Repeating code on an EPROM

[NeXT]

I know that when you generate an image file prior to burning you need to fill the unused memory addresses with 0's or it will not burn correctly. I spent part of my night dumping two very crusty 2716's and both have their code repeating itself 8 times. I don't see any variations on a glance so it's not like it's executing a different version of the code under certain conditions but I don't entirely see why one would do this beyond to fill space. The only other reason I can think of is someone was thinking REALLY far into the future and it's insurance against bitrot because I guess if you have it, a dump and a compare would help recreate the missing code from the other seven local copies.

 

[SiriusHardware]

First, there is no requirement for the unused bytes in an EPROM to be either 00 or FF. All that matters is that the bytes of the programmed code are correct. The only case where the values in the unused bytes may matter are if they have been included in any checksum written on the device - an EPROM with the leftover bytes filled with 00 will have a different whole device checksum to an EPROM with the same code and the leftover bytes filled with FF.

Assuming the device is actually being read out correctly and it's not just the first 512 bytes which are being read out 8 times, the most likely reason anyone would programme a large device with multiple identical images is if the EPROM were being used as a substitute for a smaller device like a 512 byte PROM. When this is done it will not matter what state the higher address lines are held in, the same 512 bytes will be seen by whatever is trying to read it. It would still be considered bad practice not to tie those upper address lines into known states.

Depending on your EPROM programmer's software, it may have a 'compare' function where you can compare successive blocks of the code to see if they really are identical.