# CuArray + GLMakie

**URL:** <https://discourse.julialang.org/t/cuarray-glmakie/52461>\
**Category:** Visualization\
**Tags:** plotting\
**Created:** [December 27, 2020, 3:37pm UTC](https://discourse.julialang.org/t/cuarray-glmakie/52461 "2020-12-27T15:37:13Z")\
**Posts on this page:** 1\
**Showing post:** 4

<div class="post-metadata">

**Author:** ![LaurentPlagne](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/laurentplagne/32/10103_2.png) [@LaurentPlagne](https://discourse.julialang.org/u/LaurentPlagne)\
**Post date:** [December 28, 2020, 4:41pm UTC](https://discourse.julialang.org/t/cuarray-glmakie/52461/4 "2020-12-28T16:41:37Z")

</div>

Hi Simon,  
Here is an example which is unfortunately a bit too long for the 7-lines thread…  
Anyway, It takes about 5s on my machine to run and show a rather convincing combination of `CUDA.jl` and `Makie.jl`.  
In the plot (can be `heatmap` or `surface`), I copy the `CuArray zs` to a standard `Array czs`…

![toto](https://global.discourse-cdn.com/julialang/original/3X/7/6/76ae1d2c297b7968245a3435f83275cf3feb8bd2.gif)

```julia
using CUDA,GLMakie,AbstractPlotting
const T=Float32
const V=CuArray{T,2}
function go()
    CUDA.allowscalar(false)
    ls,ns,ms,λ,dt,sr,k,nt,two,eight,fr =1.e-2,800,1.e-4,0.1,2.e-7,0.1,5e7,5000,T(2),T(8),10
    dt2sm=dt^2/ms
    r,r0,r1,r2=1:ns,1:ns-2,2:ns-1,3:ns
    cxs = [T(j-1)*ls for i in r, j in r]
    dx(i,j) = cxs[i,j]-cxs[ns÷2,ns÷2]
    cxc = [cxs[i,j]+sr*(dx(i,j))*exp(-0.5(dx(i,j)^2+dx(j,i)^2)/λ^2)/λ for i in r, j in r]
    xs,xc,ys,yc=V(cxs),V(cxc),V(cxs'),V(cxc')
    # ys,yc=collect(xs'),collect(xc')
    xp,yp,xt,yt,fx,fy,zs=copy(xc),copy(yc),copy(xc),copy(yc),zero(xc),zero(xc),zero(xc)
    @. zs = sqrt((xc-xs)^2+(yc-ys)^2) 
    czs=Array(zs) ; zn=Node(czs) ; zl=sr*0.05
    scene = Scene(resolution = (400,400)) ; scale!(scene, 1, 1, 10000)
    heatmap!(scene,1:ns,1:ns,lift(z->z,zn),colorrange = (-zl,zl))
    # surface!(scene,1:ns,1:ns,lift(z->z,zn),colorrange = (-zl,zl))
    zlims!((-zl,zl))
    display(scene)
    GLMakie.record(scene, "output.gif", 1:nt÷fr, framerate=30) do j
    # for j in 1:nt÷fr
        for i in 1:fr
            @views @. fx[r1,r1] = -k*(eight*xc[r1,r1]-xc[r0,r0]-xc[r1,r0]-xc[r2,r0]-xc[r0,r1]-xc[r2,r1]-xc[r0,r2]-xc[r1,r2]-xc[r2,r2])
            @views @. fy[r1,r1] = -k*(eight*yc[r1,r1]-yc[r0,r0]-yc[r1,r0]-yc[r2,r0]-yc[r0,r1]-yc[r2,r1]-yc[r0,r2]-yc[r1,r2]-yc[r2,r2])
            @. xt = two*xc-xp+fx*dt2sm
            @. yt = two*yc-yp+fy*dt2sm
            xc,xp,xt = xt,xc,xp
            yc,yp,yt = yt,yc,yp
        end
        @. zs = sqrt((xc-xs)^2+(yc-ys)^2)
        czs=Array(zs)
        zn[] = czs
        yield()
    end
end
@time go()

```

![toto](https://global.discourse-cdn.com/julialang/original/3X/1/c/1c0a8283dd2e18237210ae53fd6b400de8cffaa6.gif)

---

_[View the full topic](https://discourse.julialang.org/t/cuarray-glmakie/52461)._