Uploaded March 2021 | Updated September 2026, 2 weeks ago

---

This video is sped up 128 times (the original took just over 35 minutes).
The CPU is calculating each number and then converting it to base 10 (BCD) then it prints out the values while clipping the leading zeros.

It only takes 12 cycles to calculate each pair of fib values but the reason the program is very slow is due to it taking 20 cycles to convert the 32 bit value to BCD and then it takes 5 cycles to print each char (so 50 cycles as it checks all 10 chars for each value). That makes it take 132 cycles per pair of values, so roughly 66 cycles (33 seconds) per value.

MPU6.0 Specs:
5 tick clock speed (0.5 seconds)
4 stage pipe (with dual instruction fetching to speed up conditional instructions)
8 bit word length
uses a custom version of VLIW arch
10 general purpose registers
32 bytes of RAM
128 word instruction ROM (split into 2 lots of 64 for dual fetching)
8 bit hardware multiplier
32 bit BCD converter
16x8 Pixel Display
10x5, 8 Segment Char Display (using the MPU-64 character set)

Join the URCL (Universal Redstone Computer Language) discord:
discord.gg/jWRr2vx