Neon instruction, vsub_f32(a, b), is it a-b or b-a?

Question

In this neon instruction (from here ):

float32x2_t vsub_f32(float32x2_t a, float32x2_t b);  // VSUB.F32 d0,d0,d0

Does it return a - b or b - a? I cannot find it in the ARM documentation...

Answer 1

The VSUB instruction subtracts the value in the second operand register from the value in the first operand register, and places the result in the destination register.

from http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0489c/Bcffigdj.html

So, it´s ab

Answer 2

You may find the expanded documentation for the Neon Intrinsics more useful. It also gives the mapping from functions parameters to instruction operands, and is available from:

http://infocenter.arm.com/help/topic/com.arm.doc.ihi0073a/IHI0073A_arm_neon_intrinsics_ref.pdf

For vsub_f32 , that document gives the following:

float32x2_t vsub_f32 (float32x2_t a, float32x2_t b)
a → Vn.2S b → Vm.2S
FSUB Vd.2S,Vn.2S,Vm.2S