Generating the Keyboard Matrix

Hello Alex Crush ┌─- CODE ─── ;SRC - source map ;DST - destination map, 256-aligned ld de,SRC ld hl,DST fillloop; ld; (hl),-1 inc l jr nz,filloop ld c,1 ;step cpyloop; ld; a,c cpl ld l,a ld b,c fillal; ld; a,(de) and (hl) ld(hl),a dec l djnz fillal ld a,l sub c ld l,a jr nc,fillal inc de rl c jr nc, cpyloop └── CODE ─── I threw it in hastily. Not sure if it works correctly :) The algorithm, I think, is clear.

Hello, All I was once (4 years ago) creating a PS/2 keyboard controller based on second spec ;-). The controller was wait-free by reading the key code the main spec directly from the memory of the spec controller (yes, I’m probably a pervert :) ). Consequently, the task was to generate a 256-byte keyboard table. That is, the 0th byte is what the main spec will receive when reading from port #00FE, the 1st - #01FE... Etc.. If anyone doesn’t know, each byte in this table is calculated as AND of those rows keyboards whose numbers in the number of this byte are zero. That is, the value of cell number #D6= %11010110 is equal to the AND of the 0th, 3rd and 5th rows keyboards. The value of the last cell is port #FFFE, ess = #FF So, I propose (as a study, perhaps) to write a procedure for calculating such a table. Naturally, you want the procedure to be as quick as possible. Input data - 8 bytes (keyboard rows, only the 5 least significant bits have meaning, but this is kind of on the side), the output is 256 bytes. There are no more restrictions. P.S. I remember I got a result of about 7000 cycles. Today's reflections seem to allow us to reduce this number to ~4500 cycles (although I I thought for sure and didn’t check my thoughts at all). Go for it if anyone is interested :-)

Hello Vitamin Vit> I think the algorithm is clear. The algorithm never caused any problems initially :-). I suggest quickly practice ;-) . And you can do it (although I can’t already at 3 a.m. it’s normal to count something) approximately “quite a lot of ticks”

Hello Alex Crush Speaking of keyboards, a few years ago I disassembled AT keyboards for repairs (!) (yes, they were repaired once...) so there was actually something there as a controller VERY similar in appearance and with the inscription Z80... So it doesn’t look like a pervert AlexCrush pulls and can sleep peacefully :)))) and if I’m lucky, I’ll try to find a rarity and take a picture.

Hello, Witch Doctor > there really was something VERY similar in appearance and with the controller > labeled Z80... Zilog released keyboard controllers Z8602/14/15/C15/E23 Used in many AT keyboards.

Hello caro 2caro - as I understand it in your ZXMC when the spec keyboard is polled by wait used, does it slow down much when loading from a cassette, for example?

Hello caro I suggest not going off topic. Isn't anyone interested in trying to come up with a quick way? Or is everything too strong? Are you busy with something?

Hello Alex Crush Well, you can do this... ┌─- CODE ─── ld de,dst allkc; ld; hl,src ld a,-1 exa ld a,e cpl keyc; and; a jr z,nokey rra jr nc,nochk exa and (hl) exa nochk; inc; l ;?? jr keyc nokey; exa; ld(de),a inc e jr nz,allkc └── CODE ─── I admit, I forgot whether rra affects the zero flag. If yes, then you can reduce. I didn’t count how much it weighs in cycles :)