2-ibm-logo.8o

# Disassembly of the famous "IBM logo" program, published here before:
# https://github.com/Timendus/chip-8/blob/master/octo/IBM%20logo.8o

# Annotated and converted to Octo mnemonics by Timendus, in the hope that it
# will be useful to people trying to debug their CHIP-8 interpreters
#
# Original "IBM logo" MD5 hash:       2dbace8066709ac9a264d23281820d32
# MD5 hash of binary from this code:  2dbace8066709ac9a264d23281820d32
#
# Unfortunately adding the version number to the image data changed the MD5
# hash of the final binary, but functionally absolutely nothing changed.

: main
  clear                       # Address 512 / 0x200

  i := ibm-0-0                # Address 514 / 0x202
  v0 := 12                    # Address 516 / 0x204
  v1 := 8
  sprite v0 v1 15             # Address 520 / 0x208

  v0 += 9                     # Address 522 / 0x20A
  i := ibm-1-0
  sprite v0 v1 15

  i := ibm-2-0                # Interesting mixup here, swapping the operations
  v0 += 8
  sprite v0 v1 15

  v0 += 4
  i := ibm-3-0
  sprite v0 v1 15

  v0 += 8
  i := ibm-4-0
  sprite v0 v1 15

  v0 += 8
  i := ibm-5-0
  sprite v0 v1 15

  loop again                  # Address 552 / 0x228

:segment data

# What's interesting about this image data is that the author could have
# precisely stored the full image in 5 by 15 bytes. Yet they have chosen a way
# to pack the image which needs 6 by 15 bytes. The most obvious explanation for
# this that I can come up with is that they approached the challenge one letter
# at a time. Maybe it was easier to do the conversions from bits to hexadecimal
# this way.

:include "../../pictures/ibm.png" 8x15