8 Mb CHIP RAM on All Amigas

[Piru]

@utri007

fblit what it does? would it be possible to to assign more memory to graphics with it ?

No.

FBlit just makes all non-displayable bitmaps to reside in fast memory and replace hardware blitting with CPU operations. It doesn't give you any more chip memory, but it reduces usage of it, leaving it for other purposes, and speeding up system (fast memory is, err, faster).

So in a sense FBlit is the best you can get without replacing the whole hardware.

And really, fast gfxcard is much better option, you get truecolor modes aswell.

[Piru]

Then if it is not possible without a gfx card. it is possible with one as we already know so my solution in hardware would be create a gfx card thats accessed through RTG I suppose, but isnt a gfx card but only the ChipMem part of it and how it is accessed.

This is not possible. External hardware cannot take over the custom chips inside the amiga.

The only way would be to replace the whole amiga motherboard and chipset.

[Tricky]

Piru wrote:
@Tricky
Nope, only CPU accesses use MMU. Custom chip accesses always access the physical mapping.

I thought they might.  This doesn't affect the four virtual amigas scenario though, as on switching you'd turn off the display and sound, copy the entirety of Chip ram into Fast ram, copy the new Chip ram block out of Fast ram, and restore the display.  The MMU can map access to the hardware registers though, I take it, so it would just point the previous virtual machine's DFF000 range into some reserved fast ram, where it will happily carry on without doing anything... as long as it doesn't try to use the blitter, of course!  But then there's FBlit.

It would also rely on software being fully system friendly in order to work properly.

[Piru]

@leirbag28

Before continuing with this debate, would you be kind enough to explain what would one use the 8MB chip memory for? What benefits does 8MB chip memory offer?

[Lemmink]

Then if it is not possible without a gfx card....it is possible with one as we already know..............

Where do you get the strange idea that you could use even one kb more ChipRAM with an installed GFX-card.

so my solution in hardware would be.............create a gfx card thats accessed through RTG I suppose, but isnt a gfx card but only the ChipMem part of it and how it is accessed.

Ah, now I see. You confuse the GFX RAM on an graphicscard with Chip RAM. GFX RAM ist more like Fast RAM, so it is not surprising that there is as (PPC only) CGX driver for AGA.
Though Chip RAM is used to store GFX-data it is only a fraction of it's duties. This part can easyly taken off of the chip RAM, thats how RTG systems work. But it only works for software that goes through the OS.

As for AmigaUAE. Three is an UAE-port for AmigaOS but only the fastest classic PPCs are barely able to reach simpe A500 speed.


Anyway whats the point of haveing more then 8 MB of chip RAM ? You might be able to open more screens in higher Resolution in PAL/NTSC but the chipset wouldn't get any faster, so eventually you could watch at the monitor an see how the screen is updated line by line.

[Tricky]

Piru wrote:
@leirbag28

Before continuing with this debate, would you be kind enough to explain what would one use the 8MB chip memory for? What benefits does 8MB chip memory offer?

Hahahaha!  I was thinking exactly this, but I'm more interested in just solving the problem!  I'm addicted to thinking.

It's driving me crazy!

[tonyvdb]

Here is another option.

Why not just use the FBLit hack and have some of the stuff normaly loaded into the chipram get moved into the fast ram. This is really the only option at this point.
I use it and I get 1.8 meg of free chipram with my OS3.5 64 colour workbench in Highrez NTSC and a nice workbench background.

[Doppie1200]

Piru wrote:
FBlit just makes all non-displayable bitmaps to reside in fast memory and replace hardware blitting with CPU operations.

That is why fblit only speeds up a system when the CPU is beefy enough. In fact it is an example of the lacking chipset. The whole idea about amiga being offloading the CPU which made it a better computer than it's competitors in the day. FBlit does the opposite. It offloads the blitter and lets the cpu take care of things. If you cant beat em join em I'd say

Piru wrote:
@leirbag28

Before continuing with this debate, would you be kind enough to explain what would one use the 8MB chip memory for? What benefits does 8MB chip memory offer?

I think only workbench related things can benefit from this. More screens open, webbrowser graphics that sort of thing. Games and such count on only having 2MB so most of the time it is useless having 8MB.

Powerhungry addicts want to drool over their workbench when it reports the ammount of RAM they don't use. More unused RAM feels better I guess.

[tonyvdb]

"Powerhungry addicts want to drool over their workbench when it reports the ammount of RAM they don't use. More unused RAM feels better I guess."

For myself its needed because The Toaster/Flyer software needs at least 1.6mb of chipram in order to function properly. I dont like running my workbench under NTSC Highres mode because it looks aweful as everything is so large and bulky.

[xeron]

But I believe it is possible.

Then you are wrong.

Ok then, it is time to create UAE for Amiga....to emulate the 32 bit needed for the chipRAM and have the rest of the emulation for everything else run natively

The problem is this: The chipset requires that any bitmap that is displayed or blitted, and any sound samples must be in physical chip ram. The chips that create the display and output the sound have a physical limitation on the amount of memory they can actually read. They have no memory mapping unit, they just have a memory bus that is limited to 512Kb, 1Mb or 2Mb depending on which revision. This limitation is to do with physical pins on physical chips and wires between them. To increase the amount of chip memory, you'd have to make a new revision of the Amiga chipset with extra chip ram address pins, and make a new motherboard to use them.

To emulate 8Mb chip ram on the Amiga, you have to emulate the whole Amiga chipset, and the 68k just isn't fast enough to do that, especially emulating AGA on AGA would be really slow, and actually really crap since you'd be limited to 256 colours (or possibly you could use HAM-8, but you're talking minutes per frame here now, rather than frames per second). You can't emulate 8Mb chip ram and expect the real chipset to use the emulated chip ram, because they physically cannot address that much ram. To them there isn't that much ram in the whole world.

since it is a real Amiga.............keeping everything still running smooth :-)

No. Its not possible.

Then if it is not possible without a gfx card....it is possible with one as we already know

No. An A1200 with a 32Mb gfx card still only has 2Mb chip ram. The Amiga chipset can't directly access any RAM on the gfx card at all. The graphics chipset on the graphics card can't access the amigas chip ram usually either.

Well then....one will have to be created to take care of that :-)

What? A piece of code that physically adds pins to chips and tracks to motherboards?

 If you had an MMU I guess you could use the Fast ram as "virtual" chip ram in the same way it would normally use HD

Ah......now we are getting somewhere :-)  thanks for open mindedness :-)

Nope, you're not getting anywhere. It is simply not possible to get the chipset to see outside of chip ram, it is physically directly attached to chip ram and not through any kind of memory management unit. No amount of software can fix this.

[KThunder]

i think the last time this argument ran through amiga.org i tried to get the arguer to understand that the problem really is with the hardware. the os has no problem with 8megs of chipram. the hardware is incapable of it. any software to remidy the situation would result in an enormously slow bloated mess. as in probably 10 to 12 megs of fast ram being used and data being shifted back and forth from actual chip to fast. i told them to prove it possible. no proof yet.

emulating chip ram is worse than the lack of chip ram in the first place. there are ways around 2meg of chip problems. and if you absolutly have to have 8megs, use winuae.

[Tricky]

KThunder wrote:
any software to remidy the situation would result in an enormously slow bloated mess. as in probably 10 to 12 megs of fast ram being used and data being shifted back and forth from actual chip to fast.

Depends how often you have to switch between frames.  It takes 31/50 of a second to copy 2megs of chip ram to fast ram, or back again (I just timed it!! although that's on my 25MHz '060 but I'd expect chip ram to be a bottleneck anyway).  I guess that's acceptable if it only happens when you switch between one app and another (or one set of apps and another), if you don't have to switch all that often.

[Zac67]

@Tricky: There's no point in using a MMU for the chipset (not the normal, CPU integrated one of course, but one you build for that purpose).

1) An MMU can help mapping logical memory to physical memory - provided the address space of logical memory is larger than the amount of physical memory, so when a currently not existing page is accessed, you stop access, load the contents to physical RAM and then map it to where you need it. The Amiga's chip mem address space is equal to the physical RAM (2MB), so there's no sense in that.
2) If a page miss occurs, you need to stop access (using an exception) for loading the data to RAM. So how do you do that while the frame is displayed or the sound is played?
3) Chip RAM access by the chipset is realtime, that's the whole point of creating chip RAM in the first place: define memory that can be accessed in a defined manner. The CPU is stalled when the chips need the bus. Fast RAM on the other hand can be accessed by the CPU without being slowed by chip access.

[Tricky]

@Zac67
That's not what I just suggested.

I'm suggesting a scheme where you can MANUALLY switch between completely different 2Mb banks of chip RAM, that are used by entirely independed sets of programs.  When you switch, all the programs in your current frame stop.

The MMU only comes in if you want the programs to still run in the background.  You remap the custom chip hardware register address range (which is written to by the processor, not the hardware) to an area of fast ram, so the other programs aren't accessing the hardware anymore, even though they don't know this.  You also map the chip ram into fast RAM.  The hardware NEVER ACCESSES this fast RAM while the program is in the background, because the hardware registers are mapped to empty space for these processes, although the program itself can still write to what it thinks is chip ram.  You only see or hear the frame you're working on at the time.

[Zac67]

OK, but
- How does the OS now which part of code owns which parts of chipram? What happens if the user drags a screen?
- Chip registers ARE being written to be chips (i.e. by Copper). No big deal when it comes to color registers, but what about Copper controlled blits? Of course you could use the MMU to supervise access to the 'virtual' registers, but then again you'd have to emulate chipset behavior...

Apart from that I don't like the idea of [d]paging[/d] banking chipmem very much, sorry. I'd rather use an RTG card for graphics and a DMA capable sound card. What else would you need chipram for? The whole scheme was engineered for a highly efficient, low cost, yet high performance graphics machine. Technology has moved on since.

[Hyperspeed]

A lot of things like scandoublers/ff and '030 accelerators for the A600 feature a chip that sits on top of another chip and bypasses the original.

I wish people would stop saying "can't be done". With enough resources and ideas I'm sure something could be worked out.

Trouble is, with BVision and PIV cards the urge to develop a hardware based solution to the 2MB limit is not strong enough.

I heard once there was an '040 sidecar for the A500 and the Bodega Bay tower for A500 allowed A2000 GFX cards right?

I can see that more ChipMem would be handy for classic apps, especially killer apps you'd like to 'nourish' with all the tricks you can use!

FBlit conflicts with things though and is a severe hack, I don't think Shapeshifter likes it.

Shame the AGA plans were lost, it could have made for great fun over the last decade. Didn't Mick Tinker have the plans for BoXeR? What happened to his work...

The fake GPU idea is a good one, tricking the Amiga into using GFX card memory as ChipMem - maybe with some sort of FBlit/Hardware combo. The board could have lightning fast DDR memory.