(Vec4 x Mat4x4)使用SIMD和改进的产品

I am writing a complex simulation program and it apprears that the most time consumming routine is the one for multiplying a four-vector (float4) with a 4x4 matrix. I need to run this program on several computers, which are more or less old. That is why I tried to check SIMD capabilities of such operations in the following code :I am writing a complex simulation program and i