Thanks! I can't help but read these kinds of articles. They are so cool. Frankly, I don't understand fiction books because this kind of writing is so much more interesting.
for some reason D6 takes 14-18 more cycles than D7
Probably because this is implemented in microcode and that path was never tested.
If a byte of memory is modified after being read into the queue, the original value in the queue will determine the instruction executed, not the modified value
I'm sure many bugs can be traced back to that behavior.
Thanks! I can't help but read these kinds of articles. They are so cool. Frankly, I don't understand fiction books because this kind of writing is so much more interesting.
>for some reason D6 takes 14-18 more cycles than D7
Probably because this is implemented in microcode and that path was never tested.
>If a byte of memory is modified after being read into the queue, the original value in the queue will determine the instruction executed, not the modified value
I'm sure many bugs can be traced back to that behavior.
The V20 was a cool chip for it's time. I had one in an early PC, used it as a bridge between CP/M and DOS since it had kind of an 8080 mode as well.
The V20 was a cool chip for it's time. I had one in an early PC, used it as a bridge between CP/M and DOS since it had kind of an 8080 mode as well.
(post is archived)