer to the next c schedulable process if there is one and returns with a nonzero c value when there is a process to schedule. if there are no entries c in the readyq indexed by id then the remaning ready ques are c polled in a round robin manner until schedulable process is found c or task done is recorded. if task done has been recorded the value c zero is returned in gtprb. if a nonzero value is returned in gtprb, c the integer jobtag will contain the identifier of the unit of c computation that is to be executed. c c input parameter c c id an integer specifying which readyq to access first c for work to do. c c output parameters c c jobtag an integer containing the next process to be executed c c gtprb the return value of this integer function is: c c 0 if task done has been posted c c nonzero if a schedulable process has been claimed. c c c*************************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c common block description: c c for a complete common block description see the routine libopn c c nspins = 0 iwrkr = id 10 continue mhead = -1 c c gain access to head of readyq. if task done has not been recorded c then increment the head of the readyq. otherwise the head pointer c is left fixed so the next active process will receive task done. c call s_lock(hrlock(iwrkr)) if (rhead(iwrkr) .lt. rtail(iwrkr)) then mhead = rhead(iwrkr) rhead(iwrkr) = rhead(iwrkr) + 1 endif call s_unlock(hrlock(iwrkr)) c if (mhead .gt. 0) then c c there was a work unit on the readyq c jobtag = readyq(mhead,iwrkr) call s_lock(qlock(mxprcs)) if (endgrf .gt. nbuffr) call dump(endgrf,igraph) insrt = endgrf endgrf = endgrf + 1 call s_unlock(qlock(mxprcs)) if (jobtag .ge. intspn) then igraph(1,insrt) = 4 igraph(2,insrt) = parmq(6,jobtag) igraph(3,insrt) = jobtag igraph(4,insrt) = second(foo) else igraph(1,insrt) = 1 igraph(2,insrt) = jobtag igraph(3,insrt) = second(foo) endif c if (jobtag .ne. done) then c c record the subroutine call identifier in gtprb and return c the process identifier in jobtag. c gtprb = parmq(1,jobtag) if (gtprb .gt. 1 .and. parmq(5,jobtag) .eq. 0) gtprb = -1 c else c c task done has been indicated. request a return from subroutine work c by returning the value 0 in gtprb. c gtprb = 0 c endif else c jobtag = readyq(rhead(iwrkr),iwrkr) if (jobtag .eq. done) then c c task done has been posted c gtprb = 0 c else c c there was not any work on the readyq c iwrkr = mod((iwrkr+1),mxces) if (iwrkr .eq. 0) iwrkr = mxces nspins = nspins + 1 if (mod(nspins,mxces) .eq. 0) call nops go to 10 c endif endif return c c last card of gtprb c end subroutine libopn(nproc) integer nproc c************************************************************************ c c this routine sets locks and initializes variables c and then spawns nproc generic work routines. c c nproc is a positive integer. care should be taken to c match nproc to the number of physical processors c available. c c************************************************************************ parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c integer ispace(mxces), itag(mxces) c c common block description: c c common/qdata/ c c parmq is a two dimensional integer array. each column of c this array represents a schedulable process. a process is c identified by its jobtag which corresponds to a unique c column of parmq. a column of parmq has the following c entries c c parmq(1,jobtag) = nentries c a nonzero integer. if process jobtag c is on the readyq then this integer c is equal to the one plus number of times c process jobtag has been entered. c thus when work executes this process c the integer is equal to the number c of times the process has been entered. c c parmq(2,jobtag) = icango c an integer specifying the number c of processes that must check in c before this process may scheduled c (ie. be placed on the ready queue) c c parmq(3,jobtag) = nchks c an integer specifying the number c of processes that this process c must checkin to. identifiers of c these processes are recorded below. c if nchks .eq. 0 then completion of c this process signifies completion of c task. c c parmq(4,jobtag) = the number of kids spawned by this c process. if this value is zero c then this process has not spawned c any subprocesses. c c parmq(5,jobtag) = entry_flag c has the value 1 if ientry was called c has the value 0 otherwise c c parmq(6:6+nchks,jobtag) is reserved for identifiers of the c nchks c processes that must wait for completion c of this process before they can execute. c c c phead pointer to head of parmq c c intspn pointer to first spawned process. all jobtags c with values greater than or equal to intspn will c be spawned processes. c c readyq a two dimensional integer array that holds the jobtags of those c processes that are ready to execute. the k-th column of c this array serves as a readyq for the k-th work routine. c on executing gtprb, the k-th work routine will look for work c in the k-th readyq first and then the others (round robin). c if readyq(j,id) .eq. done has been set then a return from c subroutine work(*,*,id) is indicated. c c rhead an integer array. the i-th entry of rhead is a pointer to the c head of the i-th column of readyq c c rtail an integer array. the i-th entry of rtail is a pointer to the c tail of the i-th column of readyq c c c common/sync/ c c qlock is an integer array of locks. there is one lock for each c column of parmq. the purpose of this lock is to ensure c unique access to a column of parmq during the checkin operation. c c hrlock is an integer lock. the purpose of this lock is to ensure c unique access to the pointer rhead to the head of the readyq. c c trlock is an integer lock. the purpose of this lock is to ensure c unique access to the pointer rtail to the tail of the readyq. c c done is a unique non positive integer set in libopn to indicate c task done. c c common /gphout/ c c endgrf is an integer pointing to the next available c slot in igraph c c igraph is a two dimensional integer array c used as a buffer for graphics output c each column of igraph records an event. c if (nproc .gt. mxces-1) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' user asking for more physical processors' write(6,*) ' than are currently available on this system ' write(6,*) ' the maximum avilable is ', mxces write(6,*) ' ' call dump(endgrf,igraph) stop endif c ierr = m_set_procs(nproc) c c if all goes well in the calls to dep (one and only one process c has ncheks equal to zero), done will be set to zero before the c processes begin execution. c done = -1 c c set qlocks off c initialize readyq(*) = -1 to set done sequence c initialize reentry indicator in parmq(5,*) c do 50 j = 1,mxces call s_init_lock(hrlock(j)) call s_init_lock(trlock(j)) rhead(j) = 1 rtail(j) = 1 do 20 i = 1,iprcs readyq(i,j) = -1 20 continue 50 continue c do 100 j = 1,mxprcs call s_init_lock(qlock(j)) parmq(5,j) = 0 100 continue c c set readyq locks off c call s_init_lock(qtlock) c c initialize queue pointers c phead = 1 intspn = 1 qtail = 2 next = 1 endgrf = 1 open( file='trace.graph',unit=3) c c set lock on pointer to head of readyq so c no process may start until all process and data dependencies c have been specified by the user supplied routine driver. c do 150 j = 1,mxces call s_lock(hrlock(j)) 150 continue c c now spawn virtual processors. these generic work routines will c assume the identity of any schedulable process specified by driver. c C$DOACROSS share(ispace) do 200 j = 1,nproc call work(j,ispace(j)) 200 continue call dump(endgrf,igraph) return c c last card of libopn c end subroutine nxtag(mypar,jobtag) integer mypar,jobtag c************************************************************************* c c c this subroutine puts data dependencies for problem on the queue. c no synchronization c is necessary because each index of a column of parmq is associated c with a jobtag specified by the user and associated with a unique c schedulable process. the arguments of putq specify a process and are c placed in a column of jobq according to the menue specified in the c common block description given below. c c jobtag is an integer specifying a unique schedulable process c c c icango is a positive integer specifying how many processes must check in c to this process before it can be placed on the readyq. c c nchks is the number of processes that depend upon the completion of c this process. c c mchkin is an integer array specifying the jobtags of the processes c which depend upon completion of this process. c c************************************************************************* c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c c common block description: c c for a complete common block description see the subroutine libopn c c c c place this process on the next slot in the problem queue c call s_lock(qtlock) next = qtail qtail = qtail + 1 call s_unlock(qtlock) c call s_lock(qlock(mxprcs)) if (endgrf .gt. nbuffr) call dump(endgrf,igraph) insrt = endgrf endgrf = endgrf + 1 call s_unlock(qlock(mxprcs)) igraph(1,insrt) = 3 igraph(2,insrt) = mypar igraph(3,insrt) = next c if ( next .gt. mxprcs) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' user attempt to create too many processes' write(6,*) ' through dynamic spawning ' write(6,*) ' the maximum allowed is ',mxprcs write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif c jobtag = next parmq(1,next) = 1 parmq(2,next) = 0 parmq(3,next) = 1 parmq(6,next) = mypar c c update the icango counter of the parent process c by adding 2 to parmq(2,mypar)... prevents race condition. c add 1 to the number of kids spawned by parent mypar c call s_lock(qlock(mypar)) parmq(2,mypar) = parmq(2,mypar) + 2 parmq(4,mypar) = parmq(4,mypar) + 1 call s_unlock(qlock(mypar)) c c set number of kids spawned by next to zero c parmq(4,next) = 0 c return c c last card of nxtag c end subroutine start2 c c this routine allows parallel processing to start after user supplied c driver has completed by unlocking the head of the readyq c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock logical nostrt integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c c for common block description see subroutine libopn. c if (done .ne. 0) then write(6,*) '*************SCHED USER ERROR********************' if (done .eq. -1 ) then write(6,*) ' no process has set nchks equal to 0 ' else write(6,*) ' more than one process has set nchks to 0 ' endif write(6,*) ' SCHEDULE will not be able to terminate job' write(6,*) ' correctly ' write(6,*) ' ' write(6,*) ' check subroutine passed to initial call to' write(6,*) ' to see that at exactly one call to DEP has ' write(6,*) ' set nchks = 0 ' write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif c nostrt = .true. do 100 iwrkr = 1,mxces if (rhead(iwrkr) .ne. rtail(iwrkr)) nostrt = .false. 100 continue if (nostrt) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' no process had an intitial icango of 0 ' write(6,*) ' SCHEDULE could not begin ' write(6,*) ' ' write(6,*) ' check subroutine passed to initial call to' write(6,*) ' to see that at least one call to DEP has ' write(6,*) ' set icango = 0 ' write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif c intspn = qtail do 200 iwrkr = 1,mxces call s_unlock(hrlock(iwrkr)) 200 continue c return c c last card of start2 c end subroutine place(jobtag) integer jobtag c************************************************************************* c c c this subroutine places a problem on the readyq c c jobtag is an integer specifying a unique schedulable process c c c icango is a positive integer specifying how many processes must check in c to this process before it can be placed on the readyq. c c nchks is the number of processes that depend upon the completion of c this process. c c mchkin is an integer array specifying the jobtags of the processes c which depend upon completion of this process. c c************************************************************************* c parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c c common block description: c c for a complete common block description see the subroutine libopn c c c c place this process on readyq if icango is 0 c when icango .eq. 0 this process does not depend on any c others. c icango = parmq(2,jobtag) if (icango .eq. 0 ) then iwrkr = mod(jobtag,mxces) + 1 call s_lock(trlock(iwrkr)) readyq(rtail(iwrkr),iwrkr) = jobtag rtail(iwrkr) = rtail(iwrkr) + 1 call s_unlock(trlock(iwrkr)) endif c c last card of place c return end integer function ientry(jobtag,nentrs) c integer jobtag c***************************************************************************** c c this routine will allow process jobtag to continue after c spawned processes have all checked in. it should only be called if c processes have been spawned by jobtag through the use of c the subroutine spawn. c c go to (1000,2000,...,N000), ientry(jobtag,N) c 1000 continue c . c . c . c do 10 j = 1,nproc c . c . (set parameters to define spawned process) c . c call nxtag(myid,jobtag) c call spawn(myid,jobtag,subname,) c 10 continue c return c 2000 continue c . c . c . c return c N000 continue c c return c end c c this subroutine returns the number of times process jobtag c has been entered. if that number is equal to the total c number nentrs of expected reentries then parmq(5,jobtag) c is set to 0 indicating no more reentries required. c c***************************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c report the entry point where process jobtag should resume c computation c if (nentrs .lt. 2) then write(6,*) '*************SCHED USER ERROR********************' write(6,*) ' user call to IENTRY with number of ' write(6,*) ' labels in nentrs set less than 2 ' write(6,*) ' from process ',jobtag write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif c ientry = parmq(1,jobtag) if (ientry .lt. nentrs) then parmq(5,jobtag) = nentrs else parmq(5,jobtag) = 0 endif c return c c last card of ientry c end subroutine dump(endgrf,igraph) parameter (nslots = 30,nbuffr = 500) integer endgrf real igraph(nslots,nbuffr) integer ievent(nslots) c c this routine writes graphics output to a file c and resets endgrf to 1 c c do 100 j = 1,endgrf c write(6,15) (igraph(i,j),i = 1,nslots) c 100 continue c 15 format('dump: igraph',30f5.1) do 300 j = 1,endgrf-1 do 302 i = 1,nslots ievent(i) = igraph(i,j) 302 continue if( ievent(1) .eq. 0 ) $ write(3,301) (ievent(i),i=1,ievent(4)+4) if( ievent(1) .eq. 1 ) $ write(3,303) (ievent(i),i=1,2),igraph(3,j) if( ievent(1) .eq. 2 ) $ write(3,303) (ievent(i),i=1,2),igraph(3,j) if( ievent(1) .eq. 3 ) $ write(3,301) (ievent(i),i=1,3) if( ievent(1) .eq. 4 ) $ write(3,304) (ievent(i),i=1,3),igraph(4,j) if( ievent(1) .eq. 5 ) $ write(3,304) (ievent(i),i=1,3),igraph(4,j) 301 format(14i8) 303 format(2i8,1pe16.8) 304 format(3i8,1pe16.8) 300 continue c endgrf = 1 c return c c last card of dump c end logical function wait(jobtag,ienter) c integer jobtag,ienter c***************************************************************************** c c this routine will allow process jobtag to continue after c spawned processes have all checked in. it should only be called if c processes have been spawned by jobtag through the use of c the subroutine spawn. this routine must be used in conjunction with c subroutine prtspn. the required syntax is c c go to (1000,...,L000,...,N000), ientry(mytag,N) c 1000 continue c . c . c . c do 100 j = 1,nproc c . c . (set parameters to define spawned process) c . c call nxtag(mytag,jobtag) c call spawn(mytag,jobtag,subname,) c 100 continue c label = L c if (wait(jobtag,label)) return c L000 continue c . c . c . c c if this subroutine returns a value of .true. then the calling process c jobtag should issue a return. if a value of .false. is returned then c the calling process jobtag should resume execution at the c statement immediately following the reference to wait (ie. at L000 in c the example above. a return value .true. indicates that some spawned c process has not yet completed and checked in. a return value .false. c indicates all spawned processes have checked in. c c***************************************************************************** parameter (mxprcs = 1000,iprcs = 120,mxces = 24,nslots = 30) parameter (nbuffr = 500) integer parmq,readyq,phead,intspn,rhead,rtail,done,qtail integer*1 qlock,hrlock,trlock,qtlock common /qdata/ parmq(nslots,mxprcs),phead,intspn, & readyq(iprcs,mxces),rhead(mxces),rtail(mxces), & qtail common /qsync/ qlock(mxprcs),hrlock(mxces),trlock(mxces), & done,qtlock integer endgrf real igraph common /gphout/ endgrf,igraph(nslots,nbuffr) c c c check the icango counter to see if all spawned processes (kids) c have checked in. c icango = 1 call s_lock(qlock(jobtag)) icango = parmq(2,jobtag) - parmq(4,jobtag) call s_unlock(qlock(jobtag)) c if (icango .eq. 0) then c c all kids are done ... dont wait (ie return false) c wait = .false. c c record re_entry label where computation is to c resume after wait is complete c parmq(1,jobtag) = ienter c if (ienter .gt. parmq(5,jobtag)) then write(6,*) '*************SCHED USER ERROR*****************' write(6,*) ' executing SCHEDULE function WAIT ' write(6,*) ' return label larger than the maximum ' write(6,*) ' specified by user in call to ientry ' write(6,*) ' from process jobtag ',jobtag write(6,*) ' ' write(6,*) ' the maximum reentry number is ' write(6,*) ' ', parmq(5,jobtag) write(6,*) ' ' write(6,*) ' EXECUTION TERMINATED BY SCHED ' call dump(endgrf,igraph) stop endif c c set last re_entry indication (parmq(5,jobtag) = 0) c if this reentry point corresponds to last one c (recorded in parmq(5,jobtag) during call to ientry) c if (ienter .eq. parmq(5,jobtag)) parmq(5,jobtag) = 0 c else c c kids are not done c wait = .true. c c a checkin will be made so set the number of c entries to return label ienter - 1 to get c correct entry point after checkin c parmq(1,jobtag) = ienter - 1 c endif c return c c last card of wait c end c subroutine lckasn(ilock) integer*1 ilock call s_init_lock(ilock) return end c subroutine lockon(ilock) integer*1 ilock call s_lock(ilock) return end c subroutine lockoff(ilock) integer*1 ilock call s_unlock(ilock) return end c subroutine nops j = 1 return end SHAR_EOF if test -f 'maindp.f' then echo shar: over-writing existing file "'maindp.f'" fi cat << \SHAR_EOF > 'maindp.f' $STDUNIT double precision dd(500),ee(500),qq(500,500) double precision esave(500),dsave(500) double precision s,s2,t,t2,tnorm,tnorm2,res,res2 integer icase c double precision d(500),e(500),q(500,500) integer n,ldq common /prbdef/n,ldq,q,d,e common/prfpms/ksect,kgran double precision enorm,dummy external treeql c ldq = 500 c write(6,*) ' input nproc ... the number of processors ' read(5,*) nproc do 9999 n = 100,100 ksect = n/10 kgran = n/20 write(6,*) 'nprocs = ',nproc write(6,*) 'ksect = ',ksect, ' kgran = ',kgran write(6,*)'==============================' write(6,*)' n = ',n do 9998 icase = 1,1 write(6,*)'++++++++++++++++++++++++++++++' if( icase .eq. 1 ) write(6,*)' twos on diagonal' if( icase .eq. 2 ) write(6,*)' random numbers on diagonal' if( icase .eq. 3 ) write(6,*)' glued wilks eps = 1.d-8 ' if( icase .eq. 4 ) write(6,*)' glued wilks eps = 1.d-14 ' nsect = 2**ksect nd2 = n/2 go to (112,113,114,115), icase 112 do 13 i = 1,n d(i) = 2.0 e(i) = 1 13 continue go to 445 113 do 14 i = 1,n d(i) = rand(foo) e(i) = rand(foo) 14 continue go to 445 114 continue eps = 1.e-8 do 151 inum = 1,num do 15 i = 1,21 d((inum-1)*21+i) = iabs(11-i) e((inum-1)*21+i) = 1 15 continue 151 continue go to 444 115 continue eps = 1.e-14 do 161 inum = 1,num do 16 i = 1,n d((inum-1)*21+i) = iabs(11-i) e((inum-1)*21+i) = 1 16 continue 161 continue 444 continue do 443 inum = 1,num e(inum*21+1) = eps 443 continue 445 continue do 10 j = 1,n do 5 i = 1,n q(i,j) = 0.0 qq(i,j) = 0.0 5 continue q(j,j) = 1.0 qq(j,j) = 1.0 dd(j) = d(j) ee(j) = e(j) dsave(j) = d(j) esave(j) = e(j) 10 continue c if (n .eq. 50) go to 9999 t1 = second(gtime) c call tql2(ldq,n,dd,ee,qq,ierr) t2t = second(gtime) - t1 write(6,*) ' time for tql ',t2t c c the test problem has been defined now comes the numerical c refinements that will avoid cancellation c t1 = second(gtime) c call sched(nproc,treeql,n,ldq,q,d,e,ifail) c t2 = second(gtime) - t1 write(6,*) ' time for sesupd ',t2 if( ifail .gt. 1 ) write(6,*)' deflate from sesupd',ifail write(6,*)' ratio of tql2/new',t2t/t2 tnorm = 0.0d0 tnorm2 = 0.0d0 res = 0.0d0 res2 = 0.0d0 do 530 j = 1,n e(1) = dsave(1)*q(1,j) + $ esave(2)*q(2,j) - d(j)*q(1,j) ee(1) = dsave(1)*qq(1,j) + $ esave(2)*qq(2,j) - dd(j)*qq(1,j) do 400 i = 2,n-1 e(i) = esave(i)*q(i-1,j) + dsave(i)*q(i,j) + $ esave(i+1)*q(i+1,j) - d(j)*q(i,j) ee(i) = esave(i)*qq(i-1,j) + dsave(i)*qq(i,j) + $ esave(i+1)*qq(i+1,j) - dd(j)*qq(i,j) 400 continue e(n) = esave(n)*q(n-1,j) + dsave(n)*q(n,j) $ - d(j)*q(n,j) ee(i) = esave(n)*qq(n-1,j) + dsave(n)*qq(n,j) $ - dd(j)*qq(n,j) t = enorm(n,e) t2 = enorm(n,ee) res = max(res,t) res2 = max(res2,t2) if (t .gt. 1.0d-13) then c write(6,*)' j ',j, ' d ',d(j),' er ',t,' dd ',dd(j),' eer ',t2 c write(6,*) ' ev ',q(1,j),' eev ',qq(1,j) endif do 520 i = 1,n t = 0.0 s = 0.0 t2 = 0.0 s2 = 0.0 do 510 k = 1,n s2 = s2 + qq(k,i)*qq(k,j) s = s + q(k,i)*q(k,j) 510 continue if (i .eq. j) s2 = s2 - 1.0 if (i .eq. j) s = s - 1.0 t2 = abs(s2) t = abs(s) tnorm2 = max(tnorm2,t2) tnorm = max(tnorm,t) c if (t .gt. 1.0d-13) write(6,*) ' ij ',i,j,' tnorm ',tnorm 520 continue 530 continue write(6,*)' the residual for the tql values and vectors',res2 write(6,*)' the residual for the updated values and vectors',res write(6,*)' the tql norm of q*q sup t is',tnorm2 write(6,*)' the upd norm of q*q sup t is',tnorm write (6,*) ' spectrum [ ',d(1) ,',',d(n),']' 9998 continue 9999 continue stop end SHAR_EOF if test -f 'make1' then echo shar: over-writing existing file "'make1'" fi cat << \SHAR_EOF > 'make1' FILES = ftsubs.o putq.o second.o cputm.o FILES2 = ftsubs.graph.o putq.o second.o cputm.o sched : $(FILES) rm -f sched.a; ar q sched.a $(FILES); ranlib sched.a graph : $(FILES2) rm -f graph.a; ar q graph.a $(FILES2); ranlib graph.a .f.o : ; fortran -g -c -mp $*.f .c.o : ; cc -g -c $*.c SHAR_EOF if test -f 'makefile' then echo shar: over-writing existing file "'makefile'" fi cat << \SHAR_EOF > 'makefile' FILES = stuffspawn.o xtest : $(FILES) fortran -g -mp $(FILES) sched.a -o xtest gtest : $(FILES) fortran -g -mp $(FILES) graph.a -o gtest FILES2 = d_and_c.o xdandc : $(FILES2) fortran -g -mp $(FILES2) sched.a -o xdandc gdandc : $(FILES2) fortran -g -mp $(FILES2) graph.a -o gdandc FILES4 = maindp.o newevdp0.o stateig.o statses.o xeig : $(FILES4) fortran -g -mp $(FILES4) sched.a -o xeig geig : $(FILES4) fortran -g -mp $(FILES4) graph.a -o geig FILES5 = maindp.o newevdp0.o stateig.o stseswait.o xwait : $(FILES5) fortran -g -mp $(FILES5) sched.a -o xwait gwait : $(FILES5) fortran -g -mp $(FILES5) graph.a -o gwait FILES6 = example.o xexample : $(FILES6) fortran -g -mp $(FILES6) sched.a -o xexample gexample : $(FILES6) fortran -g -mp $(FILES6) graph.a -o gexample sched: make -f make1 sched graph: make -f make1 graph clean: /bin/rm -f *.o *.fpp *.dbg *~ .f.o : ; fortran -g -c -mp $*.f SHAR_EOF if test -f 'newevdp0.f' then echo shar: over-writing existing file "'newevdp0.f'" fi cat << \SHAR_EOF > 'newevdp0.f' $STDUNIT subroutine evupd(n,i,d,z,delta,rho,dlam,ifail) implicit real*8 (a-h,o-z) CVD$G NOCONCUR c*********************************************************************** c this subroutine computes the updated eigenvalues of a c rank one modification to a symmetric matrix. it is assumed c that the eigenvalues are in the array d, and that c c d(i) .ne. d(j) for i .ne. j c c it is also assumed that the eigenvalues are in increasing c order and that the value of rho is positive. this is c arranged by the calling subroutine sesupd, and is no loss c in generality. c it is also assumed that the values in the array z are c the squares of the components of the updatingvector. c c c the method consists of approximating the rational functions c c rho = sum(z(j)/((d(j)-d(i))/rho - lamda): j = i+1,n) c c phi =sum(z(j)/(d(j)-d(i))/rho - lamda): j = 1,i) c c by simple interpolating rational functions. this avoids c the need for safeguarding by bisection since c the convergence is monotone, and quadratic from any starting c point that is greater than zero but less than the solution c c c input variables... c n the length of all arrays c c i the i - th eigenvalue is computed c c d the original eigenvalues. it is assumed that they are c in order, d(i) .lt. d(j) for i .lt. j. c c z this array of length n contains the squares of c of the components of the updating vector c c delta this array of length n contains (d(j) - lamda(i))/rho c in its j - th component. these values will be used c to update the eigenvectors in sesupd. c c rho this is the scalar in the the symmetric updating c formula. c c dlam this scalar will contain the value of the i - th c updated eigenvalue on return c c ifail this integer variable indicates failure of the c updating process with value 1, and success c with value 0. c c c c************************************************************************** integer i,n,im1,ip1,ip2,niter dimension d(n),z(n), delta(n) double precision d,z,rho,delta,zero,one,two,tsave, 1 del,phi,dphi,psi,dpsi,lambda,oldlam, 2 t,temp,a,b,d1,w,eps,eps1,eta,dlam,dmax double precision epslon,enorm zero = 0.0d0 one = 1.0d0 two = 2.0d0 c c eps is machine precision c eta is the relative accuracy requirement on the roots. c c these values should be adjusted for the particular machine. c eps = epslon(one) eta = eps*8. c im1 = i - 1 ip1 = i + 1 ip2 = i + 2 niter = 1 lambda = zero oldlam = zero del = d(i) do 100 j = 1,n delta(j) = (d(j) - del)/rho 100 continue dmax = max(abs(d(1)), abs(d(n))) dmax = abs(rho) + dmax c c calculate initial guess c if (i .lt. n) * then del = d(ip1) else del = d(n) + rho endif a = zero do 200 j = 1,im1 a = a + rho*z(j)/(d(j) - del) 200 continue b = zero do 220 j = ip2,n b = b + rho*z(j)/(d(j) - del) 220 continue a = a + (b + one) if (ip1 .gt. n) * then t = z(i)/abs(a) else t = a*delta(ip1) b = t + z(i) + z(ip1) if (b .ge. zero) * then t = two*z(i)*delta(ip1)/ * (b + sqrt(abs(b*b - 4.0d0*t*z(i)))) else t = (b - sqrt(b*b - 4.0d0*t*z(i)))/(two*a) endif t = t/two endif c c test to see that the initial guess is not too close to endpoint c if (ip1 .le. n .and. t .ge. .9*delta(ip1)) * t = .9*delta(ip1) c c update the values of the array delta c 250 continue tsave = abs(t) do 300 j = 1,n delta(j) = delta(j) - t 300 continue lambda = lambda + t dlam = d(i) + rho*lambda c c evaluate psi and the derivative dpsi c dpsi = zero psi = zero do 400 j = 1,i t = z(j)/delta(j) psi = psi + t dpsi = dpsi + t/delta(j) 400 continue c c evaluate phi and the derivative dphi c dphi = zero phi = zero if (i .eq. n) go to 600 do 500 j = ip1,n t = z(j)/delta(j) phi = phi + t dphi = dphi + t/delta(j) 500 continue c c test for convergence c return if the test is satisfied c 600 continue w = one + phi + psi eps1 = eps*dmax*sqrt(dphi + dpsi) if ((abs(w) .le. eps1) .and. (abs(lambda - oldlam) .le. 1 eta*oldlam) ) then return endif c c return with ifail = -1 if convergence has not ocurred c within 45 iterations. c if (niter .lt. 45) go to 650 ifail = -1 return 650 continue niter = niter + 1 oldlam = lambda c c calculate the new step c if (i .ne. n) go to 700 c c otherwise c calculate the step for the special case i = n c t = (w*psi)/dpsi go to 250 700 continue del = delta(ip1) temp = psi/dpsi d1 = one + phi - del*dphi a = (del*(one + phi) + psi*temp)/d1 + temp b = (two*temp*del*w)/d1 t = sqrt(abs(a*a - two*b)) t = b/(a + t) if (t .lt. -eps) t = -tsave/two go to 250 c c last card of evupd c end subroutine deflat(j,jlam,jdel,indx,n,ldq,q,d,z,tol1) implicit real*8 (a-h,o-z) CVD$G NOCONCUR c c c rotate z components to 0 c assumes norm(z) .eq. 1 c integer indx(*),jdel,j,jlam,n,ldq double precision q(ldq,*),d(*),z(*),gamma,sigma,tau,t1,t2, * tol,tol1 c tol = tol1 gamma = z(jlam) sigma = z(j) tau = pythag(gamma,sigma) jdel = -1 if (tau .gt. 0.0d0) * then delta = d(j) - d(jlam) gamma = gamma/tau sigma = sigma/tau if (abs(delta*gamma*sigma) .le. tol) * then if (abs(gamma) .le. abs(sigma)) * then z(j) = tau z(jlam) = 0.0d0 jjlam = indx(jlam) jj = indx(j) do 100 i = 1,n t1 = q(i,jjlam) t2 = q(i,jj) q(i,jjlam) = -t1*sigma + t2*gamma q(i,jj) = t1*gamma + t2*sigma 100 continue temp = d(jlam)*sigma**2 + d(j)*gamma**2 d(j) = d(jlam)*gamma**2 + d(j)*sigma**2 d(jlam) = temp jdel = jlam jlam = j else z(jlam) = tau z(j) = 0.0d0 jjlam = indx(jlam) jj = indx(j) do 200 i = 1,n t1 = q(i,jjlam) t2 = q(i,jj) q(i,jjlam) = t1*gamma + t2*sigma q(i,jj) = -t1*sigma + t2*gamma 200 continue temp = d(jlam)*gamma**2 + d(j)*sigma**2 d(j) = d(jlam)*sigma**2 + d(j)*gamma**2 d(jlam) = temp jdel = j endif endif endif c return c c last card of deflat end function rand(foo) CVD$G NOCONCUR data init/1365/ init = mod(3125*init,65536) rand = (init - 32768.)/(2.**15) return end double precision function enorm(n,x) CVD$G NOCONCUR integer n double precision x(n),temp,tmax tmax = 0.0d0 do 100 j = 1,n temp = max(tmax,abs(x(j))) if (temp .gt. tmax) tmax = temp 100 continue if (tmax .eq. 0.0d0) then enorm = tmax else temp = 0.0d0 do 200 j = 1,n temp = temp + (x(j)/tmax)**2 200 continue enorm = tmax*sqrt(temp) endif return end subroutine tql2(nm,n,d,e,z,ierr) implicit real*8 (a-h,o-z) CVD$G NOCONCUR c integer i,j,k,l,m,n,ii,l1,l2,nm,mml,ierr double precision d(n),e(n),z(nm,n) double precision c,c2,c3,dl1,el1,f,g,h,p,r,s,s2,tst1,tst2,pythag c c this subroutine is a translation of the algol procedure tql2, c num. math. 11, 293-306(1968) by bowdler, martin, reinsch, and c wilkinson. c handbook for auto. comp., vol.ii-linear algebra, 227-240(1971). c c this subroutine finds the eigenvalues and eigenvectors c of a symmetric tridiagonal matrix by the ql method. c the eigenvectors of a full symmetric matrix can also c be found if tred2 has been used to reduce this c full matrix to tridiagonal form. c c on input c c nm must be set to the row dimension of two-dimensional c array parameters as declared in the calling program c dimension statement. c c n is the order of the matrix. c c d contains the diagonal elements of the input matrix. c c e contains the subdiagonal elements of the input matrix c in its last n-1 positions. e(1) is arbitrary. c c z contains the transformation matrix produced in the c reduction by tred2, if performed. if the eigenvectors c of the tridiagonal matrix are desired, z must contain c the identity matrix. c c on output c c d contains the eigenvalues in ascending order. if an c error exit is made, the eigenvalues are correct but c unordered for indices 1,2,...,ierr-1. c c e has been destroyed. c c z contains orthonormal eigenvectors of the symmetric c tridiagonal (or full) matrix. if an error exit is made, c z contains the eigenvectors associated with the stored c eigenvalues. c c ierr is set to c zero for normal return, c j if the j-th eigenvalue has not been c determined after 30 iterations. c c calls pythag for sqrt(a*a + b*b) . c c questions and comments should be directed to burton s. garbow, c mathematics and computer science div, argonne national laboratory c c this version dated august 1983. c c ------------------------------------------------------------------ c ierr = 0 if (n .eq. 1) go to 1001 c do 100 i = 2, n 100 e(i-1) = e(i) c f = 0.0d0 tst1 = 0.0d0 e(n) = 0.0d0 c do 240 l = 1, n j = 0 h = abs(d(l)) + abs(e(l)) if (tst1 .lt. h) tst1 = h c .......... look for small sub-diagonal element .......... do 110 m = l, n tst2 = tst1 + abs(e(m)) if (tst2 .eq. tst1) go to 120 c .......... e(n) is always zero, so there is no exit c through the bottom of the loop .......... 110 continue c 120 if (m .eq. l) go to 220 130 if (j .eq. 30) go to 1000 j = j + 1 c .......... form shift .......... l1 = l + 1 l2 = l1 + 1 g = d(l) p = (d(l1) - g) / (2.0d0 * e(l)) r = pythag(p,1.0d0) d(l) = e(l) / (p + sign(r,p)) d(l1) = e(l) * (p + sign(r,p)) dl1 = d(l1) h = g - d(l) if (l2 .gt. n) go to 145 c do 140 i = l2, n 140 d(i) = d(i) - h c 145 f = f + h c .......... ql transformation .......... p = d(m) c = 1.0d0 c2 = c el1 = e(l1) s = 0.0d0 mml = m - l c .......... for i=m-1 step -1 until l do -- .......... do 200 ii = 1, mml c3 = c2 c2 = c s2 = s i = m - ii g = c * e(i) h = c * p r = pythag(p,e(i)) e(i+1) = s * r s = e(i) / r c = p / r p = c * d(i) - s * g d(i+1) = h + s * (c * g + s * d(i)) c .......... form vector .......... do 180 k = 1, n h = z(k,i+1) z(k,i+1) = s * z(k,i) + c * h z(k,i) = c * z(k,i) - s * h 180 continue c 200 continue c p = -s * s2 * c3 * el1 * e(l) / dl1 e(l) = s * p d(l) = c * p tst2 = tst1 + abs(e(l)) if (tst2 .gt. tst1) go to 130 220 d(l) = d(l) + f 240 continue c .......... order eigenvalues and eigenvectors .......... do 300 ii = 2, n i = ii - 1 k = i p = d(i) c do 260 j = ii, n if (d(j) .ge. p) go to 260 k = j p = d(j) 260 continue c if (k .eq. i) go to 300 d(k) = d(i) d(i) = p c do 280 j = 1, n p = z(j,i) z(j,i) = z(j,k) z(j,k) = p 280 continue c 300 continue c go to 1001 c .......... set error -- no convergence to an c eigenvalue after 30 iterations .......... 1000 ierr = l 1001 return end double precision function pythag(a,b) double precision a,b c c finds sqrt(a**2+b**2) without overflow or destructive underflow c double precision p,r,s,t,u p = max(abs(a),abs(b)) if (p .eq. 0.0d0) go to 20 r = (min(abs(a),abs(b))/p)**2 10 continue t = 4.0d0 + r if (t .eq. 4.0d0) go to 20 s = r/t u = 1.0d0 + 2.0d0*s p = u*p r = (s/u)**2 * r go to 10 20 pythag = p return end double precision function epslon (x) double precision x c c estimate unit roundoff in quantities of size x. c double precision a,b,c,eps c c this program should function properly on all systems c satisfying the following two assumptions, c 1. the base used in representing floating point c numbers is not a power of three. c 2. the quantity a in statement 10 is represented to c the accuracy used in floating point variables c that are stored in memory. c the statement number 10 and the go to 10 are intended to c force optimizing compilers to generate code satisfying c assumption 2. c under these assumptions, it should be true that, c a is not exactly equal to four-thirds, c b has a zero for its last bit or digit, c c is not exactly equal to one, c eps measures the separation of 1.0 from c the next larger floating point number. c the developers of eispack would appreciate being informed c about any systems where these assumptions do not hold. c c this version dated 4/6/83. c a = 4.0d0/3.0d0 10 b = a - 1.0d0 c = b + b + b eps = abs(c-1.0d0) if (eps .eq. 0.0d0) go to 10 epslon = eps*abs(x) epslon = 1.0d-16 return end subroutine evdrv(k,kstart,kstop,ldq,n,q,d,rho,z,dlamda,q2, * w,indxp,indx,ksnrho,ifail) implicit real*8 (a-h,o-z) CVD$G NOCONCUR c c integer k,kstart,kstop,ldq,n,ifail integer indx(n),indxp(n),ksnrho logical rhoge0 double precision * q(ldq,n),d(n),z(n),q2(ldq,n),dlamda(n), * w(n),rho c c local variables c double precision x(300),delta(300) double precision t,s,dlam double precision enorm rhoge0 = .true. if (ksnrho .lt. 0) rhoge0 = .false. c do 2700 j = kstart,kstop i = j if (.not. rhoge0) i = k - i + 1 call evupd(k,i,dlamda,w,delta,rho,dlam,ifail) c c if the zero finder failed the computation is terminated c if (ifail .eq. 1) return c if (rhoge0) * then jp = indxp(j) else dlam = -dlam jp = indxp(k-j+1) endif c d(jp) = dlam c c compute the updated eigenvectors c do 2200 i1 = 1,n x(i1) = 0.0d0 2200 continue do 2500 jj = 1,k j3 = jj if (.not. rhoge0) j3 = k - jj + 1 jp = indxp(j3) jjpp = indx(jp) s = z(jp)/delta(j3) c qq(jj,j) = s do 2300 i1 = 1,n x(i1) = x(i1) + q(i1,jjpp)*s 2300 continue 2500 continue t = enorm(n,x) j3 = j if (.not. rhoge0) j3 = k - j + 1 jp = indxp(j3) do 2600 i1 = 1,n q2(i1,jp) = x(i1)/t 2600 continue 2700 continue c return c c last card of evdrv c end SHAR_EOF if test -f 'oldtest' then echo shar: over-writing existing file "'oldtest'" fi cat << \SHAR_EOF > 'oldtest' make -f make1 sched fortran -g -c -mp ftsubs.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 645 lines. File ftsubs.fpp cc -g -c putq.c fortran -g -c -mp second.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 23 lines. File second.fpp cc -g -c cputm.c rm -f sched.a; ar q sched.a ftsubs.o putq.o second.o cputm.o; ranlib sched.a ar: creating sched.a make -f make1 graph fortran -g -c -mp ftsubs.graph.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 787 lines. File ftsubs.graph.fpp rm -f graph.a; ar q graph.a ftsubs.graph.o putq.o second.o cputm.o; ranlib graph.a ar: creating graph.a fortran -g -c -mp stuffspawn.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 149 lines. File stuffspawn.fpp fortran -g -mp stuffspawn.o sched.a -o xtest input nprocs time = +3.870000 nprocs = 12 +1.00000000000000 -2.00000000000000 -6.00000000000000 -8.00000000000000 -1.00000000000000E+001 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +2.00000000000000 -3.00000000000000 -8.00000000000000 -1.00000000000000E+001 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +3.00000000000000 -4.00000000000000 -1.00000000000000E+001 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +4.00000000000000 -5.00000000000000 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +5.00000000000000 -6.00000000000000 -1.40000000000000E+001 -1.60000000000000E+001 +6.00000000000000 -7.00000000000000 -1.60000000000000E+001 +7.00000000000000 -8.00000000000000 +8.00000000000000 Programmed STOP fortran -g -mp stuffspawn.o graph.a -o gtest input nprocs time = +6.040000 nprocs = 12 +1.00000000000000 -2.00000000000000 -6.00000000000000 -8.00000000000000 -1.00000000000000E+001 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +2.00000000000000 -3.00000000000000 -8.00000000000000 -1.00000000000000E+001 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +3.00000000000000 -4.00000000000000 -1.00000000000000E+001 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +4.00000000000000 -5.00000000000000 -1.20000000000000E+001 -1.40000000000000E+001 -1.60000000000000E+001 +5.00000000000000 -6.00000000000000 -1.40000000000000E+001 -1.60000000000000E+001 +6.00000000000000 -7.00000000000000 -1.60000000000000E+001 +7.00000000000000 -8.00000000000000 +8.00000000000000 Programmed STOP fortran -g -c -mp d_and_c.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 55 lines. File d_and_c.fpp fortran -g -mp d_and_c.o sched.a -o xdandc input nprocs nlevls 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Programmed STOP fortran -g -mp d_and_c.o graph.a -o gdandc input nprocs nlevls 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Programmed STOP fortran -g -c -mp maindp.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 150 lines. File maindp.fpp fortran -g -c -mp newevdp0.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 465 lines. File newevdp0.fpp fortran -g -c -mp stateig.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 118 lines. File stateig.fpp fortran -g -c -mp statses.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 277 lines. File statses.fpp fortran -g -mp maindp.o newevdp0.o stateig.o statses.o sched.a -o xeig input nproc ... the number of processors nprocs = 12 ksect = 10 kgran = 5 ============================== n = 100 ++++++++++++++++++++++++++++++ twos on diagonal time for tql +0.000000E+00 time for sesupd +7.69999980926514 ratio of tql2/new+0.00000000000000E+000 the residual for the tql values and vectors+1.41421356237310 the residual for the updated values and vectors+3.55342003405926E-015 the tql norm of q*q sup t is+0.00000000000000E+000 the upd norm of q*q sup t is+1.42929017146048E-015 spectrum [ +3.86880573281076E-003,+3.99613119426719] Programmed STOP fortran -g -mp maindp.o newevdp0.o stateig.o statses.o graph.a -o geig input nproc ... the number of processors nprocs = 12 ksect = 10 kgran = 5 ============================== n = 100 ++++++++++++++++++++++++++++++ twos on diagonal time for tql +0.000000E+00 time for sesupd +8.53999996185303 ratio of tql2/new+0.00000000000000E+000 the residual for the tql values and vectors+1.41421356237310 the residual for the updated values and vectors+3.55342003405926E-015 the tql norm of q*q sup t is+0.00000000000000E+000 the upd norm of q*q sup t is+1.42929017146048E-015 spectrum [ +3.86880573281076E-003,+3.99613119426719] Programmed STOP fortran -g -c -mp stseswait.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 307 lines. File stseswait.fpp fortran -g -mp maindp.o newevdp0.o stateig.o stseswait.o sched.a -o xwait input nproc ... the number of processors nprocs = 12 ksect = 10 kgran = 5 ============================== n = 100 ++++++++++++++++++++++++++++++ twos on diagonal time for tql +0.000000E+00 time for sesupd +7.04999971389771 ratio of tql2/new+0.00000000000000E+000 the residual for the tql values and vectors+1.41421356237310 the residual for the updated values and vectors+3.55342003405926E-015 the tql norm of q*q sup t is+0.00000000000000E+000 the upd norm of q*q sup t is+1.42929017146048E-015 spectrum [ +3.86880573281076E-003,+3.99613119426719] Programmed STOP fortran -g -mp maindp.o newevdp0.o stateig.o stseswait.o graph.a -o gwait input nproc ... the number of processors nprocs = 12 ksect = 10 kgran = 5 ============================== n = 100 ++++++++++++++++++++++++++++++ twos on diagonal time for tql +0.000000E+00 time for sesupd +8.84000015258789 ratio of tql2/new+0.00000000000000E+000 the residual for the tql values and vectors+1.41421356237310 the residual for the updated values and vectors+3.55342003405926E-015 the tql norm of q*q sup t is+0.00000000000000E+000 the upd norm of q*q sup t is+1.42929017146048E-015 spectrum [ +3.86880573281076E-003,+3.99613119426719] Programmed STOP fortran -g -c -mp example.f FORTRAN 77 V3.1.3 (C) Copyright 1981, 1988 Silicon Valley Software Inc. 0 errors. 86 lines. File example.fpp fortran -g -mp example.o sched.a -o xexample sigma = +5.050000E+03 Programmed STOP fortran -g -mp example.o graph.a -o gexample sigma = +5.050000E+03 Programmed STOP SHAR_EOF if test -f 'putq.c' then echo shar: over-writing existing file "'putq.c'" fi cat << \SHAR_EOF > 'putq.c' #include #define MAXPARMS 20 struct parms { int static_link; int (*subname)(); long *parms[MAXPARMS]; }; shared struct parms indx[1000]; sched_(nprocs,parms) int *nprocs; struct parms parms; /* this procedure obtains nprocs physical processors devoted to the the execution of the parallel program indicated through parms which is a structure whose first entry is a subroutine name and whose remaining entries are parameters appearing in the calling sequence of that subroutine. */ { int libopn_(); bcopy(&parms, &indx[0], sizeof(struct parms)); /* the subroutine name and prameter list have been copied and placed in a special slot on the parmq then libopn is invoked to initialize pointers, grab physical processors and begin the computation */ libopn_(nprocs); return(0); } putq_(jobtag,parms) int *jobtag; struct parms parms; /* this procedure puts the descriptor of a schedulable process onto the problem queue. this process will be scheduled for execution when its data dependencies have been satisfied (indicated by icango==0). the argument parms is a structure whose first entry is a subroutine name and whose remaining entries are parameters appearing in the calling sequence of that subroutine. */ { register int j; int place_(); j = *jobtag; bcopy(&parms, &indx[j], sizeof(struct parms)); /* first the parms block is copied into the slot pointed to by by jobtag and then this descriptor is placed on the problem queue */ place_(jobtag); return(0); } spawn_(parent,jobtag,parms) int *parent,*jobtag; struct parms parms; /* this procedure puts the descriptor of a schedulable process onto the problem queue. this process will be scheduled for execution when its data dependencies have been satisfied (indicated by icango==0). the argument parms is a structure whose first entry is a subroutine name and whose remaining entries are parameters appearing in the calling sequence of that subroutine. the action of this procedure differs from putq in that the user does not assign jobtags or data dependencies. a parent may spawn any number of children but these child processes only report to the parent. */ { register int j,i; int place_(),clone_(); j = *jobtag; i = *parent; bcopy(&parms, &indx[j], sizeof(struct parms)); /* first the parms block is copied into the slot pointed to by by jobtag and then this descriptor is placed on the problem queue */ if (indx[j].subname == clone_) indx[j].subname = indx[i].subname; /* here we ask if this is a recursive spawning. if so the name clone has been called instead of subname so we replace the name clone by subname. */ place_(jobtag); return(0); } clone_() { /* this is a dummy routine to satisfy unresolved external */ return(0); } work_(id,jobtag) int *id,*jobtag; { int start2_(),gtprb_(); register int j,myjob; j = *id; if (j == 1) /* the worker whose id is 1 will execute the subroutine passed to sched. this subroutine executes the static data dependency graph. this graph must have at least one node. */ { indx[0].subname(indx[0].parms[0], indx[0].parms[1], indx[0].parms[2], indx[0].parms[3], indx[0].parms[4], indx[0].parms[5], indx[0].parms[6], indx[0].parms[7], indx[0].parms[8], indx[0].parms[9], indx[0].parms[10], indx[0].parms[11], indx[0].parms[12], indx[0].parms[13], indx[0].parms[14], indx[0].parms[15], indx[0].parms[16], indx[0].parms[17], indx[0].parms[18], indx[0].parms[19]); start2_(); } myjob = gtprb_(id,jobtag); while (myjob != 0) { j = *jobtag; if (myjob <= -1 ) { /* reenter... simple spawning was done all kids completed and no reentry is required. this indicates jobtag is all done and checkin can proceed. */ chekin_(jobtag); myjob = gtprb_(id,jobtag); } else { /* call subname().......... */ indx[j].subname(indx[j].parms[0], indx[j].parms[1], indx[j].parms[2], indx[j].parms[3], indx[j].parms[4], indx[j].parms[5], indx[j].parms[6], indx[j].parms[7], indx[j].parms[8], indx[j].parms[9], indx[j].parms[10], indx[j].parms[11], indx[j].parms[12], indx[j].parms[13], indx[j].parms[14], indx[j].parms[15], indx[j].parms[16], indx[j].parms[17], indx[j].parms[18], indx[j].parms[19]); chekin_(jobtag); myjob = gtprb_(id,jobtag); } } return(0); } SHAR_EOF if test -f 'second.f' then echo shar: over-writing existing file "'second.f'" fi cat << \SHAR_EOF > 'second.f' $SYSTEM $STDUNIT real function second(t) c c this routine will gather the user time for a process. c it has resolution of a millisecond c and uses the unix system call getrusage. c see the unix manual for details c returns time in seconds. c integer cputm c itime = cputm() second = float(itime)/1000.0 c c this statement is here to bump the time by a bit c incase no the interval was too small. c c second = second + second*1.0e-6 c return end SHAR_EOF if test -f 'stateig.f' then echo shar: over-writing existing file "'stateig.f'" fi cat << \SHAR_EOF > 'stateig.f' $STDUNIT subroutine treeql(n,ldq,q,d,e,ifail) implicit real*8 (a-h,o-z) CVD$G NOCONCUR c * integer n,ldq double precision d(*),e(*),q(ldq,*) c double precision z(500),dlamda(500),x(500),q2(500,500),w(500), * delta(500),rho(300) integer indx(500),indxp(500),nn(3,300),kwork(50,300),ifail common /wspace/z,dlamda,x,q2,w,delta,rho,indx,indxp,nn,kwork common/prfpms/nsmall,kgran external sesupd,tql2 c c this subroutine splits a problem in two parts c nn(1,1) = n nn(2,1) = 1 nn(3,1) = 1 ifail = 0 nsize = n nlevl = -1 1111 continue nlevl = nlevl + 1 if (nsize .gt. 2*nsmall) then c c the problem is large enough to split one more level c define the splits here in the 100 loop and then c place a call to sesupd on the queue to glue the results c together c nsize = nsize/2 do 100 id = 2**nlevl,2**(nlevl+1) - 1 c idl = 2*id idr = idl + 1 n1 = nn(1,id)/2 c nn(1,idl) = n1 nn(2,idl) = idl nn(3,idl) = nn(3,id) c nn(1,idr) = nn(1,id) - n1 nn(2,idr) = idr nn(3,idr) = nn(3,id) + n1 c isplt1 = nn(3,idr) isplt = isplt1 - 1 n1p1 = n1 + 1 rho(id) = e(isplt1) alpha = d(isplt) alphap = d(isplt1) sigma = 1.0d0 c if (sign(1.0d0,alpha)*sign(1.0d0,alphap) .ge. 0.0d0) * then if (alpha .lt. 0.0d0 .or. alphap .lt. 0.0d0) * then if(rho(id) .ne. 0.0d0) * then rho(id) = - rho(id) sigma = -sigma endif endif endif kwork(50,id) = sigma d(isplt) = d(isplt) - rho(id) d(isplt1) = d(isplt1) - rho(id) c c the problem is large enough to split into two parts c isplt = nn(3,id) c jobtag = id icango = 2 nchks = 1 if (jobtag .eq. 1) nchks = 0 list = id/2 call dep(jobtag,icango,nchks,list) call putq(jobtag,sesupd,nn(2,id),ldq,nn(1,id), * q(isplt,isplt),d(isplt),rho(id),z(isplt), * x(isplt),dlamda(isplt),q2(isplt,isplt), * delta(isplt),w(isplt),indxp(isplt), * indx(isplt),kwork(1,id),ifail) c 100 continue go to 1111 else c c we have reached the lowest level c granularity now warrants call to tql2 c do 200 id = 2**nlevl,2**(nlevl+1) - 1 c isplt = nn(3,id) c jobtag = id icango = 0 ncheks = 1 list = id/2 c call dep(jobtag,icango,ncheks,list) call putq(jobtag,tql2,ldq,nn(1,id),d(isplt),e(isplt), * q(isplt,isplt),ifail) c 200 continue endif c return c c last card of split c end SHAR_EOF if test -f 'statses.f' then echo shar: over-writing existing file "'statses.f'" fi cat << \SHAR_EOF > 'statses.f' $STDUNIT subroutine sesupd(myid,ldq,n,q,d,rho,z,x,dlamda,q2,delta,w,indxp, * indx,kwork,ifail) implicit real*8 (a-h,o-z) CVD$G NOCONCUR c*********************************************************************** c c c this subroutine will compute the updated eigensystem of a c of a symmetric matrix after modification by a rank one c symmetric matrix. c c a = qdq' + rho*z*z' c c it is assumed that the eigenvectors of the original matrix c are stored in q, and the eigenvalues are in d. c the algorithm consists of three stages... c c c the first stage constists of deflating the size of c the problem when there multiple eigenvalues or if there c zero of the vector q'z. for each such ocurrence the dimension c is reduced by one. c c the second stage consists of calculating the updated c eigenvalues of the reduced problem. this requires a call c to the zero finding routine evupd. c c the final stage consists of computing the updated eigenvectors c directly using the updated eigenvalue. c c c the algorithm requires o(n**2) operations to update the c eigenvectors, but n**3 + o(n**2) to update the eigenvectors. c c c input variables... c c n the dimension of the problem. q is n x n c c q an n x n matrix that contains the eigenvectors of c the original matrix on input and the updated c eigenvectors on output. c c d a vector of length n. the original eigenvalues are c contained in d on input. the updated eigenvectors c are contained on output. c c rho a scalar c c z a vector of length n. on input this vector c containes the updating vector. the contents of z c are destroyed during the updating process. c c x a working array of length n. c c dlamda a working array of length n c c q2 a working array of dimension n x n c c delta a working array of length n c c w a working array of length n. c c indx an integer array of length n. c c ifail this integer variable indicates failure of the c updating process with value 1, and success c with value 0. c c called subroutines... c c evupd a subroutine for calculating the updated eigenvalues. c c c*********************************************************************** integer ldq,n,ifail integer indx(n),indxp(n),kwork(*) integer ksize(20),kstart(20),kstop(20),kbins,msize,krem,prcsys integer kgran,ksect common/prfpms/ksect,kgran logical rhoge0 double precision * q(ldq,n),d(n),z(n),x(n),dlamda(n),q2(ldq,n),delta(n) * ,w(n),rho double precision eps,zero,one,two,s,t,evsprd,dmax,dlam double precision epslon,enorm external evdrv c c c eps is machine precision c ifail = 0 zero = 0.0d0 one = 1.0d0 two = 2.0d0 eps = epslon(one) rhoge0 = (rho .ge. zero) go to (1111,2222),ientry(myid,2) 1111 continue c c compute q(transpose)*z c n1 = n/2 n1p1 = n1 + 1 do 100 j = 1,n1 z(j) = q(n1,j) 100 continue sigma = kwork(50) if (sigma .gt. 0.0) then do 200 j = n1p1,n z(j) = q(n1p1,j) 200 continue else do 201 j = n1p1,n z(j) = -q(n1p1,j) 201 continue endif c c normalize z so that norm(z) = 1 c t = enorm(n,z) do 300 j = 1,n z(j) = z(j)/t indx(j) = j 300 continue rho = rho*t*t c c calculate the allowable deflation tolerence c tol = (1.0d2)*eps*max(abs(d(1)),abs(d(n))) if (abs(rho) .le. tol) return c c c order the eigenvalues c nm1 = n - 1 do 600 j = 1,nm1 t = d(1) s = z(1) inx = indx(1) k = n - j + 1 do 500 i = 2,k im1 = i - 1 if (d(i) .lt. t) go to 400 t = d(i) s = z(i) inx = indx(i) go to 500 400 continue d(im1) = d(i) d(i) = t z(im1) = z(i) z(i) = s indx(im1) = indx(i) indx(i) = inx 500 continue 600 continue c c if there multiple eigenvalues then the problem deflates. c here the number of equal eigenvalues are found. c then an elementary reflector is computed to rotate the c corresponding eigensubspace so that certain components of c z are zero in this new basis. c if (rhoge0) *then k = 0 k2 = n + 1 do 670 j = 1,n if(rho*abs(z(j)) .le. tol) * then c c deflate due to small z component c k2 = k2 - 1 indxp(k2) = j if (j .eq. n) go to 1111 else jlam = j go to 700 endif 670 continue 700 continue j = j + 1 if (j .gt. n) go to 800 if(rho*abs(z(j)) .le. tol) c if(rho*abs(z(j)) .le. 1.0d-10) * then c c deflate due to small z component c k2 = k2 - 1 indxp(k2) = j else call deflat(j,jlam,jdef,indx,n,ldq,q,d,z,tol) if (jdef .le. 0) * then k = k + 1 w(k) = z(jlam)**2 dlamda(k) = d(jlam) indxp(k) = jlam jlam = j else k2 = k2 - 1 indxp(k2) = jdef endif endif go to 700 800 continue c c record the last eigenvalue c k = k + 1 w(k) = z(jlam)**2 dlamda(k) = d(jlam) indxp(k) = jlam else jlam = n k = n + 1 k2 = 0 do 671 j = n,1,-1 if(abs(rho*z(j)) .le. tol) c if(abs(rho*z(j)) .le. 1.0d-10) * then c c deflate due to small z component c k2 = k2 + 1 indxp(n-k2+1) = j if (j .eq. 1) go to 1111 else jlam = j go to 701 endif 671 continue 701 continue j = j - 1 if (j .lt. 1) go to 801 if(abs(rho*z(j)) .le. tol) * then c c deflate due to small z component c k2 = k2 + 1 indxp(n-k2+1) = j else call deflat(j,jlam,jdef,indx,n,ldq,q,d,z,tol) if (jdef .le. 0) * then k = k - 1 w(n-k+1) = z(jlam)**2 dlamda(n-k+1) = -d(jlam) indxp(n-k+1) = jlam jlam = j else k2 = k2 + 1 indxp(n-k2+1) = jdef endif endif go to 701 801 continue c c record the last eigenvalue c k = k - 1 w(n-k+1) = z(jlam)**2 dlamda(n-k+1) = -d(jlam) indxp(n-k+1) = jlam k = n - k + 1 rho = -rho endif c c compute the updated eigenvalues of the deflated problem c c kbins = k/kgran + 1 kbins = min(kbins,20) krem = mod(k,kbins) msize = (k - krem)/kbins do 900 j = 1,kbins ksize(j) = msize 900 continue do 910 j = 1,krem ksize(j) = ksize(j) + 1 910 continue kstart(1) = 1 kstop(1) = ksize(1) kwork(49) = k kwork(48) = -1 if (rhoge0) kwork(48) = 1 if ( kbins .gt. 1 ) then c do 920 j = 2,kbins kstart(j) = kstop(j-1) + 1 kstop(j) = kstop(j-1) + ksize(j) kwork(j) = kstart(j) kwork(kbins+j) = kstop(j) call nxtag(myid,jtemp) call spawn(myid,jtemp,evdrv,kwork(49), * kwork(j),kwork(kbins+j), * ldq,n,q,d,rho,z,dlamda,q2, * w,indxp,indx,kwork(48),ifail) 920 continue call evdrv(k,kstart(1),kstop(1),ldq,n,q,d,rho,z, * dlamda,q2,w,indxp,indx,kwork(48),ifail) c c kbins-1 processes were spawned c place a join point here and resume when all spawned c processes have checked in c c else c c there was not sufficient granularity to warrant spawning c of additional processes to compute roots c kstop(1) = k call evdrv(k,kstart(1),kstop(1),ldq,n,q,d,rho,z, * dlamda,q2,w,indxp,indx,kwork(48),ifail) endif return c 2222 continue c c store the updated eigenvectors back into q c k = kwork(49) do 1100 j = k+1,n jp = indxp(j) jjpp = indx(jp) do 1000 i = 1,n q2(i,jp) = q(i,jjpp) 1000 continue 1100 continue do 1300 j = 1,n do 1200 i = 1,n q(i,j) = q2(i,j) 1200 continue 1300 continue c return c c last card of sesupd c end SHAR_EOF if test -f 'stseswait.f' then echo shar: over-writing existing file "'stseswait.f'" fi cat << \SHAR_EOF > 'stseswait.f' $STDUNIT subroutine sesupd(myid,ldq,n,q,d,rho,z,x,dlamda,q2,delta,w,indxp, * indx,kwork,ifail) implicit real*8 (a-h,o-z) CVD$G NOCONCUR c*********************************************************************** c c c this subroutine will compute the updated eigensystem of a c of a symmetric matrix after modification by a rank one c symmetric matrix. c c a = qdq' + rho*z*z' c c it is assumed that the eigenvectors of the original matrix c are stored in q, and the eigenvalues are in d. c the algorithm consists of three stages... c c c the first stage constists of deflating the size of c the problem when there multiple eigenvalues or if there c zero of the vector q'z. for each such ocurrence the dimension c is reduced by one. c c the second stage consists of calculating the updated c eigenvalues of the reduced problem. this requires a call c to the zero finding routine evupd. c c the final stage consists of computing the updated eigenvectors c directly using the updated eigenvalue. c c c the algorithm requires o(n**2) operations to update the c eigenvectors, but n**3 + o(n**2) to update the eigenvectors. c c c input variables... c c n the dimension of the problem. q is n x n c c q an n x n matrix that contains the eigenvectors of c the original matrix on input and the updated c eigenvectors on output. c c d a vector of length n. the original eigenvalues are c contained in d on input. the updated eigenvectors c are contained on output. c c rho a scalar c c z a vector of length n. on input this vector c containes the updating vector. the contents of z c are destroyed during the updating process. c c x a working array of length n. c c dlamda a working array of length n c c q2 a working array of dimension n x n c c delta a working array of length n c c w a working array of length n. c c indx an integer array of length n. c c ifail this integer variable indicates failure of the c updating process with value 1, and success c with value 0. c c called subroutines... c c evupd a subroutine for calculating the updated eigenvalues. c c c*********************************************************************** integer ldq,n,ifail integer indx(n),indxp(n),kwork(*) integer ksize(20),kstart(20),kstop(20),kbins,msize,krem,prcsys integer kgran,ksect common/prfpms/ksect,kgran common/wsync/iwrite logical rhoge0,wait double precision eps,zero,one,two,s,t,evsprd,dmax,dlam double precision epslon,enorm double precision * q(ldq,n),d(n),z(n),x(n),dlamda(n),q2(ldq,n),delta(n) * ,w(n),rho external evdrv c c c eps is machine precision c c write(6,*) ' in ses ldq n rho ifail ',ldq,n,rho,ifail c write(6,*) 'fses myid rho q d z x ',myid,rho,q(1,1),d(1),z(1),x(1) ifail = 0 zero = 0.0d0 one = 1.0d0 two = 2.0d0 eps = epslon(one) rhoge0 = (rho .ge. zero) go to (1111,2222),ientry(myid,2) 1111 continue c c compute q(transpose)*z c n1 = n/2 n1p1 = n1 + 1 do 100 j = 1,n1 z(j) = q(n1,j) 100 continue sigma = kwork(50) c write(6,*) ' sigma at 2 ',sigma if (sigma .gt. 0.0) then do 200 j = n1p1,n z(j) = q(n1p1,j) 200 continue else do 201 j = n1p1,n z(j) = -q(n1p1,j) 201 continue endif c c normalize z so that norm(z) = 1 c t = enorm(n,z) c call lockon(iwrite) c write(6,*) ' n t z1 zn ',n,t,z(1),z(n) c call lockoff(iwrite) do 300 j = 1,n z(j) = z(j)/t indx(j) = j 300 continue rho = rho*t*t c c calculate the allowable deflation tolerence c tol = (1.0d2)*eps*max(abs(d(1)),abs(d(n))) if (abs(rho) .le. tol) return c c c order the eigenvalues c c write(6,*) ' before order evs ',d(1),d(n) nm1 = n - 1 do 600 j = 1,nm1 t = d(1) s = z(1) inx = indx(1) k = n - j + 1 do 500 i = 2,k im1 = i - 1 if (d(i) .lt. t) go to 400 t = d(i) s = z(i) inx = indx(i) go to 500 400 continue d(im1) = d(i) d(i) = t z(im1) = z(i) z(i) = s indx(im1) = indx(i) indx(i) = inx 500 continue 600 continue c write(6,*) ' after order evs ',d(1),d(n) c c if there multiple eigenvalues then the problem deflates. c here the number of equal eigenvalues are found. c then an elementary reflector is computed to rotate the c corresponding eigensubspace so that certain components of c z are zero in this new basis. c if (rhoge0) *then k = 0 k2 = n + 1 do 670 j = 1,n if(rho*abs(z(j)) .le. tol) * then c c deflate due to small z component c k2 = k2 - 1 indxp(k2) = j if (j .eq. n) go to 1111 else jlam = j go to 700 endif 670 continue 700 continue j = j + 1 if (j .gt. n) go to 800 if(rho*abs(z(j)) .le. tol) c if(rho*abs(z(j)) .le. 1.0d-10) * then c c deflate due to small z component c k2 = k2 - 1 indxp(k2) = j else call deflat(j,jlam,jdef,indx,n,ldq,q,d,z,tol) if (jdef .le. 0) * then k = k + 1 w(k) = z(jlam)**2 dlamda(k) = d(jlam) indxp(k) = jlam jlam = j else k2 = k2 - 1 indxp(k2) = jdef endif endif go to 700 800 continue c c record the last eigenvalue c k = k + 1 w(k) = z(jlam)**2 dlamda(k) = d(jlam) indxp(k) = jlam else jlam = n k = n + 1 k2 = 0 do 671 j = n,1,-1 if(abs(rho*z(j)) .le. tol) c if(abs(rho*z(j)) .le. 1.0d-10) * then c c deflate due to small z component c k2 = k2 + 1 indxp(n-k2+1) = j if (j .eq. 1) go to 1111 else jlam = j go to 701 endif 671 continue 701 continue j = j - 1 if (j .lt. 1) go to 801 if(abs(rho*z(j)) .le. tol) * then c c deflate due to small z component c k2 = k2 + 1 indxp(n-k2+1) = j else call deflat(j,jlam,jdef,indx,n,ldq,q,d,z,tol) if (jdef .le. 0) * then k = k - 1 w(n-k+1) = z(jlam)**2 dlamda(n-k+1) = -d(jlam) indxp(n-k+1) = jlam jlam = j else k2 = k2 + 1 indxp(n-k2+1) = jdef endif endif go to 701 801 continue c c record the last eigenvalue c k = k - 1 w(n-k+1) = z(jlam)**2 dlamda(n-k+1) = -d(jlam) indxp(n-k+1) = jlam k = n - k + 1 rho = -rho endif c write(6,*) ' begin a problem ------------------------------- ' c write(6,*) ' ' c write(6,*) k,rho c do 6666 jj = 1,k c write(6,*) z(indxp(jj)),w(jj),dlamda(jj) c6666 continue c write(6,*) ' begin a problem ------------------------------- ' c c c compute the updated eigenvalues of the deflated problem c c kbins = k/kgran + 1 kbins = min(kbins,20) krem = mod(k,kbins) msize = (k - krem)/kbins do 900 j = 1,kbins ksize(j) = msize 900 continue do 910 j = 1,krem ksize(j) = ksize(j) + 1 910 continue c call lockon(iwrite) c write(6,*) ' from ses w ',w(1),w(2),w(k) c call lockoff(iwrite) kstart(1) = 1 kstop(1) = ksize(1) kwork(49) = k kwork(48) = -1 if (rhoge0) kwork(48) = 1 c write(6,*) ' rhoge0 kwork ',rhoge0,kwork(48) c if ( kbins .gt. 1 ) then c write(6,*) ' about to spawn',kbins,' evdrvs from ses ',myid c do 920 j = 2,kbins kstart(j) = kstop(j-1) + 1 kstop(j) = kstop(j-1) + ksize(j) kwork(j) = kstart(j) kwork(kbins+j) = kstop(j) c write(6,*) ' bef spawn evdrv ',kwork(49), c * kwork(j),kwork(kbins+j),kwork(48) c call lockon(iwrite) c write(6,*) ' about to spawn evdrv ',k,kstart(j),kstop(j) c call lockoff(iwrite) call nxtag(myid,jtemp) call spawn(myid,jtemp,evdrv,kwork(49), * kwork(j),kwork(kbins+j), * ldq,n,q,d,rho,z,dlamda,q2, * w,indxp,indx,kwork(48),ifail) c write(6,*) ' tag assigned ',jtemp 920 continue c call lockon(iwrite) c write(6,*) ' about to call evdrv ',k,kstart(1),kstop(1) c call lockoff(iwrite) call evdrv(k,kstart(1),kstop(1),ldq,n,q,d,rho,z, * dlamda,q2,w,indxp,indx,kwork(48),ifail) c c kbins-1 processes were spawned c place a join point here and resume when all spawned c processes have checked in c c else c c there was not sufficient granularity to warrant spawning c of additional processes to compute roots c kstop(1) = k c write(6,*) 'bef rho q d z x ',rho,q(1,1),d(1),z(1),x(1) c call lockon(iwrite) c write(6,*) ' about to call evdrv ',k,kstart(1),kstop(1) c call lockoff(iwrite) call evdrv(k,kstart(1),kstop(1),ldq,n,q,d,rho,z, * dlamda,q2,w,indxp,indx,kwork(48),ifail) endif if( wait(myid,2) ) return c 2222 continue c c store the updated eigenvectors back into q c k = kwork(49) do 1100 j = k+1,n jp = indxp(j) jjpp = indx(jp) do 1000 i = 1,n q2(i,jp) = q(i,jjpp) 1000 continue 1100 continue do 1300 j = 1,n do 1200 i = 1,n q(i,j) = q2(i,j) 1200 continue 1300 continue c call lockon(iwrite) c write(6,*) ' after ret in ses ',k,n c write(6,*) ' k n z1 zn ',k,n,q(1,1),q(n,1) c call lockoff(iwrite) c ifail = n-k c return c c last card of sesupd c end SHAR_EOF if test -f 'stuffspawn.f' then echo shar: over-writing existing file "'stuffspawn.f'" fi cat << \SHAR_EOF > 'stuffspawn.f' $STDUNIT double precision a,b common /prbdef/ a(1000),b(100),itmp(10),jtmp(10),myid(10) EXTERNAL PARALG nblks = 8 do 10 j = 1,10 itmp(j) = j jtmp(j) = j 10 continue write(6,*) ' input nprocs ' read (5,*) nprocs t1 = second(foo) c CALL SCHED(nprocs,paralg,nblks,a,b,itmp,jtmp,myid) c t2 = second(foo) write(6,*) ' time = ',t2-t1,' nprocs = ',nprocs nn = 36 do 100 j = 1,nn write(6,*) a(j) 100 continue stop end c subroutine paralg(n,a,b,itmp,jtmp,myid) integer itmp(*),jtmp(*),myid(*) double precision a(*),b(*) integer mychkn(1) EXTERNAL STUFF1 c c this is the driver for filling a packed triangular matrix with c i on the i-th diagonal and -(i+j) in the (i,j) off diagonal position c icount = 1 do 200 j = 1,n-1 c c c the j-th diagonal waits for the diagonal above to complete c c the j-th diagonal completion will allow c the (j+1)-st diagonal to start c c jobtag = j icango = 1 if (jobtag .eq. 1) icango = 0 nchks = 1 mychkn(1) = j+1 c c we just set up data dependencies and are ready to put c this process on the queue c CALL DEP(jobtag,icango,nchks,mychkn) myid(jobtag) = jobtag CALL PUTQ(jobtag,stuff1,myid(jobtag),n, & a(icount),jtmp(j),itmp(1)) c c when the data dependencies for process jobtag are satisfied c the following call will be made c c call stuff1(myid....,itmp(1)) c icount = icount + (n-j+1) 200 continue c icango = 1 nchks = 0 mychkn(1) = n+1 c CALL DEP(n,icango,nchks,mychkn) myid(n) = n CALL PUTQ(n,stuff1,myid(n),n,a(icount),jtmp(n),itmp(1)) c return end c subroutine stuff1(myid,n,a,j,itmp) double precision a(*) integer myid,n,j,itmp(*) logical wait EXTERNAL STUFF2,STUFF3 c c write(6,*) ' enter stuff1 ',myid,j c c write(6,*) ' enter stuff1 ',ientry(myid),myid go to (1111,2222,3333),ientry(myid,3) 1111 continue ii = 2 do 100 i = 2,n c CALL NXTAG(myid,jdummy) c write(6,*) ' about to spawn id jd ',myid,jdummy CALL SPAWN(myid,jdummy,stuff2,a(ii + 1),itmp(i),itmp(j)) c c this spawns a process that will execute a call to stuff2 c and report completion to process MYID c ii = ii + 1 100 continue call stuff2(a(2),itmp(1),itmp(j)) if (wait(myid,2)) return c c return to help out and then return here (at label 2222) c on the next reentry c 2222 continue ii = 2 do 200 i = 2,n c CALL NXTAG(myid,jdummy) c write(6,*) ' about to spawn 3 id jd ',myid,jdummy CALL SPAWN(myid,jdummy,stuff3,a(ii + 1),itmp(i),itmp(j)) c c this spawns a process that will execute a call to stuff2 c and report completion to process MYID c ii = ii + 1 200 continue call stuff2(a(2),itmp(1),itmp(j)) if (wait(myid,3)) return 3333 continue a(1) = j c return end subroutine stuff2(a,i,j) double precision a(*),one one = 1.0d0 a(1) = 0.0d0 c write(6,*) ' enter stuff2 ',i,j do 100 kk = 1,10000 a(1) = a(1) + one 100 continue a(1) = -(i+j) return end subroutine stuff3(a,i,j) double precision a(*),one,save one = 1.0d0 save = a(1) c write(6,*) ' enter stuff3 ',i,j do 100 kk = 1,10000 a(1) = a(1) + one 100 continue a(1) = -(i+j) + save return end c SHAR_EOF if test -f 'testrun' then echo shar: over-writing existing file "'testrun'" fi cat << \SHAR_EOF > 'testrun' make sched; make graph make xtest; xtest < data make gtest; gtest < data make xdandc; xdandc < data make gdandc; gdandc < data make xeig; xeig < data make geig; geig < data make xwait; xwait < data make gwait; gwait < data make xexample; xexample < data make gexample; gexample < data SHAR_EOF chmod +x 'testrun' # End of shell archive exit 0