the diagram reads clearly and there is no error, it is not a comparison between, – Z80 and 8080A – it is instead a comparison of the machine times related to Intel’s 16 bit 8086 cpu with respect to the machine times of the Z80 CPU. –
I also found “wrong” textbooks that indicate data that does not conform to intel specifications, and I prefer not to name texts and authors, in particular I read one by a Ligurian professor who surprised me a little in the negative sense.
But I prefer to say no more about this matter, I just say that the Z80 does not need the active-high “ALE” signal on PC-IBM to allow for address decoding. – This sign that cripples “the INTEL architecture” is a legacy of the Intel 8080A and has not yet been eliminated. –
Note: – On the 8080A the ALE signal was called SYNC, it was an active-high signal, this signal gets the name of ALE for the first time on the Intel 8085 processor. – Only to take care of an external contribution clarification.
(Isaac Luongo, 15 November 2019) –
The Z80, on the other hand, has two separate buses from birth and this allows it to have a transfer rate on I / O equal to 3 clock cycles against the 4 cycles + 1 delay imposed by the standard INTEL and IBM. – So unlike what happened in the comparison between the CPU Z80 with the MOS 6502, I believe that there is no need to write a text that traps between the different characteristics of the processors [intel 8086 / intel 8088], it clearly appears the beating defeat of the jewel 16 bit intel on input / output. –
Also in this case the palm of victory is assigned by scientifically rigorous tests to the Z80. – And every practical experiment conducted on the same bus as the intel 8086 shows that the latter is clearly slower. –
First data transmission via IBM PC-XT BUS with Z80 CPU
We were deceived, wrong books and hired professors?
Meanwhile, the first problem I had when I brought the Z80 to work with a very common PC-XT expansion card was the OVERRUN-SERIAL problem. – the Z80 is too fast at the same clock rate with a normal 16-bit PC-XT. –
Is it possible that an 8-bit processor with multiplexed 4-bit ALU can be faster on the INPUT / OUTPUT than on an INTEL 8086? – In the meantime, to get the video you see without some few “CR+LF” being broadcast, I had to do various tests, with the mechanical detachment of some control signal put in place to delay the “start-of-transmission”. –
Then after some laboratory research on the signals, the truth comes out of the diagrams. –
*** PLEASE NOTE ***
Thanking to an external contribute, — coming from China, I must review some position, — I was supposing of 3 times low, (the contrib. supposes that in this three down was comprised a “wait” state). — I’m surprised from the big observation capacity of this (Chinese contrib.) —
So we can focus on a little detail, … the CPU Z80 can read/write from MEMORY in only 2 tk low. — I’ve measured and yes, it’s true. — The Z80 is more faster than I can imagine. — The 8086 require always 4 tk of clock + extra wait — ISACCO’S NOTES. —