39     VLDMIA      r1!,{s0-s4}         // Segment and load to shorten stall times due to data hazards  in MTX33CopyAsm()
42     VSTMIA      r2!,{s0-s4}         // Store result  in MTX33CopyAsm()
53     VMLA.F32    s1,s10,s0  in MTX33MAddAsm()
54     VMLA.F32    s2,s11,s0  in MTX33MAddAsm()
55     VMLA.F32    s3,s12,s0  in MTX33MAddAsm()
59     VMLA.F32    s4,s13,s0  in MTX33MAddAsm()
60     VMLA.F32    s5,s14,s0  in MTX33MAddAsm()
61     VMLA.F32    s6,s15,s0  in MTX33MAddAsm()
63     VMLA.F32    s7,s10,s0  in MTX33MAddAsm()
64     VMLA.F32    s8,s11,s0  in MTX33MAddAsm()
65     VMLA.F32    s9,s12,s0  in MTX33MAddAsm()
75     VLDMIA      r1,{s0-s8}          // First line of matrix pM to register [S0-S2]  in VEC3TransformAsm()
78     VMUL.F32    s12,s0,s9  in VEC3TransformAsm()