[Mark SiblyKeymaster]

Had a quick go at speeding this via render-to-pixmap and it’s not bad – was going 22FPS-ish, now does 60 if the mouse is at the edges!

A bit of hackery involved – it draws directly to a pixel ‘ptr’ instead of SetPixel( x,y ) so it’s not recalculating pixel address all the time.

Monkey

' ' ' Julia Set example by abakobo ' ' a monkey 2 example-learning exercise ' next steps/points of interest is to optimise: -the drawpoint speed ' -the choice for globals, fields, locals ' ' Of course the use of the imagearray should finally be skipped for better perfomances ' #Import "<std>" #Import "<mojo>" Using std.. Using mojo.. Global w_width:=800 'window size Global w_height:=600 Global MaxIt:=32 'Julia's calculation precision Global Threshold:=2 'Julia's escape Threshold Global zoom:=2.7 'inverted zoom factor Global Palette:UInt[] Class Julia Extends Window Field pixmap:=New Pixmap( w_width,w_height ) Field image:=New Image( w_width,w_height,TextureFlags.Dynamic ) Method New( title:String,width:Int,height:Int,flags:WindowFlags=WindowFlags.Resizable ) Super.New( title,width,height,WindowFlags.Center ) End Method OnRender( canvas:Canvas ) Override App.RequestRender() ' ' local vars declaration ' Local Cr:Float, Ci:Float, Zr:Float, Zi:Float ' the complex number parts (should probably use struct...) Local ta:Int,tb:Int,tj:Int 'the timer vars Local x:Int,y:Int 'coord iterators ' ' get the mouse coord and transform it to the complex constant used in Julia set ' Cr=1+(App.MouseLocation.x-w_width)/(w_width/2.0) Ci=1+(App.MouseLocation.y-w_height)/(w_height/2.0) ' ' Calculates Julia in indexed colors (iterations up to MaxIt-1 due to array starting at 0) ' and copy it to the imagearray array ' ta=Millisecs() For y=0 Until w_height Local p:=Cast<UInt Ptr>( pixmap.PixelPtr( 0,y ) ) Zi=(y-(w_height/2.0))*(zoom/w_height) For x=0 Until w_width Zr=(x-(w_width/2.0))*(zoom/w_width) Local t:=NbIter( Zr,Zi,MaxIt,Threshold,Cr,Ci ) p[x]=Palette[t] Next Next tb=Millisecs() tj=tb-ta 'Records the time it took to calculate Julia and copy it to the array image.Texture.PastePixmap( pixmap,0,0 ) canvas.DrawImage( image,0,0 ) ' Prints timers and FPS ' canvas.Color=Color.Black canvas.DrawText("Move mouse arroud to see various Julia sets",0,0) canvas.DrawText("Julia to array time: "+tj,0,15) canvas.DrawText("array to canvas time: "+tb,0,30) canvas.DrawText("FPS:"+App.FPS,0,45) End End ' ' squared complex+const series (Julia) ' Function NbIter:Int(Zr:Float,Zi:Float,max_iter:Int,exit_radius:Float,Cr:Float,Ci:Float) Local i:=0 Local Zr2:Float, Zi2:Float, er2:Float er2=exit_radius*exit_radius Zr2=Zr*Zr Zi2=Zi*Zi While i<max_iter-1 And Zr2+Zi2<er2 Zi=2*Zr*Zi + Ci Zr=Zr2-Zi2 +Cr Zr2=Zr*Zr Zi2=Zi*Zi i+=1 Wend Return i End Function ColorToBGRA:UInt( color:Color ) Return UInt(color.a*255) Shl 24 | UInt(color.b*255) Shl 16 | UInt(color.g*255) Shl 8 | UInt(color.r*255) End ' ' Create the indexed color Palette ' Function CreateGlobalPalette:Void() Palette=New UInt[MaxIt] 'Palette is Global, declared at top For Local i:=0 Until MaxIt-1 Palette[i]=ColorToBGRA( New Color(Abs(Sin(2.0*Pi*i/MaxIt)),Abs(Sin(1.2*Pi*i/MaxIt)),0.6-(i/(MaxIt*1.0))*0.3) ) Next Palette[MaxIt-1]=ColorToBGRA( Color.Black ) End Function Main() CreateGlobalPalette() New AppInstance New Julia( "Julia",w_width,w_height ) App.Run() End

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141

' ' ' Julia Set example by abakobo ' '     a monkey 2 example-learning exercise '     next steps/points of interest is to optimise: -the drawpoint speed '                                                   -the choice for globals, fields, locals ' '             Of course the use of the imagearray should finally be skipped for better perfomances ' #Import "<std>" #Import "<mojo>" Using std.. Using mojo.. Global w_width:=800 'window size Global w_height:=600 Global MaxIt:=32 'Julia's calculation precision Global Threshold:=2 'Julia's escape Threshold Global zoom:=2.7 'inverted zoom factor Global Palette:UInt[]   Class Julia Extends Window   Field pixmap:=New Pixmap( w_width,w_height ) Field image:=New Image( w_width,w_height,TextureFlags.Dynamic )   Method New( title:String,width:Int,height:Int,flags:WindowFlags=WindowFlags.Resizable ) Super.New( title,width,height,WindowFlags.Center ) End Method OnRender( canvas:Canvas ) Override App.RequestRender() ' '  local vars declaration ' Local Cr:Float, Ci:Float, Zr:Float, Zi:Float ' the complex number parts (should probably use struct...) Local ta:Int,tb:Int,tj:Int 'the timer vars Local x:Int,y:Int 'coord iterators               '     ' get the mouse coord and transform it to the complex constant used in Julia set     ' Cr=1+(App.MouseLocation.x-w_width)/(w_width/2.0) Ci=1+(App.MouseLocation.y-w_height)/(w_height/2.0)     '     ' Calculates Julia in indexed colors (iterations up to MaxIt-1 due to array starting at 0)     ' and copy it to the imagearray array     ' ta=Millisecs() For y=0 Until w_height   Local p:=Cast<UInt Ptr>( pixmap.PixelPtr( 0,y ) ) Zi=(y-(w_height/2.0))*(zoom/w_height) For x=0 Until w_width Zr=(x-(w_width/2.0))*(zoom/w_width) Local t:=NbIter( Zr,Zi,MaxIt,Threshold,Cr,Ci ) p[x]=Palette[t] Next Next tb=Millisecs() tj=tb-ta 'Records the time it took to calculate Julia and copy it to the array      image.Texture.PastePixmap( pixmap,0,0 ) canvas.DrawImage( image,0,0 )      ' Prints timers and FPS ' canvas.Color=Color.Black      canvas.DrawText("Move mouse arroud to see various Julia sets",0,0) canvas.DrawText("Julia to array time: "+tj,0,15) canvas.DrawText("array to canvas time: "+tb,0,30) canvas.DrawText("FPS:"+App.FPS,0,45) End End ' ' squared complex+const series (Julia) ' Function NbIter:Int(Zr:Float,Zi:Float,max_iter:Int,exit_radius:Float,Cr:Float,Ci:Float) Local i:=0 Local Zr2:Float, Zi2:Float, er2:Float er2=exit_radius*exit_radius Zr2=Zr*Zr Zi2=Zi*Zi While i<max_iter-1 And Zr2+Zi2<er2 Zi=2*Zr*Zi + Ci Zr=Zr2-Zi2 +Cr Zr2=Zr*Zr Zi2=Zi*Zi i+=1 Wend Return i End Function ColorToBGRA:UInt( color:Color ) Return UInt(color.a*255) Shl 24 | UInt(color.b*255) Shl 16 | UInt(color.g*255) Shl 8 | UInt(color.r*255) End ' ' Create the indexed color Palette ' Function CreateGlobalPalette:Void()   Palette=New UInt[MaxIt] 'Palette is Global, declared at top    For Local i:=0 Until MaxIt-1 Palette[i]=ColorToBGRA( New Color(Abs(Sin(2.0*Pi*i/MaxIt)),Abs(Sin(1.2*Pi*i/MaxIt)),0.6-(i/(MaxIt*1.0))*0.3) ) Next   Palette[MaxIt-1]=ColorToBGRA( Color.Black )   End Function Main() CreateGlobalPalette() New AppInstance New Julia( "Julia",w_width,w_height ) App.Run() End

[Mark SiblyKeymaster]

> What I would really like to see is some kind of direct access to the audio buffer in mojo.

There is actually limited/experimental support for this in now – see my blog update, but note there are issues with doing realtime stuff like audio in a single threaded app.

There’s a little demo in the vsynth bananas folder called vsynth_mojo.monkey2. It basically does the same thing as vsynth – ie: realtime audio generation – only through a single openal audio channel, so you can mix it with stuff playing through other audio channels easily. I have no idea how well it runs on other machines though!

[Mark SiblyKeymaster]

Mojo uses the GLES20 API exclusively.

However, Windows builds use ‘angle’ to convert GLES20 to D3D9/11 calls.

You could theoretically switch to using native GLES20 on Windows, but I don’t think you’d gain much, if any, performance, and you’d lose a lot of compatiblity.

[Mark SiblyKeymaster]

Thanks, will fix, but please post bugs in the ‘issues’ thing at github in future, or I may miss them.

[Mark SiblyKeymaster]

This turned out ot be a ‘spaces in filename’ build issue, but was kind of fun to find.

The main problem was that mx2cc wasn’t recognizing a file had changed so wasn’t rebuilding it.

But each time I tried to add something to the monkey module to do some testing, this was enough to trigger a rebuild! So I ended up having to add a new foobar field after each modules update – I ended up with about 5 foobars before it clicked what was happening…

For a while it seemed like the ultimate heisenbug – each time I tried to measure it, it very predictably disappeared!

[Mark SiblyKeymaster]

Nope sorry.

I no longer even have a CD of Windows 7, although I guess I’m gonna need one sooner or later if there are problems with mx2 apps not running on it.

And to be honest, I am not really keen on supporting Windows 7 as a dev platform. It’s officially reached ‘end of mainstream support’ and more and more stuff is just gonna break on it over time.

There’s also the time factor for me – I’m already supporting 3 (soon to be 5) dev platforms and I just don’t like the idea of having to support another.

Yeah, it sucks if you don’t like Windows 10, but that’s just the way it is if you’re a developer. I’m kind of lucky – I hate all versions of Windows equally so changing up has never been a big deal.

Of course, if anyone can come up with fixes for Windows 7 that don’t break Windows 8/10  I’m happy to implement them – I could sure do with the help!

[Mark SiblyKeymaster]

Monkey2 currently only implements GLES2.0, which means you need to create shaders and so on before you can draw anything.

Also, you can’t yet render directly to OpenGL in mojo – coming soon though.

[Mark SiblyKeymaster]

Looking forward to next screenshot!

[Mark SiblyKeymaster]

As for lines (ie. DrawLine) they don’t appear affected by this change for glowy line style drawing.

I think they are – the intersections between the bright and dark blue lines appear to be slighty brighter.

But that may be an optical illusion! Another thing to remember is that blending results are clamped to [0,1]  so adding Color( 0,0,1 ) to Color( 0,0,.5 ) will still give you Color( 0,0,1 ).

I just hacked in a LineWidth too, now up at github, although it’s not very clever. It basically just draws a rotated rect if LineWidth>1 and, unless blendmode is Opaque, attempts to ‘smooth’ it. The ends of the line are ‘hard clipped’ so it looks more like a cylinder if LineWidth is too big, but it’s not bad for a quick hack IMO.

I really want to wait until shaders are going before getting too carried away with this stuff, but this should give people something to play with until that happens.

[Mark SiblyKeymaster]

What error do you get?

How do I reproduce the error?

Also, there’s a version of Deque coming to std soon..

[Mark SiblyKeymaster]

Ok, possible fix is now up at github.

Copy from console is indeed a little broken. Try ‘copy’ from the edit menu instead of ‘ctrl-c’, works here.

[Mark SiblyKeymaster]

Can you post the error?

I’m guessing it’s the 8K command line character limit, will fix ASAP.

[Mark SiblyKeymaster]

I should perhaps clarify a bit…

What I’d really like to add instead of #define is ‘StaticIf’.

This would be a bit like a normal If, except it can appear *anywhere* in a program, not just in a block of statements.

StaticIf expressions must be constant, but they can also include type comparisons which makes StaticIf also useful for kludging up generic code when necessary.

The big issue with #define is that (the way it’s implemented in bmx and mx1 anyway) it happens during parsing, so is sensitive to ‘order of imports’ issues, eg:

Monkey

#If TEST #Print "YES" #EndIf #TEST=1

1 2 3 4 5 6

#If TEST #Print "YES" #EndIf   #TEST=1

…will not print “YES” on bmx (I think) or mx1.

It’s obvious in this case, but when you have a #define in one file and a dependant #if in another (and potential cyclic file dependancies) it can all get a bit hairy.

‘StaticIf’ solves this because it happens on the semant phase, after parsing, so this:

Monkey

StaticIf Test #Print "YES" 'hypothetical #Print used here.. EndIf Const Test:=True

1 2 3 4 5

StaticIf Test #Print "YES"   'hypothetical #Print used here.. EndIf   Const Test:=True

…will actually print “YES”.

Another thing I like about StaticIf is that code that is not generated because the StaticIf fails is still actually parsed, whereas #If just skips the code altogether, even if it’s full of garbage.

Once StaticIf is in, the current __TARGET__, __HOSTOS__ etc symbols could also be implemented as simple Consts inside, say, a monkey.build namespace. This way, you’d be able to use them for both conditional compilation and directly in code.

Not sure when this’ll happen though – and if anyone can come up with a better name than StaticIf…

[Mark SiblyKeymaster]

You can’t yet…

[Mark SiblyKeymaster]

Try BlendMode.Additive.

[Author]

Posts