turbo-ZX
ZXNet echo conference «real.speccy»
09 Nov 99 18:19, Alexander Kormishin wrote to Kirill Frolov:
KF>> You forgot one thing - after reading the COP you will have 1/2
KF>> clock cycle after
KF>> removing WAIT, and after reading the data 1 clock cycle. Because diagram of
AK> Come on, show me and All here “after reading the COP there will be 1/2 clock cycle
AK> after removing WAIT, and after reading the data - 1 clock cycle”:
AK> ·┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌·
AK> Ft-7M·└─┘ └─┘
AK> ____·--──┐ ? ? ? ┌─?───────────┐ ┌─?───────┐ ? ┌─?──·
AK> WAIT· └─────────────┘ └─┘ └─────┘ ·
Highlighted with the letter X.
AK> You forgot that we are talking about the subject in Leningrad, where the highest priority
AK> memory access is given to the video controller. And if you somehow
AK> suddenly win your half-beat, the video controller will still take them away from you later
AK>.
This is the problem and it needs to be fixed.
AK> Access windows run at the same frequency and are not shifted in phase.
On the curb - once. Only CAS RAS WE memory should be turned on.
KF>> 1/2 beat on each reading. Only it may not exist,
AK> It’s not “maybe”, but it simply won’t be that’s all.
It’s possible on the curb!
AK> Again, inattention: I didn’t mention it in the originalmessage, acceleration on the curb.
AK> Therefore, with the above principle, this was not lying anywhere.
And then you can write that you only mentioned the kind of acceleration that
you mentioned and there can be no other.
KF>> There is little hope. It won’t work anyway.
AK> If you set it up this way from the beginning, then it’s better not to do it at all.
And this is so!
Kirill Frolov. [ZX]