Re: [PATCH vkd3d v4 2/6] tests: Test a number of simple HLSL operations.

[Giovanni Mascellani]

Hi,

Il 28/01/22 15:53, Francisco Casas ha scritto:
> Probably because float doesn't have enough precision to represent 4294967300.

Precisely. Technically the native compiler is a bit more correct than 
what I wrote, because it's true that the actually representable value is 
4294967296, not 4294967300. But they both map to the same float value.

A useful tool to experiment with this stuff is 
https://evanw.github.io/float-toy/.

Giovanni.