Uniform buffer array alignment in HLSL vs SPIR-V

Sebastian Ahlman · 2017-10-07T19:54:36

Hi! I have the following HLSL shader code: struct Light { float4 positionWS; //-------------------------- ( 16 bytes ) float4 directionWS; //-------------------------- ( 16 bytes ) float4 positionVS; //-------------------------- ( 16 bytes ) float4 directionVS; //-------------------------- ( 16 bytes ) float4 color; //-------------------------- ( 16 bytes ) float spotlightAngle; float range; float intensity; uint type; //-------------------------- ( 16 bytes ) bool enabled; float3 padding1; //-------------------------- ( 16 bytes ) //-------------------------- ( 16 * 7 = 112 bytes ) }; layout(set=0,binding=0) cbuffer cbPerRenderOperationData : register(b0) { Light gLights[MAX_LIGHT_COUNT]; }; MAX_LIGHT_COUNT is 16. On DX11/12 the code gets compiled to D3D bytecode, and on Vulkan I compile it to SPIR-V using the HLSL frontend for glslang. On DX everything works fine. Looking at the buffer in RenderDoc shows me that it occupies 112 * 16 = 1792 bytes as I would expect. However, on Vulkan only the first light is valid. The rest are garbage. RenderDoc shows that the uniform buffer occupies 2048 bytes rather than 1792, which suggests to me that the size of Light in SPIR-V is 128 and not 112. Is there some power-of-two requirement in SPIR-V I am not aware of? So far I have just made sure every element in an array is a multiple of 16 bytes. Is that not enough? Thanks!

Graphics and GPU Programming Programming

Started by GuyWithBeard October 07, 2017 06:17 AM

12 comments, last by GuyWithBeard 7 years, 4 months ago

JoeJ

4,406

October 07, 2017 06:49 PM

So using

bool enabled;

int padd1; int padd2; int padd3;

instead would fix it? Makes sense! And it isn't that bad as i remember it.

Jihodg

1,160

October 07, 2017 07:00 PM

yep!... it is quite simple actually... you just have to think that the padding goes before the given type to produce the proper alignment of the variable or struct, instead of going after it as a filling!... float is aligned to 4 bytes, float2 aligned to 8 bytes, while both float3 and float4 aligned to 16 bytes

GuyWithBeard

Author

1,932

October 07, 2017 07:54 PM

I see, never would have guessed float3 and float4 would have the same requirement, but it makes sense.

Quote
...you just have to think that the padding goes before the given type...

That's pure gold! Thanks man!

Uniform buffer array alignment in HLSL vs SPIR-V

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Uniform buffer array alignment in HLSL vs SPIR-V

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines