Smonson wrote: 10 Dec 2021 23:47
mrbombermillzy wrote: 09 Dec 2021 09:01
I would be more than happy to re-purpose my 'Shifter Killer' routines from the TT. Any relevant data gleaned there can be fed back to the HDMI Shifter project for compatibility purposes (if that's a needed thing?)
Yes please, that is definitely needed. What we need for test cases is a short sequence of code (one scanline or so) that twiddles the shifter in some advanced way and causes it to output data differently. {Closure} is doing this sort of thing but we don't have enough info to build a test case.
Well, probably best to explain from the beginning, so you know where Im coming from...
The routines I am working on (1st display method) are actually code for high colour static display output for 8bpp palette machines (i.e the TT, but will likely transfer to F030, A1200, etc). These routines are similar to Spectrum512/Quantum but with extra colour write avoidance code. The more advanced stuff (2nd display method) is similar, but almost like a poor mans MJPG codec for the use of action games with a similar high colour content available (IF I can pull it off and don't find any minor detail, of many, that becomes a major show stopper in the meantime).
I've blogged the initial testing here:
https://www.exxosforum.co.uk/forum/viewt ... =18&t=3936 but things have slowly advanced a bit more since then (nowhere near getting complete however, but some of the errors in the explanations in that post have been corrected or better ways of doing things have been realised) and more importantly, I believe the end goals of both display methods can be achieved, especially if boosters are included in the mix with display method 2.
Anyhow, (finally!) to my point... a 'side effect' of this research is some crazy experiments on the poor TT shifter to try and realise the above end goal.
I have several simple sync locked test programs, write limit stress tests, etc; mainly visual aids as to the working of the shifter essentially.
What I can attempt to do is transfer these mainly visualisation tests to an ST shifter. But is the one you are creating an exact copy of one of the ST/F/Mega shifters or will it be an 'augmented copy' i.e. fully 100% compatible, but with extra features? Let me know any differences and I will factor that in.
While I have pretty much done my own thing with the TT shifter because of lack of detailed info (wish the Atari Profibuch had been translated! lol), the ST one has been far more thoroughly documented.
As you have already mentioned, Troed is one of the major sources of accurate detailed information for the workings of the ST shifter. I'm sure he will have far more knowledge than I do (having never actually coded for the ST yet! :o ), but I would maybe be able to chip in if no one better is available...happy to help if I can at all.
One last thing to note... If you DO need/want my help, my work for this will be done/tested from my end on the only ST I own, which is a Mega2 with 4Mb. Now I'm not sure if it either needs attention, as 'Closure' displays graphical artifacts, or perhaps its the fact that the Mega has its own special 'close to but not quite the same as' an STFM 8Mhz osc. Perhaps
@troed knows if its the demo not suitable for the slightly out clock, or my machine that needs attention! :lol:
P.S. Sorry about the overly long post guys. Just thought it best to explain everything properly to ascertain whether what I can contribute will be useful/relevant at all.
Perhaps any further correspondence can be done via PM to avoid completely derailing this interesting topic? :)