64-BitBrainstorm_2026/AMSS-NCKU
8
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add OpenMP parallelization to BSSN RHS hot-path stencil routines

Browse Source 
Enable OpenMP threading for the dominant computational kernels:
- makefile.inc: add -qopenmp to f90appflags
- diff_new.f90: split fderivs/fdderivs into OpenMP interior + serial boundary
- kodiss.f90: split kodis into OpenMP interior + serial boundary
- lopsidediff.f90: add OMP PARALLEL DO COLLAPSE(2) to lopsided
- fmisc.f90: parallelize symmetry_bd bulk array copy
- bssn_rhs.f90: add OMP WORKSHARE to array-syntax operations

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
CGH0S7
2026-02-07 13:58:55 +08:00
parent 09ffdb553d
commit c6e4d4ab71
6 changed files with 162 additions and 56 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
AMSS_NCKU_source/bssn_rhs.f90
View File

@@ -168,6 +168,8 @@
call fderivs(ex,gyz,gyzx,gyzy,gyzz,X,Y,Z,SYM ,ANTI,ANTI,Symmetry,Lev)
call fderivs(ex,dzz,gzzx,gzzy,gzzz,X,Y,Z,SYM ,SYM ,SYM ,Symmetry,Lev)
!$OMP PARALLEL
!$OMP WORKSHARE
gxx_rhs = - TWO * alpn1 * Axx - F2o3 * gxx * div_beta + &
TWO *( gxx * betaxx + gxy * betayx + gxz * betazx)
@@ -186,7 +188,7 @@
gxy * betaxz + gyy * betayz + &
gxz * betaxy + gzz * betazy &
- gyz * betaxx
gxz_rhs = - TWO * alpn1 * Axz + F1o3 * gxz * div_beta + &
gxx * betaxz + gxy * betayz + &
gyz * betayx + gzz * betazx &
@@ -201,6 +203,8 @@
gupyy = ( gxx * gzz - gxz * gxz ) / gupzz
gupyz = - ( gxx * gyz - gxy * gxz ) / gupzz
gupzz = ( gxx * gyy - gxy * gxy ) / gupzz
!$OMP END WORKSHARE
!$OMP END PARALLEL
if(co == 0)then
! Gam^i_Res = Gam^i + gup^ij_,j
@@ -234,6 +238,8 @@
endif
! second kind of connection
!$OMP PARALLEL
!$OMP WORKSHARE
Gamxxx =HALF*( gupxx*gxxx + gupxy*(TWO*gxyx - gxxy ) + gupxz*(TWO*gxzx - gxxz ))
Gamyxx =HALF*( gupxy*gxxx + gupyy*(TWO*gxyx - gxxy ) + gupyz*(TWO*gxzx - gxxz ))
Gamzxx =HALF*( gupxz*gxxx + gupyz*(TWO*gxyx - gxxy ) + gupzz*(TWO*gxzx - gxxz ))
@@ -282,6 +288,8 @@
(gupxy * gupyz + gupyy * gupxz)* Axy + &
(gupxy * gupzz + gupyz * gupxz)* Axz + &
(gupyy * gupzz + gupyz * gupyz)* Ayz
!$OMP END WORKSHARE
!$OMP END PARALLEL
! Right hand side for Gam^i without shift terms...
call fderivs(ex,Lap,Lapx,Lapy,Lapz,X,Y,Z,SYM,SYM,SYM,Symmetry,Lev)
@@ -336,6 +344,8 @@
call fderivs(ex,Gamy,Gamyx,Gamyy,Gamyz,X,Y,Z,SYM ,ANTI,SYM ,Symmetry,Lev)
call fderivs(ex,Gamz,Gamzx,Gamzy,Gamzz,X,Y,Z,SYM ,SYM ,ANTI,Symmetry,Lev)
!$OMP PARALLEL
!$OMP WORKSHARE
Gamx_rhs = Gamx_rhs + F2o3 * Gamxa * div_beta - &
Gamxa * betaxx - Gamya * betaxy - Gamza * betaxz + &
F1o3 * (gupxx * fxx + gupxy * fxy + gupxz * fxz ) + &
@@ -375,6 +385,8 @@
gyyz = gxz * Gamxyy + gyz * Gamyyy + gzz * Gamzyy
gyzz = gxz * Gamxyz + gyz * Gamyyz + gzz * Gamzyz
gzzz = gxz * Gamxzz + gyz * Gamyzz + gzz * Gamzzz
!$OMP END WORKSHARE
!$OMP END PARALLEL
!compute Ricci tensor for tilted metric
call fdderivs(ex,dxx,fxx,fxy,fxz,fyy,fyz,fzz,X,Y,Z,SYM ,SYM ,SYM ,symmetry,Lev)
@@ -401,6 +413,8 @@
Ryz = gupxx * fxx + gupyy * fyy + gupzz * fzz + &
( gupxy * fxy + gupxz * fxz + gupyz * fyz ) * TWO
!$OMP PARALLEL
!$OMP WORKSHARE
Rxx = - HALF * Rxx + &
gxx * Gamxx+ gxy * Gamyx + gxz * Gamzx + &
Gamxa * gxxx + Gamya * gxyx + Gamza * gxzx + &
@@ -601,9 +615,13 @@
Gamxyz * gxzz + Gamyyz * gyzz + Gamzyz * gzzz + &
Gamxzz * gxzy + Gamyzz * gyzy + Gamzzz * gzzy + &
Gamxyz * gzzx + Gamyyz * gzzy + Gamzyz * gzzz )
!$OMP END WORKSHARE
!$OMP END PARALLEL
!covariant second derivative of chi respect to tilted metric
call fdderivs(ex,chi,fxx,fxy,fxz,fyy,fyz,fzz,X,Y,Z,SYM,SYM,SYM,Symmetry,Lev)
!$OMP PARALLEL
!$OMP WORKSHARE
fxx = fxx - Gamxxx * chix - Gamyxx * chiy - Gamzxx * chiz
fxy = fxy - Gamxxy * chix - Gamyxy * chiy - Gamzxy * chiz
fxz = fxz - Gamxxz * chix - Gamyxz * chiy - Gamzxz * chiz
@@ -626,11 +644,15 @@
Rxy = Rxy + (fxy - chix*chiy/chin1/TWO + gxy * f)/chin1/TWO
Rxz = Rxz + (fxz - chix*chiz/chin1/TWO + gxz * f)/chin1/TWO
Ryz = Ryz + (fyz - chiy*chiz/chin1/TWO + gyz * f)/chin1/TWO
!$OMP END WORKSHARE
!$OMP END PARALLEL
! covariant second derivatives of the lapse respect to physical metric
call fdderivs(ex,Lap,fxx,fxy,fxz,fyy,fyz,fzz,X,Y,Z, &
SYM,SYM,SYM,symmetry,Lev)
!$OMP PARALLEL
!$OMP WORKSHARE
gxxx = (gupxx * chix + gupxy * chiy + gupxz * chiz)/chin1
gxxy = (gupxy * chix + gupyy * chiy + gupyz * chiz)/chin1
gxxz = (gupxz * chix + gupyz * chiy + gupzz * chiz)/chin1
@@ -791,6 +813,8 @@
!!!! gauge variable part
Lap_rhs = -TWO*alpn1*trK
!$OMP END WORKSHARE
!$OMP END PARALLEL
#if (GAUGE == 0)
betax_rhs = FF*dtSfx
betay_rhs = FF*dtSfy