|
|---|
| | 
Eric Gus
Posts 477
04 May 2017 01:25
| Rafal C wrote:
|
Pedro Cotter wrote:
|
I would be very surprised if they don't as it would be against the Vampire "spirit" :-)
Just wait a little longer.... and, damn, why only 2mb?
|
You have to think thirst about how amiga works. The chipset doesn't have access to Vampires RAM, so if a program place a music in the area after 512kB, Paula wouldn't be able to play that, because that area doesn't physicly exist. It exists only in Vampire, so in order to do that you have to implement Paula in Vampire too and any other device that uses CHIPRAM through chipset. The only problem with that might be a floppy drive i think.
|
This is why I am perfectly happy with the Vampire re-implementing the custom chips in the FPGA to work around such convoluted scenarios without having to rely on overly complicated bizarre schemes to try to keep utilizing the original chips.. There is no technical reason not to do just offload the functionality to the FPGA other than for nostalgia reasons. I know some folks might have adverse reactions to that, and thats OK, sometimes we have to make compromises to move forward. If the Vampire team can re-implement technically superior replacements in every aspect for the original chips there is no reason not to use it more so if the gain is significant for what you might even possibly be giving up.. Its just like someone with an old 512k chip Amiga modding their system with a 2meg A3000 Agnus, flicker fixer, 16bit audio , USB .. the only real difference is they are using actual commodore chips and various 3rd party boards. But the end result is the same, the original chips are being sidelined for technically superior replacements.
| |
| | 
Michael R
Posts 281
04 May 2017 09:06
| eric gus wrote:
|
Rafal C wrote:
|
Pedro Cotter wrote:
|
I would be very surprised if they don't as it would be against the Vampire "spirit" :-)
Just wait a little longer.... and, damn, why only 2mb?
|
You have to think thirst about how amiga works. The chipset doesn't have access to Vampires RAM, so if a program place a music in the area after 512kB, Paula wouldn't be able to play that, because that area doesn't physicly exist. It exists only in Vampire, so in order to do that you have to implement Paula in Vampire too and any other device that uses CHIPRAM through chipset. The only problem with that might be a floppy drive i think.
|
This is why I am perfectly happy with the Vampire re-implementing the custom chips in the FPGA to work around such convoluted scenarios without having to rely on overly complicated bizarre schemes to try to keep utilizing the original chips.. There is no technical reason not to do just offload the functionality to the FPGA other than for nostalgia reasons. I know some folks might have adverse reactions to that, and thats OK, sometimes we have to make compromises to move forward. If the Vampire team can re-implement technically superior replacements in every aspect for the original chips there is no reason not to use it more so if the gain is significant for what you might even possibly be giving up.. Its just like someone with an old 512k chip Amiga modding their system with a 2meg A3000 Agnus, flicker fixer, 16bit audio , USB .. the only real difference is they are using actual commodore chips and various 3rd party boards. But the end result is the same, the original chips are being sidelined for technically superior replacements.
|
I absolutely agree with you there! It is totally necessary to re-implement and thereby replace and enhance the functionality of the original Amiga chipsets. That will benefit everyone by bringing us more modern hardware and a much better Super AGA chipset. The standalone board requires everything to be in FPGA. The original chips are just hardware. The Vampire accelerators and even the standalone motherboard are a very close approximation of the original "Amiga Experience". The old hardware won't be missed. But the new hardware opens up lots of possibilities!
| |
| | 
Olaf Schoenweiss
Posts 690
04 May 2017 09:21
| 10 points for candidate :-) you just realized what Vampire/Apollo is about if you need just a accellerator then there are other options
| |
| | 
E Penguin
Posts 46
04 May 2017 10:00
| It would be quite nice (although I have no idea how feasible) if it could be configurable; for compatibility or pure nostalgia, to select whether the FPGA version or original chip is in-use. I imagine it's mostly a case of configuring the memory map so wouldn't be too difficult.
| |
| | 
Rafal C
Posts 3
04 May 2017 14:45
| Mark Smith wrote:
|
But thinking on that how would the Vampire know if Paula or some other device writes to the real ChipRAM ?
|
The chipram mirroring would be for the graphics chip sake only and blitter/copper would operate on the Vampire side. The problem might appear if a programmer used some weird techniques to transfer data directly to the framebuffer with something else than CPU/BLITTER (does somebody know anything about that?). It's just an idea (may be wrong) and propably a temporary solution till the Vampire get full chipset inside FPGA.
| |
| |
Eric Gus
Posts 477
04 May 2017 15:59
| Rafal C wrote:
|
Mark Smith wrote:
|
But thinking on that how would the Vampire know if Paula or some other device writes to the real ChipRAM ?
|
The chipram mirroring would be for the graphics chip sake only and blitter/copper would operate on the Vampire side. The problem might appear if a programmer used some weird techniques to transfer data directly to the framebuffer with something else than CPU/BLITTER (does somebody know anything about that?).
It's just an idea (may be wrong) and propably a temporary solution till the Vampire get full chipset inside FPGA.
|
I think thats why they came up with the parasite ...
| |
| | 
Saladriel Amrael
Posts 166
05 May 2017 13:11
| eric gus wrote:
|
[...]
I think thats why they came up with the parasite ...
|
As far as I can remember Gunnar clearly stated that's it.
| |
| | 
M Rickan
Posts 177
05 May 2017 20:09
| Saladriel Amrael wrote:
|
As far as I can remember Gunnar clearly stated that's it.
|
Precisely... the Parasite has a limited shelf life.
| |
| | 
Pedro Cotter
(Apollo Team Member)
Posts 308
05 May 2017 21:25
| They are getting there... EXTERNAL LINK
| |
| | 
Gunnar von Boehn
(Apollo Team Member)
Posts 6214
06 May 2017 10:00
| AGA in A600
| |
| | 
Aksel Andersen
Posts 120
06 May 2017 10:06
| Gunnar von Boehn wrote:
|
AGA in A600
|
4mb chipram? Hmm..
| |
| |
Gunnar von Boehn
(Apollo Team Member)
Posts 6214
06 May 2017 10:50
| Compare rendering
AGA HAM8 on A600 with APOLLO
with rendering on
AGA HAM8 A1200 68060 CPU EXTERNAL LINK
| |
| |
Saladriel Amrael
Posts 166
06 May 2017 10:56
| Whew! That's almost 180% the speed of a 060-80Mhz Nice to see AGA planar modes to work properly on SAGA :)
| |
| |
Gunnar von Boehn
(Apollo Team Member)
Posts 6214
06 May 2017 11:02
| Measuring AGA Speed on A600
PAL 8Color
Please compare it to your A1200 !
| |
| |
Gunnar von Boehn
(Apollo Team Member)
Posts 6214
06 May 2017 11:33
| Now lets compare Text-Render Speed on
AGA A600 with other AMIGAs
| |
| | 
Simo Koivukoski
(Apollo Team Member)
Posts 601
07 May 2017 07:43
| Gunnar von Boehn wrote:
|
Compare rendering
AGA HAM8 on A600 with APOLLO
with rendering on
AGA HAM8 A1200 68060 CPU
EXTERNAL LINK |
Thanks for sharing. A bit of details: Cinema4D has no 040/060 FPU optimized code (or FPU versions of C4D are messed up). Cinema4D can be found from CU Amiga Magazine's Super CD ROM 27 and this default scene is easy to reproduce.
Apollo1260@80MHz:
Integer: 4m17s "NOFPU"
FPU: 4m21s "FPU Version & Phase 5 060 boards"
FPU2: 4m28s "Other 060 boards (Apollo, Draco)"
--
68080: 2m13s "NOFPU"
| |
| |
Eric Gus
Posts 477
08 May 2017 07:14
| Aksel Andersen wrote:
| | 4mb chipram? Hmm..
|
You saw that too.. humm indeed.. I certainly hope its what I suspect it is..
| |
| | 
David Wright
Posts 373
08 May 2017 15:15
| For now, until further development chip ram is still important.
I have been extensively using various graphic programs like personal paint, xipaint.
You start editing an image of any size and memory warnings start right away. It has little to do with the 128mb ram included since I believe even in RTG it is using the limited chipram, which in an Amiga 500 with vampire is very limited.
| |
| |
Eric Gus
Posts 477
08 May 2017 17:11
| David Wright wrote:
|
For now, until further development chip ram is still important.
I have been extensively using various graphic programs like personal paint, xipaint.
You start editing an image of any size and memory warnings start right away. It has little to do with the 128mb ram included since I believe even in RTG it is using the limited chipram, which in an Amiga 500 with vampire is very limited.
|
Well I hope the Apollo team has fixes planned for that limitation, I agree with you a 512k chip ram A500 would be rather limiting.
| |
| |
Gunnar von Boehn
(Apollo Team Member)
Posts 6214
08 May 2017 17:15
| eric gus wrote:
|
David Wright wrote:
|
For now, until further development chip ram is still important.
I have been extensively using various graphic programs like personal paint, xipaint.
You start editing an image of any size and memory warnings start right away. It has little to do with the 128mb ram included since I believe even in RTG it is using the limited chipram, which in an Amiga 500 with vampire is very limited.
|
Well I hope the Apollo team has fixes planned for that limitation, I agree with you a 512k chip ram A500 would be rather limiting.
|
Which limitation?
You saw that we can give you today 4 MB chipmem?
And if this is not enough then I can give you more, up to 128 MB if you want
| |
|
The team will post updates and news about our project here