Migration status

  • Home Page: Activity map.

intro story DELWAQ

DELWAQ

DELWAQ is the engine of the D-Water Quality and D-Ecology programmes of the Delft3D suite. It is based on a rich library from which relevant substances and processes can be selected to quickly put water and sediment quality models together.

The processes library covers many aspects of water quality and ecology, from basic tracers, dissolved oxygen, nutrients, organic matter, inorganic suspended matter, heavy metals, bacteria and organic micro-pollutants, to complex algae and macrophyte dynamics. High performance solvers enable the simulation of long periods, often required to capture the full cycles of the processes being modelled.

The finite volume approach underlying DELWAQ allows it to be coupled to both the structured grid hydrodynamics of the current Delft3D-FLOW engine and the upcoming D-Flow Flexible Mesh engine (1D-2D-3D) of the Delft3D Flexible Mesh Suite (or even other models such as TELEMAC).

'DELWAQ in open source' is our invitation to all leading experts to collaborate in further development and research in the field of water quality, ecology and morphology using Delft3D. Feel free to post your DELWAQ related questions or comments in this dedicated forum space. If you are new to DELWAQ, the tutorial (in the user manual) is a good place to start. A list of DELWAQ related publications is available here.

** PLEASE TAG YOUR POST! **

 

 

Sub groups
D-Flow Flexible Mesh
DELWAQ

Cohesive sediments & muddy systems

 


Message Boards

Parrallel simulation and hyperthreading

Richard Measures, modified 4 Years ago.

Parrallel simulation and hyperthreading

Jedi Knight Posts: 178 Join Date: 3/23/11 Recent Posts
Hi,

I'm currently running a large simulation on a multicore windows desktop (4CPUs hyperthreaded to allow 8 logical processes). I don't fully understand the theory behind hyperthreading and was wondering if it was generally quicker to only have 4 parrallel processes or 8 in this situation? Alternatively - would I be better to turn hyperthreading off? I realise I could run some tests to find this out but was wondering if anyone else has already done this?

Cheers,
Richard
Adri Mourits, modified 4 Years ago.

RE: Parrallel simulation and hyperthreading

Yoda Posts: 1224 Join Date: 1/3/11 Recent Posts
Hi Richard,

Hyperthreading is very useful to speed up your computation when the running processes are doing different things (I/O, network access, computation). Unfortunately, when doing a large simulation, all processes are doing heavy computations at the same time. The optimum number of partitions on your 4CPU machine will be 4 or sometimes 3 (allowing one CPU to take care of the communication between the partitions). Assuming that the computation time is indeed the bottle neck for your simulation.

Switching off hyperthreading will only have a marginal effect.

If you have additional findings, please post it here.

Thanks.

Regards,

Adri
Richard Measures, modified 4 Years ago.

RE: Parrallel simulation and hyperthreading

Jedi Knight Posts: 178 Join Date: 3/23/11 Recent Posts
Hi Adri,

Thanks for your informative reply. This matches up with what I am experiencing but it's good to know that turning hyper-threading off would not make a big difference (given that it is useful for other software and quite involved to turn on/off).

Cheers,
Richard
RC
Rodrigo Carballo, modified 4 Years ago.

RE: Parrallel simulation and hyperthreading

Youngling Posts: 19 Join Date: 5/27/11 Recent Posts
Adri Mourits:
Hi Richard,

Hyperthreading is very useful to speed up your computation when the running processes are doing different things (I/O, network access, computation). Unfortunately, when doing a large simulation, all processes are doing heavy computations at the same time. The optimum number of partitions on your 4CPU machine will be 4 or sometimes 3 (allowing one CPU to take care of the communication between the partitions). Assuming that the computation time is indeed the bottle neck for your simulation.

Switching off hyperthreading will only have a marginal effect.

If you have additional findings, please post it here.

Thanks.

Regards,

Adri


Hi Adri.

It's possible to run in parallel an off-line coupled wave-flow simulation? It seems that there is a problem to identify the id of the com-files with the different subdomains. I created a thread in the general forum but I didn't receive any clarification. http://oss.deltares.nl/web/delft3d/general/-/message_boards/view_message/830539

Kind regards.
Adri Mourits, modified 4 Years ago.

RE: Parrallel simulation and hyperthreading

Yoda Posts: 1224 Join Date: 1/3/11 Recent Posts
Hi Rodrigo,

Although I didn't test that, I expect that to work. But you must be very careful when producing the com-files containing the wave information: you have to do that by performing an online(!) FLOW-parallel WAVE computation with exactly the same number of partitions and at least one FLOW time step.

Regards,

Adri