There more to it than that. The -5 is really started by Intel and was really meant to indicate 8085 bus timing compatability which is slightly different than 8080. It shows up in the MCS-85 Users manual 1978. So the -5 was 8085 compatable and that part was nominally available in 3mhz speeds. So if you did anything else for 8085 at 5mhz you had to use another number other than 5 was used. The base 8255A was slower as the 8080 never got over 3mhz. A side effect of this was the -5 was usually faster and the 8255A-5 was reverse compatable to 8255A. The difference are subtle but do exist. OBTW: the -n for peripherals didn't always relate to CPU clock speed as it was bus timing related. One could argue bus speed and timing are the same but the timing overlaps for things like CE/ to READ/ or WRITE/ can be different and in the case of 82xx peripherals the -5 was aimed at the 8085 and others were maybe not 8085 but faster time wide for WR/ or RD/ active time. Typically Mistusbishi and NEC were either excatly the same or radically different for numbering. It really shows when 8085s or 8088s faster than 5mhz. the 8255 existed to 12mhz(CMOS) as uPD71055 (memory test on number). It was very common in the market that specific -n didn't mean speed and even when it did -1 or -5 could mean the fastest or slowest. especially for non memory parts. Generally this is most important if your doing fast 8085 (5-6mhz) 8086/88 at 5mhz or more or Z80s at 4mhz or faster. Allison