Welcome, Guest. Please login or register.

Author Topic: Indivision ocs/ecs  (Read 2262 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline BLTCON0

  • Jr. Member
  • **
  • Join Date: Oct 2013
  • Posts: 91
    • Show all replies
Re: Indivision ocs/ecs
« on: December 25, 2013, 03:57:10 PM »
Quote from: motrucker;755332
I received one of these adapters from Amigakit for Christmas (actually got a backoder slip, but it's one the way).
I have a question. I just noticed my A2000 has an old (ocs) denise in it. How will this effect performance? Any one have an extra ECS denise?

It won't affect the performance at all.

You can even leave the Denise chip out, the Indivision provides a hardware level Denise replacement so it doesn't matter if the "real" Denise is OCS, ECS or even present at all as far as output from the Indivision is concerned. The only reason to keep Denise installed atop the Indivision is to keep having output from the DB23 video connector.

As a Denise replacement, the Indivision ECS is of course ECS capable, provided an ECS Agnus is also used.

Of course ECS screenmodes will only be usable via the Indivision's VGA output if the "real" Denise is OCS.
 

Offline BLTCON0

  • Jr. Member
  • **
  • Join Date: Oct 2013
  • Posts: 91
    • Show all replies
Re: Indivision ocs/ecs
« Reply #1 on: December 26, 2013, 12:22:43 PM »
Quote from: Oldsmobile_Mike;755394
I've heard this before.  Personally I'd be nervous as fu...dge powering up a Miggy with one of it's custom chips missing, LOL.  I know to use the Indivision in an A600 you plug it into a card in the memory expansion port (no Denise needed), but then the A600 already has a Denise on the motherboard.  I dunno, would like to see a picture of someone using it without the Denise before I'd try it myself!  ;)

Been there, done that :) Powered it up with just the Indy installed and the Kickstart screen agreeably popped up just fine.

It isn't a case of "powering it up with a custom chip missing" anyway, as the Indy is a Denise replacement, registers and all. So as I see it, it just needs to sit somewhere with access to the CHIP memory bus, the RGA lines and a colour clock.

For example, on the A500 trapdoor connector, the RGA lines are missing, so it wouldn't be possible to install the Indy there without additional wiring. On the A600 trapdoor the RGA lines are there, so it can happily work there. Similar story with the A1000 where the Indy is installed alongside Paula via an adapter - the lines needed are there.

There's only one possible gray area here - mouse movement. Denise has inputs for the mouse movement pulses and these are of course not routed to either the A600 trapdoor or Paula. So it's obvious that in the A1000 and A600 case the Indy does rely on the actual Denise being present, to count mouse pulses.
On the A500/A2000/A3000 case it sits on the Denise socket and therefore has access to the mouse inputs, whether it implements pulse counting in this case or still relies on Denise I can't remember. I'm unsure if I booted my A500 beyond the Kickstart screen when I tested the Indy without the Denise, it's been 2+ years since and have no real desire to take that A500 apart again!
But rest assured that powering it up without the Denise is harmless, so if you feel like trying it out you can fill in my lack of memory for the possible mouse issue :)

Happy holidays!
 

Offline BLTCON0

  • Jr. Member
  • **
  • Join Date: Oct 2013
  • Posts: 91
    • Show all replies
Re: Indivision ocs/ecs
« Reply #2 on: December 26, 2013, 06:56:39 PM »
No idea if anyone has demonstrated the multimonitor setup.

But I am practically certain that under this scenario the DB23 output will be completely useless (i.e. Denise's output won't be a clone of the "1st" or "2nd" monitor, but garbage).

Another certainty is that maximum resolution and refresh rates will not be attained in a dual Indy setup, as fetched data will have to alternatively be routed to the two Indies, cutting the effective refresh rate to half for a given bandwidth requirement.

(i.e. if for a given screenmode bandwidth requirement is X, then for the same screenmode in two screens obviously bandwidth requirement would be 2X. So to keep it again down to X, only every other frame must be drawn, thus cutting the effective frame rate to half. Of course if the extra X bandwidth is both available and not needed for other stuff - CPU, copper, blitter - then the frames can be normally drawn. For example two 320 x 200 x 4 bpp screens @ 60 Hz is possible, albeit of marginal use. In ECS mode, two 640 x 200 x 2 bpp @ 60 Hz would be possible as well. But two 1280 x 200 x 2 bpp screens would bring the effective refresh rate down to 30 Hz, same for two 640 x 200 x 4 bpp screens, it'd also have to be 30 Hz in that case).

Needless to say this is just my guesswork on the subject :)