[0035] linalg::Convert calculates incorrect DstN value

[NealLou]

Which document does this relate to?

hlsl-specs/proposals/0035-linalg-matrix.md

Lines 717 to 722 in 4412bb3

	template <ComponentEnum DstTy, ComponentEnum SrcTy, int SrcN> struct DstN {
	static const int Value =
	(SrcN * ComponentTypeTraits<SrcTy>::ElementsPerScalar) /
	ComponentTypeTraits<DstTy>::ElementsPerScalar;
	};

Describe the issue you see with the spec
linalg::Convert calculates incorrect DstN value, the current formula performs integer division with truncation, but the result needs to be rounded up.
e.g. Convert F16->Fp8, SrcN=3, result in a DstN=0.

Additional context
DXC has fix that microsoft/DirectXShaderCompiler#8359, spec needs to be updated.

[llvm-beanz]

Resolved in #855