Assembler for vector calculus

The assembler calculator is enhanced with vector computing.

According to the ARM processor specifications, three operation modes are available:

• scalars
• vectors
• scalars and vectors mixed (our choice)

How do we program this :

• using a vector lenght of 4, the control register FPSCR is loaded with ( 4 – 1 ) × ²↑16
• the registers s8, s9, s10 and s11 are loaded with the first vector operand [a1 a2 a3 a4]
• if the second operand is a vector then the registers s20, s21, s22 and s23 are loaded with [b1 b2 b3 b4] ;
  else, if the operand is a scalar then the register s0 must be used and loaded with b
• the result of the operation in the registers s8, s9, s10 and s11 is stored into [c1 c2 c3 c4]