arm neon BSL instruction for floating point

Question

Is there an inline asm instruction equivalent for neon intrinsic vbslq_f32 ?

In intrinsic manual there is written BSL Vd.16B,Vn.16B,Vm.16B as asm equivalent for vbslq_f32 . But in asm manual there is not implementation for 4S, there is only "8B or 16B (though an assembler should accept any valid format)" and it doesn't work with floats.

Answer 1

You don't need any data type for bitwise operations.

Only the vector length matters, either 64bit or 128bit which you specify by 8b or 16b.