Jump to content
  • Advertisement


  • Content Count

  • Joined

  • Last visited

Community Reputation

353 Neutral

About medigames

  • Rank
  1. medigames

    bullet Box penetration

    yeap, that was my bad, the problem was a quaternion mat3 conversion. thank you all, 
  2. medigames

    bullet Box penetration

    Hi this is the first time i'm using bullet physics, and i have encounter a problem that i don't know if it's known in bullet library. the problem is that when boxes rest, they penetrate each other. im using the default btSequentialImpulseConstraintSolver. this is an image of what i mean.   this is how i get the world transform btTransform bttransform; aBtRigidBody[uIdx].getMotionState()->getWorldTransform(bttransform);
  3. medigames

    Convex Hull computation

    Dirk Gregorius i did see your presentation, only the 3d part was a bit hard for me to grasp completely, however i will be looking into the JAVA implementation. i would like to take this chance to thank you for your great presentations.
  4. Hi, for the last month i have been trying to implement SAT into a physics engine, the best and the cleanest way i could v found was the Gregorius Dirk's from GDC 2013, and i have been trying since then to get a convex hull with a half edge data structure to work, the thing that proves to be hard for me, iv tried the quick hull, but i couldn't get to implement it, please, if there is any implementation or an article with a clear algorithm of the former your aware of, share it with us, thank you,
  5. medigames

    Software Rasterizer

    i don't think so, the store implicitly changes the order, and i'v tried it but it didn't work.   EDIT: Actually that this solve those edges, but now introduces interiors
  6. Hello every one, i'm having a bit of a trouble tracking a bug, i'm getting weird edges at the end of a triangle only when using SSE, the output is just fine with scalars, here is an image tothe problem, and that's the code, it basically performs a point triangle intersection test based on fgiesen blog series for (p.y = iMinY; p.y <= iMaxY; p.y += SEdge::s_iYStepSize, iIdx += iDepthBufferWidth) { Simd::I32x4 i4w0 = i4W0Row; Simd::I32x4 i4w1 = i4W1Row; Simd::I32x4 i4w2 = i4W2Row; I32 iXIdx = iIdx; for (p.x = iMinX; p.x <= iMaxX; p.x += SEdge::s_iXStepSize, iXIdx += SEdge::s_iXStepSize, // one step to the right. i4w0 = Simd::AddPackedI32(i4w0, e12.i4StepX), i4w1 = Simd::AddPackedI32(i4w1, e20.i4StepX), i4w2 = Simd::AddPackedI32(i4w2, e01.i4StepX)) { const Simd::I32x4 i4Mask = _mm_cmplt_epi32(zero, Simd::OrPackedI32(i4w0, Simd::OrPackedI32(i4w1, i4w2))); if (_mm_test_all_zeros(i4Mask, i4Mask)) continue; Simd::F32x4 depth = z0; depth = Simd::AddPacked(depth, Simd::MulPacked(_mm_cvtepi32_ps(i4w1), z1)); depth = Simd::AddPacked(depth, Simd::MulPacked(_mm_cvtepi32_ps(i4w2), z2)); Simd::F32x4 prevDepth = _mm_load_ps(&pRenderTarget[iXIdx]); Simd::F32x4 depthMask = _mm_cmplt_ps(depth, prevDepth); Simd::I32x4 finalMask = _mm_and_si128(i4Mask, _mm_castps_si128(depthMask)); depth = Simd::BlendPacked(prevDepth, depth, _mm_castsi128_ps(finalMask)); Simd::StorePacked(&pRenderTarget[iXIdx], depth); } i4W0Row = Simd::AddPackedI32(i4W0Row, e12.i4StepY); i4W1Row = Simd::AddPackedI32(i4W1Row, e20.i4StepY); i4W2Row = Simd::AddPackedI32(i4W2Row, e01.i4StepY); }
  7. medigames

    The Framework Engine

    Hello every one, this is my first log on GameDev, so please tell me if i'm violating any rule. basically, i wanna share with you the state of an engine i'm currently developing. it's almost a year now, giving that i've started from the ground up, i think i did pretty great, and i've learned a lot of new stuff along the way, i've named it the Framework Engine, sounds like a stupid name, i know, just when i started it, Framework seem'd like a cool namespace . i've been trying lately to get people to notice it, but, it's just not getting much of an attention, i'm not trying to get people to use it though, at this stage i don't even have an editor, but, as an indie, it's seem nice to get feedback from the community. the main visual features are the following : Frustum & Occlusion culling:  Skeletal animation:  Tiled deferred rendering:  please, feel free to leave me your feedback.
  8. medigames

    Unity Singletons and Game Dev

    I don't like, nether use singletons, because they inforce a one object in the hole application, it could be fine for an audio system for exp, but, this restriction could get ugly if for some reason you tried to change that, how ever im giving access to high level system that will be alive as long as the application is running, such as the GraphicsDevice, JobManager, ScreenManager, etc, via a global Context variable, i found that it cleaner this way, instead of passing a pointer of the GraphicsDevice to every low level system that is intrested in it, and then i have to keep track of it's life, the low level systems how ever are decoupled as much i as i could them to be, and this is only my opnion.
  9. Alright, i did a find the root problem, in my engine i'm using a "layout(location = x) uniform" and  generate the locations based on the names of those uniforms based on a table, now, for some reason that i don't now, the shader cannot use a uniform buffer of such a size, so all i did was to remove that mechanism and let opengl decide the locations.
  10. Matias Goldberg i did tried that, and not working.   fleabay that's  right, but this is not the syntax i'm using my engine to declare a uniform buffer i have just write that here, what i use is "UniformBlock LightTilesBlock" instead of "layout (std140, binding = 4) uniform", thanks for pointing this out though.
  11. Hey every body, i'm in the middle of implementing tiled deferred rendering in the fragment shader, but i keep getting a gpu program link error like this: error(#468) Link error : the number of uniform symbols exceeds the maximum of system defined. this is the uniform buffer i'm using in the fragment shader. layout (std140, binding = 4) LightTilesBlock { uvec4 au4LightIndices[8*400]; } u_lightTiles; i manually calculate it's size and i found that it's 51200 bytes please correct me if i'm wrong. GL_MAX_UNIFORM_BLOCK_SIZE return the following 65536 and the same problem even with a vec4 not a problem with 1024 mat4 in the vertex shader.  
  • Advertisement

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

GameDev.net is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!