Issue Matrix multiplication in Directx vs OpenGL.

[11110]

This worked in DirectX & OpenGL platform. But it repeat the same calculate every vertex, which is not efficient.

//Shader
float4x4 matWVP = mul(World, mul(View, Proj));
o.position = mul(i.position, matWVP);

But when i precalculate on CPU and pass into GPU, the rendering become weird.

_effect.Parameters[WVP].SetValue(worldviewprojection);
shadercode

//Shader
o.position = mul(i.position, WVP);

[Charles_Humphrey]

Pass the WVP in rather than doing the calc in the shader.

I think that OpenGL and DX use different multiplaction order, that is to say one multiples a matrix by column, the other by row, switching the multiplication order may help??

[willmotil]

@11110

But when i precalculate on CPU and pass into GPU, the rendering become weird.

_effect.Parameters[WVP].SetValue(worldviewprojection);
shadercode

//Shader
o.position = mul(i.position, WVP);

try transposing on the shader though im not to confident that is the problem.

WVP = transpose(WVP);
o.position = mul(i.position,WVP);

Does sound similar to a problem i had here.

[SOLVED] Can't get this instanced shader to work right. Transpose

VSOutputInstanced MainVsInstanced(in VSVertexInputInstanced vertexInput, VSInstanceInputInstanced instanceInput) { VSOutputInstanced output; //vertexInput.Position is position of base vertex, multiple with instance transform matrix to get correct position // effectparameter setvalue do column->row for matrix but this one not >>> transpose(instanceInput.Transform) float4 wpos = mul(vertexInput.Position, transpose(instanceInput.Transform));/…

[11110]

Thanks for reply! it’s work, i did some research on it. I found this.

glsl vector*matrix different to hlsl