UEMS vs WRFEMS run time

Questions and other topics related to UEMS 15.
meteo60
Posts: 104
Joined: Tue Apr 17, 2012 4:50 pm

UEMS vs WRFEMS run time

Post by meteo60 » Tue Jan 19, 2016 12:13 pm

Hi,
I use WRFEMS 3.4.1, the run time for my domain is approximately 1h50
For the same domain on UEMS 15.52, with same parameters, the run time is 3h30!
I'm trying to change some parameters but run time is always too long
Is anyone else has the same thing?

theocarter2911
Posts: 65
Joined: Fri Mar 08, 2013 5:22 am

Re: UEMS vs WRFEMS run time

Post by theocarter2911 » Wed Jan 20, 2016 1:28 pm

Hi,
Did you use arw before as well? Reason I mention, is that arw takes significantly longer than nmm to run. If that is not the issue, then I have no idea!
Theo

sweensta
Posts: 12
Joined: Thu Jul 02, 2015 10:17 am

Re: UEMS vs WRFEMS run time

Post by sweensta » Wed Jan 20, 2016 3:09 pm

Hi meteo60,

We have seen exactly the same thing. Our domain in WRFEMS took 1 hr 30 mins. When we transferred it to UEMS the same run takes 3 hrs 30 mins. These runs use the same core, same static data, same input data.
Perhaps we are missing something new in the config files that affects runtime. If you find out what it is, please let us know!

meteoadriatic
Posts: 1567
Joined: Wed Aug 19, 2009 10:05 am

Re: UEMS vs WRFEMS run time

Post by meteoadriatic » Wed Jan 20, 2016 5:39 pm

Do you have namelist.wrfm file from old install (wrfems/runs/yourdomain/static/namelist.wfrm)?

If you do, post both old and new here, so we can check what is different and possibly suggest solution.

alfe
Posts: 86
Joined: Thu Nov 25, 2010 8:13 pm

Re: UEMS vs WRFEMS run time

Post by alfe » Wed Jan 20, 2016 7:04 pm

Hi all,
Yes I have seen the same effect. But do you use adaptive time step or not ? In my case I use adaptive time step option both for EMS and UEMS.

smartie
Posts: 97
Joined: Sat May 21, 2011 7:34 am

Re: UEMS vs WRFEMS run time

Post by smartie » Thu Jan 21, 2016 9:56 am

Possibly due to UEMS selecting a shallow convection scheme . Make sure:

Code: Select all

SHCU_PHYSICS  = 0
You almost certainly don't need it unless for specialist applications or you're using the Grell-Freitas scheme (3)

norulz
Posts: 92
Joined: Thu Mar 12, 2015 1:43 pm

Re: UEMS vs WRFEMS run time

Post by norulz » Thu Jan 21, 2016 10:34 pm

Hi,

Are you perhaps using ADAPTIVE time step? It might be worth checking if it is working or running at the minimum set times.

I'm trying to use ADAPTIVE with UEMS and so far the timing is pegged to whatever I set as minimum time step. I was previously using NMM so this is new for me.

David

meteoadriatic
Posts: 1567
Joined: Wed Aug 19, 2009 10:05 am

Re: UEMS vs WRFEMS run time

Post by meteoadriatic » Fri Jan 22, 2016 7:12 am

norulz wrote:Hi,

Are you perhaps using ADAPTIVE time step? It might be worth checking if it is working or running at the minimum set times.

I'm trying to use ADAPTIVE with UEMS and so far the timing is pegged to whatever I set as minimum time step. I was previously using NMM so this is new for me.

David
Did you tried to not change min and max values from -1 ?

norulz
Posts: 92
Joined: Thu Mar 12, 2015 1:43 pm

Re: UEMS vs WRFEMS run time

Post by norulz » Fri Jan 22, 2016 9:07 am

Hello,

I tried both leaving the defaults and also setting various max, start and min settings such as:
max = 180,60 ; start = 90,30 ; min 72,24.

In all situations the run shows that adaptive timing is on but the time steps for the entire simulation are at the value set by "min".

My feeling is that the adaptive mode is on but the test that should allow the time step to increase is not engaged / working / needs other values than the defaults ( 1.2 ; 0.84).

I have seen some other post mentioning that there has been a problem with adaptive timing since version 3.6.1

What is your feeling on this?

David

norulz
Posts: 92
Joined: Thu Mar 12, 2015 1:43 pm

Re: UEMS vs WRFEMS run time

Post by norulz » Fri Jan 22, 2016 9:33 am

...more on this.

Does anyone know how I can check the values of the horizontal and vertical Courant numbers (CFL). This can help to see if Adaptive timing should kick in or not.

David

Post Reply