decreases memory usage by 6 times for MERRA2

[AnningCheng-NOAA]

decrease of the memory usage for MERRA2 aerosol climatology by reading 2 month of data instead of 12 months. When the model is running, the new data is read in the interpolation subroutine. The performance of the model has also improved by 6%. as shown in the two figures below
fix #774
fix # ufs-community/ufs-weather-model#903

w

.

[climbfuji]

This looks good. I am wondering: In physics/aerinterp.F90, lines 231/232-288/290 and lines 437/438-494/498 are essentially the same. Should this be put in a private subroutine in the same file and called from the two places?

[climbfuji]

Suggested change

	SUBROUTINE aerinterpol( me,master,nthrds,npts,IDATE,FHOUR,iflip, jindx1,jindx2, &
	SUBROUTINE aerinterpol(me,master,nthrds,npts,IDATE,FHOUR,iflip, jindx1,jindx2, &

[AnningCheng-NOAA]

Dom, It is only repeated two/three times. I am afraid using a new subroutine might slow the code down by defining a new set of variables in the new subroutine. A few arrays occupy large memory, relocating and delocating might take time.

…

On Tue, Nov 16, 2021 at 9:02 AM Dom Heinzeller ***@***.***> wrote: ***@***.**** requested changes on this pull request. This looks good. I am wondering: In physics/aerinterp.F90, lines 231/232-288/290 and lines 437/438-494/498 are essentially the same. Should this be put in a private subroutine in the same file and called from the two places? ------------------------------ In physics/aerinterp.F90 <#774 (comment)>: > @@ -256,15 +351,20 @@ END SUBROUTINE setindxaer !********************************************************************** !********************************************************************** ! - SUBROUTINE aerinterpol(me,master,nthrds,npts,IDATE,FHOUR,jindx1,jindx2, & + SUBROUTINE aerinterpol( me,master,nthrds,npts,IDATE,FHOUR,iflip, jindx1,jindx2, & ⬇️ Suggested change - SUBROUTINE aerinterpol( me,master,nthrds,npts,IDATE,FHOUR,iflip, jindx1,jindx2, & + SUBROUTINE aerinterpol(me,master,nthrds,npts,IDATE,FHOUR,iflip, jindx1,jindx2, & ------------------------------ In physics/aerinterp.F90 <#774 (comment)>: > @@ -310,6 +415,88 @@ SUBROUTINE aerinterpol(me,master,nthrds,npts,IDATE,FHOUR,jindx1,jindx2, & endif enddo n1 = n2 - 1 + if (n2 > 12) n2 = n2 -12 +! need to read a new month + if (n1.ne.n1sv) then + if (me == master) write(*,*)"read in a new month MERRA2", n2 ⬇️ Suggested change - if (me == master) write(*,*)"read in a new month MERRA2", n2 +#ifdef DEBUG + if (me == master) write(*,'(a,i3)') "Reading in MERRA2 data for month", n2 +#endif — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#774 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALQPMIOGI742FQ237TFCPATUMJP6NANCNFSM5HVS2IFA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[climbfuji]

Dom, It is only repeated two/three times. I am afraid using a new subroutine might slow the code down by defining a new set of variables in the new subroutine. A few arrays occupy large memory, relocating and delocating might take time.
…

The compiler will easily take care of this, maybe even inline the new subroutine. Code duplication is bad and should be avoided if possible.

[AnningCheng-NOAA]

Dom, I am trying to put the code in a subroutine now. Let see how it goes,

…

On Tue, Nov 16, 2021 at 10:23 AM Dom Heinzeller ***@***.***> wrote: Dom, It is only repeated two/three times. I am afraid using a new subroutine might slow the code down by defining a new set of variables in the new subroutine. A few arrays occupy large memory, relocating and delocating might take time. … <#m_-8902789229747407544_> The compiler will easily take care of this, maybe even inline the new subroutine. Code duplication is bad and should be avoided if possible. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#774 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALQPMIOBD4LPMGR6V6UORKLUMJZPVANCNFSM5HVS2IFA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[AnningCheng-NOAA]

Dom, I have just completed the changes by putting the repeated lines in subroutine read_netfaer. The code passed regression tests and the timing is about the same. The changes have been pushed in. Great idea! Anning On Tue, Nov 16, 2021 at 11:24 AM Anning Cheng - NOAA Affiliate < ***@***.***> wrote:

…

Dom, I am trying to put the code in a subroutine now. Let see how it goes, On Tue, Nov 16, 2021 at 10:23 AM Dom Heinzeller ***@***.***> wrote: > Dom, It is only repeated two/three times. I am afraid using a new > subroutine might slow the code down by defining a new set of variables in > the new subroutine. A few arrays occupy large memory, relocating and > delocating might take time. > … <#m_-776242511803897711_m_-8902789229747407544_> > > The compiler will easily take care of this, maybe even inline the new > subroutine. Code duplication is bad and should be avoided if possible. > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub > <#774 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ALQPMIOBD4LPMGR6V6UORKLUMJZPVANCNFSM5HVS2IFA> > . > Triage notifications on the go with GitHub Mobile for iOS > <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> > or Android > <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. > >

[climbfuji]

Looking great, thanks a lot. I had made one suggestion for change in aerinterp.F90 aroind line 286, please let me know if you can't see it.

[AnningCheng-NOAA]

dom, just made the change and pushed in, which should save some time.

…

On Tue, Nov 16, 2021 at 2:52 PM Dom Heinzeller ***@***.***> wrote: ***@***.**** commented on this pull request. Looking great, thanks a lot. I had made one suggestion for change in aerinterp.F90 aroind line 286, please let me know if you can't see it. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#774 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALQPMIIVUBOK7HZJR2FMWU3UMKZAPANCNFSM5HVS2IFA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[climbfuji]

Changes look good to me. Will approve after the regression testing is completed. We need to make sure that this PR andd the parent PRs for fv3atm and ufs-weather-model are all up to date with the latest development, and that the submodule pointers are all correct.