New Velocity Potential and Stream Function Calculations #1072
Conversation
|
@WenMeng-NOAA We have updated the computation for streamfunction and velocity potential for compatibility with UPP parallelization. The subroutine now uses a poisson solver with a convergence tester to reduce runtimes. Comparison of spectral and numerical results is here |
|
The poisson solver works but is substantially slower than the gather --> stptranf --> scatter operation to solve the equation spectrally. The gather is much cheaper than it sounds because it is TWO fields, (U and V) at one level, not the several hundred such fields that comprise state plus derivative variables. A gather of two fields takes about 0.06 seconds on hera at GFS (high) resolution and it hard to even measure at CFS resolution. The relaxations take about 10 seconds on hera for chi and psi together and less than a second for the spectral solver. Spectral solver takes 27 seconds for the GFS (high) resolution case while the poisson solver takes 161 seconds. Spectral solver takes less than a second for the CFS case while the poisson solver takes about three seconds. |
sorc/ncep_post.fd/COLLECT_LOC.f
Outdated
| @@ -81,6 +84,9 @@ SUBROUTINE COLLECT_LOC ( A, B ) | |||
| deallocate(buff) | |||
| deallocate(rbufs) | |||
|
|
|||
| tb=mpi_wtime() | |||
There was a problem hiding this comment.
@KarinaAsmar-NOAA Clean up the debugging code.
There was a problem hiding this comment.
@GeorgeVandenberghe-NOAA Would you please clean up the debugging part of COLLECT_LOC.f? Let me know when done and I'll push it to this branch.
sorc/ncep_post.fd/COLLECT_LOC.f
Outdated
| @@ -104,6 +110,8 @@ SUBROUTINE COLLECT_ALL ( A, B ) | |||
| real, dimension(im,jm), intent(out) :: b | |||
| integer ierr,n | |||
| real, allocatable :: rbufs(:) | |||
| real*8 tb,ta | |||
There was a problem hiding this comment.
@KarinaAsmar-NOAA Clean up the debugging code in this routine.
KarinaAsmar-NOAA
force-pushed
the
spectral_strm_vpot
branch
from
ad772e3 to
f88bf42
Compare
WenMeng-NOAA
changed the title
sorc/ncep_post.fd/UPP_PHYSICS.f
Outdated
| IF(J>1 .and. J<JM) THEN | ||
| pval=chi(i,j) | ||
| CHI(I,J) = 0.25*( PTMP(I-1,J)+PTMP(I+1,J)+PTMP(I,J-1)+PTMP(I,J+1) & | ||
| - DIV(I,J)/(wrk2(i,j)*wrk1(i,j)*wrk3(i,j)*wrk1(i,j)*COSL(i,j)*4.)) |
There was a problem hiding this comment.
I think entire sub-expression:
- DIV(I,J)/(wrk2(i,j)*wrk1(i,j)*wrk3(i,j)*wrk1(i,j)*COSL(i,j)*4.
can be computed outside the 'DO jj=1,300000' loop. A compiler might see this and optimize the code, but it might not, so I think it's worth trying'
There was a problem hiding this comment.
@DusanJovic-NOAA @KarinaAsmar-NOAA
Good idea! Maybe can reduce a few seconds. Worth trying. Also the ABSV-F term above.
There was a problem hiding this comment.
@WenMeng-NOAA Please let us know if the new changes with suggestions from @DusanJovic-NOAA @JesseMeng-NOAA reduce the runtimes
sorc/ncep_post.fd/UPP_PHYSICS.f
Outdated
| ! npass = npass2 | ||
| ! if (j > jm-jtem+1 .or. j < jtem) npass = npass3 |
There was a problem hiding this comment.
Remove if not needed
sorc/ncep_post.fd/MDL2P.f
Outdated
| @@ -1816,6 +1819,101 @@ SUBROUTINE MDL2P(iostatusD3D) | |||
| ENDIF | |||
| ENDIF | |||
| ! | |||
| !*** CHIPSI | |||
| ! | |||
| IF (IGET(1021) > 0 .or. IGET(1022) > 0) THEN | |||
There was a problem hiding this comment.
@JesseMeng-NOAA @KarinaAsmar-NOAA This new scheme will only applied in global applications. Please change as:
IF ( (IGET(1021) > 0 .or. IGET(1022) > 0). and. MODELNAME == 'GFS' ) THEN
| @@ -4497,6 +4500,483 @@ SUBROUTINE CALSLR_UUTAH(SLR) | |||
| END SUBROUTINE CALSLR_UUTAH | |||
| ! | |||
| !------------------------------------------------------------------------------------- | |||
| ! | |||
| !> Computes streamfunction and velocity potential from absolute vorticity | |||
There was a problem hiding this comment.
@KarinaAsmar-NOAA Since this is a new subroutine incorporated in UPP, you might reformat the fortran code provided by @JesseMeng-NOAA and @GeorgeVandenberghe-NOAA readable and consistent as:
- all lowercase
- indentation for loops, conditionals and blocks
|
You should change "GWVX RELAX TIME " to something like "RELAX TIME" or
"POISSON SOLUTION TIME".
…On Fri, Dec 6, 2024 at 6:55 PM KarinaAsmar-NOAA ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In sorc/ncep_post.fd/UPP_PHYSICS.f
<#1072 (comment)>:
> + ENDDO
+
+ CHI=0.
+ tb=mpi_wtime()
+ if (me .eq. 5) print 109,' GWVX RELAX TIME ',tb-ta
+ 109 format(a,f20.10,i10)
+ DO jj=1,300000
+ call exch(chi(ista_2l:iend_2u,jsta_2l:jend_2u))
+ PTMP=CHI
+ err=0
+ DO J=JSTA,JEND
+ DO I=ISTA,IEND
+ IF(J>1 .and. J<JM) THEN
+ pval=chi(i,j)
+ CHI(I,J) = 0.25*( PTMP(I-1,J)+PTMP(I+1,J)+PTMP(I,J-1)+PTMP(I,J+1) &
+ - DIV(I,J)/(wrk2(i,j)*wrk1(i,j)*wrk3(i,j)*wrk1(i,j)*COSL(i,j)*4.))
@WenMeng-NOAA <https://github.com/WenMeng-NOAA> Please let us know if the
new changes with suggestions from @DusanJovic-NOAA
<https://github.com/DusanJovic-NOAA> @JesseMeng-NOAA
<https://github.com/JesseMeng-NOAA> reduce the runtimes
—
Reply to this email directly, view it on GitHub
<#1072 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ANDS4FUGVSAMGB7UCBLPHST2EHXIHAVCNFSM6AAAAABQK7FXP6VHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDIOBVGY2DGNRZGA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
--
George W Vandenberghe
*Lynker Technologies at * NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2141
College Park, MD 20740
***@***.***
301-683-3769(work) 3017751547(cell)
|
@GeorgeVandenberghe-NOAA Sure I can change that. Are the rest of the 'GWVX' necessary or can they be removed? |
WenMeng-NOAA
added
Ready for Review
|
@KarinaAsmar-NOAA Your recent commits have increase the runtime to 2381 S. |
@WenMeng-NOAA Which commit needs to be removed? 808bc84 or b67fb2c? |
@WenMeng-NOAA It looks like the issue was removing the if lvls block (see comment). I restored it and tested with SFS data. Please let me know if runtimes are acceptable now. |
This PR adds the CPC-requested streamfunction and velocity potential at 200mb to SFS. It is meant to resolve [Issue #902 ] and update on the runtime issues from the previous PR #951 . Job scripts used for testing are in WCOSS2: /lfs/h2/emc/vpppg/noscrub/karina.asmar/vpot_strm/UPP (submit_run_gfsv16_wcoss2.sh and submit_run_sfs_wcoss2.sh).