• Advertisement
Sign in to follow this  

Optimization Android/iOS GPU/System Debugging

Recommended Posts

(Posted this in graphics forum too, which was perhaps the wrong forum for it)

Hey, I was wondering if on mobile development (Android mainly but iOS as well if you know of it), if there is a GPUView equivalent for whole system debugging so we can figure out if the CPU/GPU are being pipelined efficiently, if there are bubbles, etc. Also slightly tangent question, but do mobile GPU's have a DMA engine exposed as a dedicated Transfer Queue for Vulkan?

Thanks!

Share this post


Link to post
Share on other sites
Advertisement
22 minutes ago, hiya83 said:

GPUView equivalent for whole system debugging so we can figure out if the CPU/GPU are being pipelined efficiently

You should have some kind of profiler. For IOS you would use Xcode for this. you need it anyway to build your game.

The way it works is you build a debug game and plug you mobile into your Mac. Then you run the game on your mobile while still plugged in, this should then give you lots of feedback on what is happening.

For Android you use the SDK that you need to build APK files, it has the tools with it.

 

If your using a engine like Unity or Unreal you will find they have profiling tools in the engine, this is what I always use.

30 minutes ago, hiya83 said:

Also slightly tangent question, but do mobile GPU's have a DMA engine exposed as a dedicated Transfer Queue for Vulkan?

I know these words but not in the order you are presenting them. In layman terms?

Share this post


Link to post
Share on other sites
1 hour ago, hiya83 said:

Hey, I was wondering if on mobile development (Android mainly but iOS as well if you know of it), if there is a GPUView equivalent for whole system debugging so we can figure out if the CPU/GPU are being pipelined efficiently,

Snapdragon Profiler if you're optimizing for a device using the Snapdragon SoC (most likely).

 

Share this post


Link to post
Share on other sites
1 hour ago, Scouting Ninja said:

I know these words but not in the order you are presenting them. In layman terms?

Vulkan exposes concept of command queues which are abstraction of different engines on the GPU. On the PC, a copy/transfer queue is basically a controller for the DMA engine that you can control independent of the graphics/compute engine. Was wondering if it works the same on mobile. 

1 hour ago, khawk said:

Snapdragon Profiler if you're optimizing for a device using the Snapdragon SoC (most likely).

That looks like what I am looking for, I'll give it a shot. Is there something equivalent for iOS?

Thanks!

Edited by hiya83

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this  

  • Advertisement
  • Advertisement
  • Popular Tags

  • Advertisement
  • Popular Now

  • Similar Content

    • By Cat's machete
      I need to pass 24 vec3, to a shader
      However glUniform3fv requires an array of GLfloat,
      Since my vec3 structure looks like:
      struct vec3{float x; float y; float z;};
      Can i just pass that safely, to float parray[24 * 3] using memcpy?
      Dont bother about GLfloat and float sizes i just gave pseudocode i just need to know if sent array will have first vertex at position 0 second one will be in pos 3, thrid one in 6 and so on, cause im not sure when even float and GLfloat match the sizes i could get some extra bytes anywhere,
       
      And another question is how then i define an uniform in shader?
      uniform vec3 box[24];. ?
       
      Cheers
    • By turanszkij
      Hi, right now building my engine in visual studio involves a shader compiling step to build hlsl 5.0 shaders. I have a separate project which only includes shader sources and the compiler is the visual studio integrated fxc compiler. I like this method because on any PC that has visual studio installed, I can just download the solution from GitHub and everything just builds without additional dependencies and using the latest version of the compiler. I also like it because the shaders are included in the solution explorer and easy to browse, and double-click to open (opening files can be really a pain in the ass in visual studio run in admin mode). Also it's nice that VS displays the build output/errors in the output window.
      But now I have the HLSL 6 compiler and want to build hlsl 6 shaders as well (and as I understand I can also compile vulkan compatible shaders with it later). Any idea how to do this nicely? I want only a single project containing shader sources, like it is now, but build them for different targets. I guess adding different building projects would be the way to go that reference the shader source project? But how would they differentiate from shader type of the sources (eg. pixel shader, compute shader,etc.)? Now the shader building project contains for each shader the shader type, how can other building projects reference that?
      Anyone with some experience in this?
    • By EddieK
      Hi I am having this problem where I am drawing 4000 squares on screen, using VBO's and IBO's but the framerate on my Huawei P9 is only 24 FPS. Considering it has 8-core CPU and a pretty powerful GPU, I don't think it is not capable of drawing 4000 textured squares at 60FPS.
      I checked the DMMS and found out that most of the time spent was by the put() method of the FloatBuffer, but the strange thing is that if I'm drawing these squares outside of the view frustum, the FPS increases. And I'm not using frustum culling. 
      If you have any ideas what could be causing this, please share them with me. Thank you in advance.
    • By EddieK
      Hi, so I am trying to implement packed VBO's with indexing on OpenGL but I have run across problems. It worked fine when I had separate buffers for vertex positions, colors and texture coordinates. But when I tried to put everything into a single packed buffer, it completely glitched out. Here's the code which I am using:
      this.vertexData.position(0); this.indexData.position(0); int stride = (3 + 4 + 2) * 4; GLES20.glBindBuffer(GLES20.GL_ARRAY_BUFFER, buffers[0]); GLES20.glBufferData(GLES20.GL_ARRAY_BUFFER, vertexData.capacity()*4, vertexData, GLES20.GL_STATIC_DRAW); ShaderAttributes attributes = graphicsSystem.getShader().getAttributes(); GLES20.glEnableVertexAttribArray(positionAttrID); GLES20.glVertexAttribPointer(positionAttrID, dimensions, GLES20.GL_FLOAT, false, stride, 0); GLES20.glEnableVertexAttribArray(colorAttrID); GLES20.glVertexAttribPointer(colorAttrID, 4, GLES20.GL_FLOAT, false, stride, dimensions * 4); GLES20.glEnableVertexAttribArray(texCoordAttrID); GLES20.glVertexAttribPointer(texCoordAttrID, 2, GLES20.GL_FLOAT, false, stride, (dimensions + 4) * 4); GLES20.glBindBuffer(GLES20.GL_ELEMENT_ARRAY_BUFFER, buffers[3]); GLES20.glBufferData(GLES20.GL_ELEMENT_ARRAY_BUFFER, indexData.capacity()*2, indexData, GLES20.GL_STATIC_DRAW); GLES20.glDrawElements(mode, count, GLES20.GL_UNSIGNED_SHORT, 0); The data in vertex buffer is ordered like this:
      Vertex X, vertex Y, vertex Z, Color r, color g, color b, color a, Tex coord x, tex coord z and so on... (And I am pretty certain that the buffer I'm using is in this order)
      This is the version of the code which worked fine:
      this.vertexData.position(0); this.vertexColorData.position(0); this.vertexTexCoordData.position(0); this.indexData.position(0); GLES20.glBindBuffer(GLES20.GL_ARRAY_BUFFER, buffers[0]); GLES20.glBufferData(GLES20.GL_ARRAY_BUFFER, vertexPositionData.capacity()*4, vertexPositionData, GLES20.GL_STATIC_DRAW); ShaderAttributes attributes = graphicsSystem.getShader().getAttributes(); GLES20.glEnableVertexAttribArray(positionAttrID); GLES20.glVertexAttribPointer(positionAttrID, 4, GLES20.GL_FLOAT, false, 0, 0); GLES20.glBindBuffer(GLES20.GL_ARRAY_BUFFER, buffers[1]); GLES20.glBufferData(GLES20.GL_ARRAY_BUFFER, vertexColorData.capacity()*4, vertexColorData, GLES20.GL_STATIC_DRAW); GLES20.glEnableVertexAttribArray(colorAttrID); GLES20.glVertexAttribPointer(colorAttrID, 4, GLES20.GL_FLOAT, false, 0, 0); GLES20.glBindBuffer(GLES20.GL_ARRAY_BUFFER, buffers[2]); GLES20.glBufferData(GLES20.GL_ARRAY_BUFFER, vertexTexCoordData.capacity()*4, vertexTexCoordData, GLES20.GL_STATIC_DRAW); GLES20.glEnableVertexAttribArray(textCoordAttrID); GLES20.glVertexAttribPointer(textCoordAttrID, 4, GLES20.GL_FLOAT, false, 0, 0); */ GLES20.glBindBuffer(GLES20.GL_ELEMENT_ARRAY_BUFFER, buffers[3]); GLES20.glBufferData(GLES20.GL_ELEMENT_ARRAY_BUFFER, indexData.capacity()*2, indexData, GLES20.GL_STATIC_DRAW); GLES20.glDrawElements(mode, count, GLES20.GL_UNSIGNED_SHORT, 0); This is the output of the non working code:

      From this picture I can see that some of the vertex positions are good, but for some reason every renderable object from the game has a at least one vertex position of value 0
      Thank in advance,
      Ed
    • By mark_braga
      I am working on a compute shader in Vulkan which does some image processing and has 1024 * 5=5120 loop iterations (5 outer and 1024 inner)
      If I do this, I get a device lost error after the succeeding call to queueSubmit after the image processing queueSubmit
      // Image processing dispatch submit(); waitForFence(); // All calls to submit after this will give the device lost error If I lower the number of loops from 1024 to 256 => 5 * 256 = 1280 loop iterations, it works fine. The shader does some pretty heavy arithmetic operations but the number of resources bound is 3 (one SRV, one UAV, and one sampler). The thread group size is x=16 ,y=16,z=1
      So my question - Is there a hardware limit to the number of loop executions/number of instructions per shader?
  • Advertisement