Improve Eady figure #145

jagoosw · 2023-09-18T16:33:38Z

This PR addresses #140 and increases the model resolution.

Closes #140

jagoosw · 2023-09-18T16:34:41Z

paper/paper.md

@@ -72,7 +72,7 @@ An example of a problem involving small-scale flow features is showcased in \aut
 ![In this simulation of baroclinic instability in the Eady problem, a background buoyancy gradient and corresponding thermal wind generates a sub-mesoscale eddy, roughly following the setup of @taylor:2016.
 To this physical setup, we added a medium complexity (9 tracers) biogeochemical model, some components of which are shown above.
 On top of this, we added particles modelling the growth of sugar kelp, which are free-floating and advected by the flow, and carbon dioxide exchange from the air.
-Thanks to Julia's speed and efficiency the above model (1 km × 1 km × 100 m with 64 × 64 × 16 grid points) took about 30 minutes of computing time to simulate 10 days of evolution on an Nvidia P100 GPU. 
+Thanks to Julia's speed and efficiency the above model (1 km × 1 km × 100 m with 512 × 512 × 64 grid points) took about X minutes of computing time to simulate 10 days of evolution on an Nvidia A100 GPU. 
 Panel (a) shows the domain with the colour representing the concentration of various biogeochemical tracer fields: inorganic carbon, organic carbon (dissolved and particulate), phytoplankton, and nutrients. 
 The increase in organic carbon concentration in the centre of the eddy can be seen, as well as carbon being subducted (most visible in the xz face in the organic carbon).


@johnryantaylor I will update this caption when I have made the new figure, but do you want to add a sentence about the physics visible at the high resolution?

Good point - I suggested a sentence above

paper/paper.md

navidcy · 2023-09-19T15:23:00Z

paper/figures/eady.jl

+
+u, v, w = model.velocities
+
+# Periodically save the velocities and vorticity to a file.


Suggested change

# Periodically save the velocities and vorticity to a file.

# Periodically save the velocities to a file.

navidcy · 2023-09-19T15:23:20Z

paper/figures/eady.jl

+
+n = 32
+
+x = arch_array(arch, [repeat([-200.0, -100.0, 0.0, 100.0, 200.0], 1, 5)...] .+ Lx / 2)


Suggested change

x = arch_array(arch, [repeat([-200.0, -100.0, 0.0, 100.0, 200.0], 1, 5)...] .+ Lx / 2)

x = arch_array(arch, [repeat([-200.0, -100.0, 0.0, 100.0, 200.0], 1, 5)...] .+ Lx / 2)

Also, where does this 5 come from? Seems hardcoded. Should we define a variable so that it's clearer how one can change it?

I'll update this since there was an issue here too

navidcy · 2023-09-19T15:24:04Z

paper/figures/eady.jl

+
+particles = SLatissima(; architecture = arch,
+                         x, y, z, 
+                         A = arch_array(arch, 5.0 .* ones(n)), N = arch_array(arch, 0.01.* ones(n)), C = arch_array(arch, 0.18.* ones(n)), 


Suggested change

A = arch_array(arch, 5.0 .* ones(n)), N = arch_array(arch, 0.01.* ones(n)), C = arch_array(arch, 0.18.* ones(n)),

A = arch_array(arch, 5.0 .* ones(n)), N = arch_array(arch, 0.01 .* ones(n)), C = arch_array(arch, 0.18 .* ones(n)),

navidcy · 2023-09-19T15:24:56Z

paper/figures/eady.jl

+                                                          schedule = TimeInterval(6hours),
+                                                          filename = "eady_particles.jld2",
+                                                          overwrite_existing = true)


Suggested change

schedule = TimeInterval(6hours),

filename = "eady_particles.jld2",

overwrite_existing = true)

schedule = TimeInterval(6hours),

filename = "eady_particles.jld2",

overwrite_existing = true)

navidcy · 2023-09-19T15:25:53Z

paper/figures/eady.jl

+particles = SLatissima(; architecture = arch,
+                         x, y, z, 
+                         A = arch_array(arch, 5.0 .* ones(n)), N = arch_array(arch, 0.01.* ones(n)), C = arch_array(arch, 0.18.* ones(n)), 
+                         latitude = 57.5,


where does this latitude value comes from? is this the value that corresponds to the f value we use above?

It was arbitrary for fast kelp growth, but I've changed now to correspond to the coriolis force.

navidcy · 2023-09-19T15:26:35Z

paper/paper.md

@@ -72,7 +72,7 @@ An example of a problem involving small-scale flow features is showcased in \aut
 ![In this simulation of baroclinic instability in the Eady problem, a background buoyancy gradient and corresponding thermal wind generates a sub-mesoscale eddy, roughly following the setup of @taylor:2016.
 To this physical setup, we added a medium complexity (9 tracers) biogeochemical model, some components of which are shown above.
 On top of this, we added particles modelling the growth of sugar kelp, which are free-floating and advected by the flow, and carbon dioxide exchange from the air.
-Thanks to Julia's speed and efficiency the above model (1 km × 1 km × 100 m with 64 × 64 × 16 grid points) took about 30 minutes of computing time to simulate 10 days of evolution on an Nvidia P100 GPU. 
+Thanks to Julia's speed and efficiency the above model (1 km × 1 km × 100 m with 512 × 512 × 64 grid points) took about X minutes of computing time to simulate 10 days of evolution on an Nvidia A100 GPU. 


"X minutes" needs to be replaced with the number

Yeah not quite finished running!

navidcy

some remarks

jagoosw · 2023-09-24T13:23:08Z

Do we think the eady plot is now sufficiently clear and large enough text?

jagoosw · 2023-09-24T13:24:03Z

It doesn't have the colorbars or kelp line plot but I don't think they add much to the demonstration value of the plot.

jagoosw · 2023-09-24T13:25:35Z

I think I would like to add to one of the sentences about performance an acknowledgement of and reference for KA/JuliaGPU

jagoosw · 2023-09-24T13:29:18Z

I've also just noticed that the reference for Oceananigans isn't at the first instance it was mentioned, should we change?

Co-authored-by: johnryantaylor <[email protected]>

navidcy · 2023-09-24T14:05:56Z

Do we think the eady plot is now sufficiently clear and large enough text?

Fontsizes for axis labels don't seem to have increased. Have they?
Also lack of colorbars makes it less clear for me.
Another suggestion is to alter a bit the colorrange for the left-most part of the figure so that the y-z structure shows up better?

navidcy · 2023-09-24T14:08:28Z

src/Boundaries/Sediments/simple_multi_G.jl

+                     sediment.denitrification_params.E * log(reactivity) ^ 2 +
+                     sediment.denitrification_params.F * log(O₂) * log(reactivity)) / (Cᵐⁱⁿ * day)
+        =#


shall we delete these commented code?

I guess so, might be better todo in #138 since I've made a lot of changes to the sediment there

src/Boundaries/Sediments/simple_multi_G.jl

Co-authored-by: Navid C. Constantinou <[email protected]>

jagoosw · 2023-09-24T17:39:04Z

Fontsizes for axis labels don't seem to have increased. Have they? Also lack of colorbars makes it less clear for me. Another suggestion is to alter a bit the colorrange for the left-most part of the figure so that the y-z structure shows up better?

I think the axis labels appear the same size as other figures now? I think the main complaint was about the ticks.

Is this better with color bars:

I've adjusted the colorrange so it clips on the visible faces (rather than the whole volumes min/max), is that okay?

navidcy · 2023-09-24T17:47:16Z

Fontsizes for axis labels don't seem to have increased. Have they? Also lack of colorbars makes it less clear for me. Another suggestion is to alter a bit the colorrange for the left-most part of the figure so that the y-z structure shows up better?

I think the axis labels appear the same size as other figures now? I think the main complaint was about the ticks.

Oh I see! Okie then!

Is this better with color bars:

Yes! Much better!

I've adjusted the colorrange so it clips on the visible faces (rather than the whole volumes min/max), is that okay?

Try 0.8*min / max ? Ideally I'd like a bit more contrast for the left-part of the figure. But it's also fine as is!

jagoosw · 2023-09-24T17:55:16Z

I think this is a bit better:

Its kind of hard to make some of these colormaps give more contrast because they're meant to be perceptually uniform

added eady script and changed resolution in text

ac93375

jagoosw commented Sep 18, 2023

View reviewed changes

johnryantaylor reviewed Sep 18, 2023

View reviewed changes

paper/paper.md Outdated Show resolved Hide resolved

navidcy reviewed Sep 19, 2023

View reviewed changes

jagoosw added 2 commits September 20, 2023 16:31

changed params from named tuple to tuple

6c87930

updated eady figure

a566ab1

jagoosw marked this pull request as ready for review September 24, 2023 13:22

added sentence about color ordering

8a70f46

jagoosw and others added 3 commits September 24, 2023 14:33

added citation for KA and CUDA

526efab

corrected citation

1257d28

Apply suggestions from code review

69eb822

Co-authored-by: johnryantaylor <[email protected]>

navidcy reviewed Sep 24, 2023

View reviewed changes

src/Boundaries/Sediments/simple_multi_G.jl Outdated Show resolved Hide resolved

Update src/Boundaries/Sediments/simple_multi_G.jl

c3506d2

Co-authored-by: Navid C. Constantinou <[email protected]>

jagoosw added 3 commits September 24, 2023 18:42

modified figire

a7fc9e0

included Julia GPU in acknowledgements

675d256

Merge branch 'joss-paper' into jsw/joss-eady-fig

d12ee7c

jagoosw added 2 commits September 24, 2023 18:56

changed colorrange

edf470a

updated eady plot file

93e655d

navidcy approved these changes Sep 24, 2023

View reviewed changes

changed colorrange again

0349288

jagoosw merged commit 591d3d2 into joss-paper Sep 24, 2023
1 check passed

jagoosw deleted the jsw/joss-eady-fig branch September 24, 2023 18:04

navidcy mentioned this pull request Sep 26, 2023

Figure 1 from JOSS paper #140

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Improve Eady figure #145

Improve Eady figure #145

jagoosw commented Sep 18, 2023

jagoosw Sep 18, 2023

johnryantaylor Sep 18, 2023

navidcy Sep 19, 2023

navidcy Sep 19, 2023 •

edited

Loading

jagoosw Sep 20, 2023

navidcy Sep 19, 2023

navidcy Sep 19, 2023

navidcy Sep 19, 2023

jagoosw Sep 20, 2023

navidcy Sep 19, 2023

jagoosw Sep 20, 2023

navidcy left a comment

jagoosw commented Sep 24, 2023

jagoosw commented Sep 24, 2023

jagoosw commented Sep 24, 2023

jagoosw commented Sep 24, 2023

navidcy commented Sep 24, 2023

navidcy Sep 24, 2023

jagoosw Sep 24, 2023

jagoosw commented Sep 24, 2023

navidcy commented Sep 24, 2023

jagoosw commented Sep 24, 2023 •

edited

Loading


		u, v, w = model.velocities

		# Periodically save the velocities and vorticity to a file.

	# Periodically save the velocities and vorticity to a file.
	# Periodically save the velocities to a file.


		n = 32

		x = arch_array(arch, [repeat([-200.0, -100.0, 0.0, 100.0, 200.0], 1, 5)...] .+ Lx / 2)

	A = arch_array(arch, 5.0 .* ones(n)), N = arch_array(arch, 0.01.* ones(n)), C = arch_array(arch, 0.18.* ones(n)),
	A = arch_array(arch, 5.0 .* ones(n)), N = arch_array(arch, 0.01 .* ones(n)), C = arch_array(arch, 0.18 .* ones(n)),

Improve Eady figure #145

Improve Eady figure #145

Conversation

jagoosw commented Sep 18, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

navidcy Sep 19, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

navidcy left a comment

Choose a reason for hiding this comment

jagoosw commented Sep 24, 2023

jagoosw commented Sep 24, 2023

jagoosw commented Sep 24, 2023

jagoosw commented Sep 24, 2023

navidcy commented Sep 24, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jagoosw commented Sep 24, 2023

navidcy commented Sep 24, 2023

jagoosw commented Sep 24, 2023 • edited Loading

navidcy Sep 19, 2023 •

edited

Loading

jagoosw commented Sep 24, 2023 •

edited

Loading