• Advertisement

Shnoutz

Member
  • Content count

    201
  • Joined

  • Last visited

Community Reputation

356 Neutral

About Shnoutz

  • Rank
    Member

Personal Information

  • Interests
    Programming
  1. Algorithm Tetrahedralization

    I found this: https://www.geometrictools.com/GTEngine/Include/Mathematics/GteDelaunay3.h Geometric tools FTW! This is exactly what I needed.
  2. Hi, I am looking for an algorithm that takes as input a point cloud and generates a list tetrahedrons (Delaunay?). I would like to use said tetrahedrons as a mean to interpolate the point cloud values at given points in space. Can anyone point me toward some resources? Cheers!
  3. Hi, quick question.. Why is the type of "fooBar" ambiguous? template< typename T > struct Bar { }; template< typename T1, typename T2 > struct Foo; template< typename T > struct Foo< T, Bar< T > > { }; template< typename T > struct Foo< float, Bar< T > > { }; Foo< float, Bar< float > > fooBar; I would expect that the second version of Foo is more specialized than the first... (I am building a compiler and I would like to understand the template specialization selection algorithm) Cheers!
  4. Thanks, interresting stuff :)
  5. I think the first step would be to define the components of the language as generic as possible.   What the most basic operations that are performed in a frame? Clear, copy, dispatch, draw (and present maybe). These operations have inputs and outputs, figuring out dependencies is quite easy.   Granted the shader code is required to figure out if a texture is used as a UAV or as a SRV but really, adding a pseudo shading language is not too scary I've done something similar in the past (maybe not the first thing I'll do ;) ). Still that would not be enough to know if, for example, a buffer was fully or partially written to but there might be a way to express sub resource regions.   I'm starting to think that the basic operations are like instructions and the shader code is like micro-instructions. With "instructions" and dependencies we can do lots of cool stuff like re-ordering and eliminating duplication.   I have started to work on a small prototype that takes code as input and spit out a graph with dependencies as output... I think its the first step. But its easy to imagine where that could lead... Automatic barriers, operation reordering, automatic async compute (given a descriptions of the hardware queues), automatic volatile resource allocation/aliasing, descriptor management.   That's a job for a full team of professional and for quite a while but ill try just for fun.
  6. I was thinking of a programming language because it is essentially a compact way to express a tree of resources and operations. I also like the idea of generating code rather than evaluating a tree like structure at run-time. (I do convert a lot of my code from a script like language to c++) Too bad I wont be attending GDC this year I would have loved to see that presentation.
  7. Hi,   I am a professional graphics programmer and I create scripting/programming languages as an hobby. I got this idea that I wanted to share with you.   I am reading and learning about low level graphics APIs and the reason why they exist. In DirectX11/OpenGL a lot of the GPU work, like resource barriers for example, is hidden and executed by the driver. Now, because the driver doesn't know what your frame looks like it has to execute the worst case scenario and execute more barriers that may be required. (I think DX11 drivers now are quite clever and do prediction to reduce that problem but you get my point)   DX12/Vulkan somewhat solves this issue by letting the programmer decide where to execute the barriers by exposing them as an API concept. That is a major plus but it is very error prone and if not done correctly can lead to major performance issues.   Now this got me thinking... What if we created a programming language that allowed to define explicitly what a full frame looks like. The steps and the resources involved in those steps. We could then look at theses steps and figure out exactly where to put the barriers. Re-order the steps for optimal performance. We could also look at what are the dependencies between steps and probably figure out a way to automatically dispatch the work on different queues (copy/dma, compute & graphics).   I have the feeling that with new low level APIs this door is now opened. Static analysis and optimization (of full frames)... Something every compilers do for CPU code. Why not GPU code?   Any thoughts on that?   Gab.
  8. I have a question about that.   I used to sort my passes (in a tree like structure) by root signatures then pipeline states then by resources all in one big command list.   Since then, I started looking at async compute to do culling of a pass while another is rendering.   Because fences are not part of the command list interface but rather on the queue I had to break my big command list into smaller ones (one per pass). Root signatures are undefined at the beginning of a command list so I have to re-set it for each pass' command list (even if I know they are the same for many consecutive pass). Is there still a point to sorting by root signature in this context? Or simply sort by pipelinestate+rootsignature like hodgman seem to suggest?
  9. Fence flags

    Hi!   I can kind of guess what D3D12_FENCE_FLAG_SHARED_CROSS_ADAPTER is for but what is the use of D3D12_FENCE_FLAG_SHARED?   Is it used for fence shared across different queues?   Thanks.
  10. D3D12CreateDevice fails

    Thank you for looking into this :)
  11. D3D12CreateDevice fails

      :wacko: I am kinda sure the code for the swap chain is fine. It works on amd & warp without any validation error/warning. It's also very straightforward. I was assuming GPU validation would not completely crash the application if there was something wrong. The message I get before the crash seems wrong, it basically complains that a resource is not in the right state but ask for the resource to be in that exact state (present 0x0). Disabling GPU validation works, the app is behaving exactly as expected. I am not excluding the possibility I messed up, but I gpu validation is a relatively new feature and I have the feeling my experience with it was worth reporting.
  12. D3D12CreateDevice fails

    Yeah, I have been lazy and using the first adapter, I will take some time and enumerate the adapters...   Back to GPU validation, I get this message on my laptop and I am not sure what it means:   (at the "Present" call) IGIESW ***.exe found in whitelist: NOIGIWHW Game ***.exe found in whitelist: NOD3D12 ERROR: GPU-BASED VALIDATION: Present, Back Buffer state invalid, Incompatible resource state: Resource: 0x0000025FF8FD0A50:'swapchain buffer', Subresource Index: [0], Resource State: D3D12_RESOURCE_STATE_[COMMON|PRESENT](0x0), Required State Bits: D3D12_RESOURCE_STATE_[COMMON|PRESENT](0x0), Draw Count [0], Dispatch Count [0], Command List: <deleted>, Resources used in COPY command lists must start out in the D3D12_RESOURCE_STATE_COMMON state.  This includes Resources created in a COPY_SOURCE or COPY_DEST state.  [ EXECUTION ERROR #942: GPU_BASED_VALIDATION_INCOMPATIBLE_RESOURCE_STATE]   If I understand correctly, 'swapchain buffer' is in the state "D3D12_RESOURCE_STATE_[COMMON|PRESENT](0x0)" but should be in the "state "D3D12_RESOURCE_STATE_[COMMON|PRESENT](0x0)"  o.O ??
  13. D3D12CreateDevice fails

    Good point. Ill check when I get off work, It tends to revert back to the intel gpu after a driver update.
  14. D3D12CreateDevice fails

    Just a quick question, is it possible that by updating windows it changes the supported level of features and resource binding tiers supported by a given gpu? I have a gtx980m and since last windows update it reports feature level 11.1 and resource binding tier 1. I am sure this was at least feature level 12 and resource binding tier 2.   My app is not working anymore on my laptop :(
  15. D3D12CreateDevice fails

    In my case it was a missing windows update.   Now, GPU based validation works like a charm on my desktop with an AMD rx480 but crashes on my notebook with a GTX980m. (The cash happens when I try to create a descriptor heap). 
  • Advertisement