porousPlate2d

[Bhuttu]

I was trying to run the file named “porousPlate2d” and I got this message in every loop.
“./tmp/imageData/data/heatMapphysTemperature.p”, line 22: Matrix does not represent a grid”

I am using GCC 7.5 and windows subsystem for linux 18.04LTS
Simulation was still running without any problem.

[mgaedtke]

Hi Bhuttu,

Good that simulation is still working. This is just a warning from the gnuplot interface and will be fixed in the upcoming release.

Thank’s for mentioning!

Best Regards,
Max

[Rookie]

Dear Max

I’m having a similar problem, after I add these lines to my code：
vtmWriterStartTime.write(iT);
SuperEuklidNorm3D<T> normVel(velocity);
BlockReduction3D2D<T> planeReduction(normVel, {0, -1, 0}, 600, BlockDataSyncMode::ReduceOnly);
// write output as JPEG
heatmap::write(planeReduction, iT);

// write output as JPEG and changing properties
heatmap::plotParam<T> jpeg_Param;
jpeg_Param.name = “velocity”;
jpeg_Param.contourlevel = 5;
jpeg_Param.colour = “blackbody”;
jpeg_Param.zoomOrigin = {0., 0.};
jpeg_Param.zoomExtend = {1., 1.};
heatmap::write(planeReduction, iT, jpeg_Param);

if I run it in parallel it runs fine, but if I don’t I get this error:
“./channel/imageData/data/_EuklidNormphysVelocityiT0000000.p” line 21: Matrix does not represent a grid
“./channel/imageData/data/_velocityiT0000000.p” line 27: Natrix does not represent a grid

But the picture for this step is actually output.Is it that the code can only run in parallel?

Best Regards,
Rookie

[Adrian]

Any OpenLB case can be executed sequentially (this is just parallel with nProcess=1 after all :-)).

We also sometimes observe such gnuplot warnings/errors on some systems – generally you should be able to ignore them. Just to confirm: The output is correct despite the error on your system?

(In general I would recommend to not rely on the gnuplot interface for any serious post-processing – it is neither good for performance (as the gnuplot process blocks execution of the simulation) nor convenient to re-plot results for large simulations without rerunning them.)

[Rookie]

Dear Adrian

Yes, it outputs the image from step 0, and also as you said, this prevents the simulation from continuing. I just want to experiment more with what the code does, so I’d better remove that part of the code then.

Best Regards,
Rookie

[Adrian]

That is not what I meant by “blocks execution” – this referred to the simulation not being able to continue while gnuplot is executed to output a frame, reducing total performance.

So the program terminates after step 0 in your case?

[Rookie]

Dear Adrian

It’s not terminated, it’s when I turn off parallelism that I get these problems, it outputs the image from step 0 and the simulation doesn’t move, it doesn’t break. Normally gnuplot itself doesn’t have problems due to serial mode. I guess it could be because of this gnuplot problem or the simulation slows down when parallelism is turned off. It could be that gnuplot is not suitable for non-parallel use in this code?

Best Regards,
Rookie

[mathias]

Dear Rookie,

if the Spring School doesn’t work for you, then maybe becoming a member of the consortium (https://www.openlb.net/consortium/) is an alternative option, because unfortunately I can’t provide such detailed support here.

Best regards,
Mathias

[Author]

Posts