This type defines move code instructions with methods for encoding, decoding, string conversion, and parsing from strings and byte arrays. For the encoding into machine words (byte arrays), see the addresses of PUs and buffers in the SCAD reference sheet. Hence, the general purpose PUs have indices \(2,\ldots,\mathsf{numPU}+1\), and their input and output buffers have addresses \(4,\ldots,2\cdot\mathsf{numPU}+3\). If these numbers are encoded as radix-2 numbers, we need \(\left\lceil{\log_2(2\cdot\mathsf{numPU}+4)}\right\rceil\) bits for the source address and the same number of bits for the target address. Together with the two bits to determine one of the four kinds of move instructions, we finally obtain instructions |mv|srcAdr|tgtAdr| where mv has two bits, and each one of srcAdr and tgtAdr has \(\left\lceil{\log_2(2\cdot\mathsf{numPU}+4)}\right\rceil\) bits. Hence, we get

• \(\mathsf{numBits(numPU)} := 2+2\cdot \left\lceil{\log_2(2\cdot\mathsf{numPU}+4)}\right\rceil\)
• \(\mathsf{maxPU(numBits)} := 2^{\frac{\mathsf{numBits}}{2}-2}-2\)