15 August 2024, 02:20 | #1 |
Registered User
Join Date: Mar 2009
Location: moon
Posts: 377
|
An alternate history without "slow" or "fake fast" memory
Most of the non-Zorro Amiga models (specifically 1000, 500, 500+, and 600) were designed to be upgraded with memory expansions. These expansions doubled the availabe memory, and except for on the 500, it was all chip memory. Even on the 500, the memory is connected to the chip memory bus (and thus slow), and can relatively easily be configured into real chip memory, as long as the Agnus chip is capable of the full size. I don't think it's too far-fetched to assume that the original plan for the 500's memory expansion was to double the amount of chip memory, but for whatever reason, Commodore didn't have a capable Agnus chip ready for the 500's launch, and worked around the problem with minimal changes.
If we imagine that the 500 followed the same pattern as the other models, meaning that developers could rely on the common configuration of 1 MB memory being all chip memory, would it have meaningfully affected the potential of the software? There are certain obvious situations, like being able to use more samples in Protracker, but what about games and demos in general? Were they typically constrained by having half the amount of chip memory, or wasn't it such a big deal? I haven't done nearly enough Amiga programming to say, but what do those of you who have think? |
15 August 2024, 02:46 | #2 |
Registered User
Join Date: Mar 2016
Location: Australia
Posts: 936
|
Even better would have been if the standard 512k expansion would have been fast ram. Games would have been much better. 512kchip 512k fast would be a dream base config for most developers.
|
15 August 2024, 08:50 | #3 |
Registered User
Join Date: May 2023
Location: Norwich
Posts: 524
|
I think it's a bit of a no brainer that more chip ram would've been preferable. Slow ram was basically a hack that gave you the worst of both worlds. Given that you needed chip ram for most things and fast ram wouldn't be that much faster on a stock A500, pretty much any developer you asked would probably have wanted more chip.
Whether it would have made a noticeable difference in the end products is harder to say, but it probably would have saved a lot of developer effort juggling things around in ram to avoid disk loading whilst keeping what is useful accessible to the chip set. |
15 August 2024, 18:30 | #4 | |
Registered User
Join Date: Feb 2017
Location: Denmark
Posts: 1,298
|
If we're dreaming, best would of course be if it was switchable (could be HW switch) between chipram and true fast ram If that's not possible true fast RAM would certainly be the most interesting "alt history" regardless of whether developers at the time would prefer it.
Quote:
It's of course true that there are many situations where you won't see a benefit, but true fast RAM allows CPU to continue executing even if chipmem is fully saturated by display/blitter. |
|
16 August 2024, 02:04 | #5 |
Registered User
Join Date: Mar 2009
Location: moon
Posts: 377
|
While that may be the case, I'm more interested in the alternate history the designers seemingly intended. I can't think of a reason for the memory expansion to be connected to the chip memory bus, if it wasn't meant to be chip memory.
|
16 August 2024, 02:52 | #6 |
Registered User
Join Date: Mar 2016
Location: Australia
Posts: 936
|
We don't need to imagine too hard what the world would have looked like if 1mb chip was the standard, as this was a "defacto" standard for many games/demo/cracks towards the end of the ECS era. I certainly recall modding my A500 for 1mb chip to open up a large number of titles previously unavailable to me.
|
16 August 2024, 07:36 | #7 |
Registered User
Join Date: May 2023
Location: Norwich
Posts: 524
|
Cost cutting, it avoids needing to implement a separate memory controller that isn't encumbered by chip ram timing.
|
Yesterday, 11:48 | #8 | |
Registered User
Join Date: Mar 2009
Location: moon
Posts: 377
|
Quote:
That's a good point. |
|
Yesterday, 17:47 | #9 |
Registered User
Join Date: May 2013
Location: Grimstad / Norway
Posts: 876
|
In the Stupid Q Department: The ROM bus is obviously separate, so doesn't it have its own "controller"? Doubly so for the A1000 that has the WOM that _is_ fastram?
|
Yesterday, 19:15 | #10 |
Registered User
Join Date: Mar 2009
Location: moon
Posts: 377
|
For the chip memory bus, Agnus multiplexes address lines to the memory chips, and handles refresh. I assume the logic chips on fast memory expansions (and the A1000 WCS daughterboard) implement similar functionality. The ROM doesn't need any of that.
|
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
"Diabolik" & "Dylan Dog" & "Tex" & "Time Runners" series | DamienD | request.Old Rare Games | 20 | 21 July 2022 16:58 |
"Voices8" 8 Channel Soundtracker "DemoSongI" song - "This is the Amiga with 8 Voices" | DemosongIHunter | request.Music | 45 | 23 May 2022 20:07 |
"Screech!! v2.41" & "Screech!! [AGA] v2.51" - "HD install" --> "ADFs" | DamienD | request.Old Rare Games | 45 | 15 June 2020 12:42 |
Difference WHDLoad-Games "normal" & "fast" | pellewolf | project.WHDLoad | 3 | 25 July 2017 15:38 |
So called "History of Videogames" on Times "Month" CD | Antiriad | Retrogaming General Discussion | 11 | 26 May 2009 15:41 |
|
|