Although the basic word size of the EDSAC was 18 bits, only 17 bits could actually be used due to circuit set-up times. The opcode (i.e. one of the instruction codes explained below) was specified in 5 bits, and the address in 10 bits (giving an address range of 1024 short words). The length bit specified the operand length : most instructions could operate on either a 17-bit short word (length = 0), or a 35-bit full word (length = 1).
The EDSAC Instruction Set 1949
The were four number formats which could be used in the EDSAC: short and long integers, and short and long fractions.
Short numbers were 17-bits in length (for reasons explained above) and long ones 35-bits.
Whereas the multiplier and multiplicand registers each had a capacity of 35 bits, the accumulator provided 71 bits. This means that long numbers can be multiplied without overflow.
|Editor Window||Output Window||Binary||Decimal|
|Pi (type in as "#")||Figure shift||01011||11|
|Erase (type in as "*")||Letter shift||01111||15|
|Blank Tape (type in as "." )||(no effect)||10000||16|
|Theta (type in as "@")||Carriage return||10010||18|
|Phi (type in as "!")||Space||10100||20|
|Delta (type in as "&")||Line feed||11000||24|