search

apengsigkarup / OceanWave3D-Fortran90

OceanWave3D - a very efficient coastal engineering research tool used worldwide for simulation of nonlinear and dispersive free surface waves in varying bathymetries from very deep to shallow water. Learn more about the model:
http://www2.compute.dtu.dk/~apek/OceanWave3D/
Other
68 stars 37 forks source link

RAM Consumption when wave spectra is used for Boundary Condition #10

Open Gapouliasis opened 3 years ago

Gapouliasis commented 3 years ago

Dear Allan,

I have encountered the following issue. When the boundary condition is set to sample from a spectra with given parameters, such as Jonswap, the memory utilization increases dramatically. After running a few tests, I was convinced that OCW3D stores the entire boundary timeseries in RAM. Is this correct? ​

If this is true then running large simulations becomes prohibitively expensive.

hbbingham commented 3 years ago

Dear George,

Yes, that is correct. This is a known issue which we hope to have time to address at some point, but for now that is a feature and a limitation.

Best regards,

Harry and Allan

From: George Pouliasis @.> Sent: 06 June 2021 16:14 To: apengsigkarup/OceanWave3D-Fortran90 @.> Cc: Subscribed @.***> Subject: [apengsigkarup/OceanWave3D-Fortran90] RAM Consumption when wave spectra is used for Boundary Condition (#10)

Dear Allan,

I have encountered the following issue. When the boundary condition is set to sample from a spectra with given parameters, such as Jonswap, the memory utilization increases dramatically. After running a few tests, I was convinced that OCW3D stores the entire boundary timeseries in RAM. Is this correct? ​

If this is true then running large simulations becomes prohibitively expensive.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/apengsigkarup/OceanWave3D-Fortran90/issues/10, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ADL6A3VTGAWGPHTG7S3AHETTRN7A5ANCNFSM46F7R2WA.