Speed up sync.

lib/ft4/ft4_decode.f90

@@ -19,7 +19,7 @@ subroutine ft4_decode(cdatetime0,tbuf,nfa,nfb,nQSOProgress,ncontest,nfqso, &
    complex cd2(0:NMAX/NDOWN-1)                  !Complex waveform
    complex cb(0:NMAX/NDOWN-1)
    complex cd(0:NN*NSS-1)                       !Complex waveform
-   complex ctwk(4*NSS),ctwk2(4*NSS)
+   complex ctwk(2*NSS),ctwk2(2*NSS,-16:16)
    complex csymb(NSS)
    complex cs(0:3,NN)
    real s4(0:3,NN)
@@ -63,7 +63,7 @@ subroutine ft4_decode(cdatetime0,tbuf,nfa,nfb,nQSOProgress,ncontest,nfqso, &
       1,0,0,1,0,1,1,0,0,0,0,1,0,0,0,1,0,1,0,0,1,1,1,1,0,0,1,0,1, &
       0,1,0,1,0,1,1,0,1,1,1,1,1,0,0,0,1,0,1/
    save fs,dt,tt,txt,twopi,h,one,first,linex,apbits,nappasses,naptypes, &
-      mycall0,hiscall0,msg0,cqstr0
+      mycall0,hiscall0,msg0,cqstr0,ctwk2
    
    call clockit('ft4_deco',0)
    hhmmss=cdatetime0(8:13)
@@ -82,6 +82,15 @@ subroutine ft4_decode(cdatetime0,tbuf,nfa,nfb,nQSOProgress,ncontest,nfqso, &
          enddo
       enddo
 
+      do idf=-16,16
+         a=0.
+         a(1)=real(idf)
+         ctwk=1.
+         call clockit('twkfreq1',0)
+         call twkfreq1(ctwk,2*NSS,fs/2.0,a,ctwk2(:,idf))
+         call clockit('twkfreq1',1)
+      enddo
+
       mrrr=2*mod(mrrr+rvec(59:77),2)-1
       m73=2*mod(m73+rvec(59:77),2)-1
       mrr73=2*mod(mrr73+rvec(59:77),2)-1
@@ -182,9 +191,8 @@ subroutine ft4_decode(cdatetime0,tbuf,nfa,nfb,nQSOProgress,ncontest,nfqso, &
    ndecodes=0
    do icand=1,ncand
       f0=candidate(1,icand)
-      if(f0.le.max(fa,10.0) .or. f0.ge.min(fb,4990.0)) cycle
       snr=candidate(3,icand)-1.0
-
+      if( f0.le.10.0 .or. f0.ge.4990.0 ) cycle
       call clockit('ft4_down',0)
       call ft4_downsample(iwave,f0,cd2)  !Downsample from 512 to 32 Sa/Symbol
       call clockit('ft4_down',1)
@@ -212,16 +220,10 @@ subroutine ft4_decode(cdatetime0,tbuf,nfa,nfb,nQSOProgress,ncontest,nfqso, &
          smax=-99.
          idfbest=0
          do idf=idfmin,idfmax,idfstp
-            a=0.
-            a(1)=real(idf)
-            ctwk=1.
-            call clockit('twkfreq1',0)
-            call twkfreq1(ctwk,4*NSS,fs,a,ctwk2)
-            call clockit('twkfreq1',1)
 
             call clockit('sync4d  ',0)
             do istart=ibmin,ibmax,ibstp
-               call sync4d(cd2,istart,ctwk2,1,sync,sync2)  !Find sync power
+               call sync4d(cd2,istart,ctwk2(:,idf),1,sync)  !Find sync power
                if(sync.gt.smax) then
                   smax=sync
                   ibest=istart

lib/ft4/sync4d.f90

@@ -1,13 +1,13 @@
-subroutine sync4d(cd0,i0,ctwk,itwk,sync,sync2)
+subroutine sync4d(cd0,i0,ctwk,itwk,sync)
 
 ! Compute sync power for a complex, downsampled FT4 signal.
 
   include 'ft4_params.f90'
   parameter(NP=NMAX/NDOWN,NSS=NSPS/NDOWN)
   complex cd0(0:NP-1)
-  complex csynca(4*NSS),csyncb(4*NSS),csyncc(4*NSS),csyncd(4*NSS)
+  complex csynca(2*NSS),csyncb(2*NSS),csyncc(2*NSS),csyncd(2*NSS)
-  complex csync2(4*NSS)
+  complex csync2(2*NSS)
-  complex ctwk(4*NSS)
+  complex ctwk(2*NSS)
   complex z1,z2,z3,z4
   complex zz1,zz2,zz3,zz4
   logical first
@@ -29,11 +29,11 @@ subroutine sync4d(cd0,i0,ctwk,itwk,sync,sync2)
     phic=0.0
     phid=0.0
     do i=0,3
-      dphia=twopi*icos4a(i)/real(NSS) 
+      dphia=2*twopi*icos4a(i)/real(NSS) 
-      dphib=twopi*icos4b(i)/real(NSS) 
+      dphib=2*twopi*icos4b(i)/real(NSS) 
-      dphic=twopi*icos4c(i)/real(NSS) 
+      dphic=2*twopi*icos4c(i)/real(NSS) 
-      dphid=twopi*icos4d(i)/real(NSS) 
+      dphid=2*twopi*icos4d(i)/real(NSS) 
-      do j=1,NSS
+      do j=1,NSS/2
         csynca(k)=cmplx(cos(phia),sin(phia)) 
         csyncb(k)=cmplx(cos(phib),sin(phib)) 
         csyncc(k)=cmplx(cos(phic),sin(phic)) 
@@ -46,7 +46,7 @@ subroutine sync4d(cd0,i0,ctwk,itwk,sync,sync2)
       enddo
     enddo
     first=.false.
-    fac=1.0/(4.0*NSS)
+    fac=1.0/(2.0*NSS)
   endif
 
   i1=i0                            !four Costas arrays
@@ -60,36 +60,18 @@ subroutine sync4d(cd0,i0,ctwk,itwk,sync,sync2)
   z4=0.
 
   if(itwk.eq.1) csync2=ctwk*csynca      !Tweak the frequency
-  if(i1.ge.0 .and. i1+4*NSS-1.le.NP-1) z1=sum(cd0(i1:i1+4*NSS-1)*conjg(csync2))
+  if(i1.ge.0 .and. i1+4*NSS-1.le.NP-1) z1=sum(cd0(i1:i1+4*NSS-1:2)*conjg(csync2))
 
   if(itwk.eq.1) csync2=ctwk*csyncb      !Tweak the frequency
-  if(i2.ge.0 .and. i2+4*NSS-1.le.NP-1) z2=sum(cd0(i2:i2+4*NSS-1)*conjg(csync2))
+  if(i2.ge.0 .and. i2+4*NSS-1.le.NP-1) z2=sum(cd0(i2:i2+4*NSS-1:2)*conjg(csync2))
 
   if(itwk.eq.1) csync2=ctwk*csyncc      !Tweak the frequency
-  if(i3.ge.0 .and. i3+4*NSS-1.le.NP-1) z3=sum(cd0(i3:i3+4*NSS-1)*conjg(csync2))
+  if(i3.ge.0 .and. i3+4*NSS-1.le.NP-1) z3=sum(cd0(i3:i3+4*NSS-1:2)*conjg(csync2))
 
   if(itwk.eq.1) csync2=ctwk*csyncd      !Tweak the frequency
-  if(i4.ge.0 .and. i4+4*NSS-1.le.NP-1) z4=sum(cd0(i4:i4+4*NSS-1)*conjg(csync2))
+  if(i4.ge.0 .and. i4+4*NSS-1.le.NP-1) z4=sum(cd0(i4:i4+4*NSS-1:2)*conjg(csync2))
 
   sync = p(z1) + p(z2) + p(z3) + p(z4)
 
-sync2=0.0
-!do i=1,4
-!  i1=i0+(i-1)*33*NSS
-!  if(i.eq.1) csync2=ctwk*csynca
-!  if(i.eq.2) csync2=ctwk*csyncb
-!  if(i.eq.3) csync2=ctwk*csyncc
-!  if(i.eq.4) csync2=ctwk*csyncd
-!  z1=sum(cd0(i1      :i1+  NSS-1)*conjg(csync2(      1:  NSS)))
-!  z2=sum(cd0(i1+  NSS:i1+2*NSS-1)*conjg(csync2(  NSS+1:2*NSS)))
-!  z3=sum(cd0(i1+2*NSS:i1+3*NSS-1)*conjg(csync2(2*NSS+1:3*NSS)))
-!  z4=sum(cd0(i1+3*NSS:i1+4*NSS-1)*conjg(csync2(3*NSS+1:4*NSS)))
-!  sync2=sync2 + abs(z1)**2+abs(z2)**2+abs(z3)**2+abs(z4)**2+&
-!       2*abs(z1*conjg(z2)+z2*conjg(z3)+z3*conjg(z4)) + &
-!       2*abs(z1*conjg(z3)+z2*conjg(z4)) + &
-!       2*abs(z1*conjg(z4))
-!enddo
-!sync2=sync2*(fac**2)
-
   return
 end subroutine sync4d