Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: trekiej on March 31, 2008, 09:04:15 PM
-
Is anyone working on a second version of minimig?
Would adding a 1 million gate chip be beneficial?
I hope xx020 cpu could be added.
-
She's called Natami...
-
There is also a new FPGA board being designed by long term FPGA Guru from FPGA-Arcade. Do a search in these forums for MikeJ's posts.
-
alexh wrote:
There is also a new FPGA board being designed by long term FPGA Guru from FPGA-Arcade. Do a search in these forums for MikeJ's posts.
FPGA-Arcade has caught my eye :-)
-
@ alexh:
I recognise that thread.
Thanks.
-
New Xilinx super chip :-)
http://biz.yahoo.com/prnews/080331/aqm029a.html?.v=7
65nm Virtex-5 FPGA, 2x PowerPC 440, 32kB instruction & 32 kB data cache, SERDES, GTX High Speed Transceivers 6.5Gbps, 190 GMACs DSP performance, 384 DSP slices, 16.5 Mb of internal memory.
Supports: XAUI, Fibre Channel, SONET, Serial RapidIO, PCI Express® 1.1, 2.0, Interlaken etc..
Could be something to drool over in a minimig2..
165 USD in 1000 qty (or hit digikey..).
-
I have heard of multi-PPC virtex fpga's before. I am suprised I have not heard of them before in the forums.
Is there enough room for MMU?
-
You can proberbly build a complete A4000 in plain hdl-code with it. So MMU, I think yes....
-
A mini Mig solution that would interest me and most of AMiga community would be the one that will support AGA modes and has to have a compact size,like original MIni mig.
we all want to revive all bad 1200 from our childhood and play platoform on our "from future" 42 inch lcds....
-
I just bought some Virtex-5 FPGA's (for work) and I think they cost almost $16,000 a pair.
-
freqmax wrote:
New Xilinx super chip :-)
http://biz.yahoo.com/prnews/080331/aqm029a.html?.v=7
65nm Virtex-5 FPGA, 2x PowerPC 440, 32kB instruction & 32 kB data cache, SERDES, GTX High Speed Transceivers 6.5Gbps, 190 GMACs DSP performance, 384 DSP slices, 16.5 Mb of internal memory.
Supports: XAUI, Fibre Channel, SONET, Serial RapidIO, PCI Express® 1.1, 2.0, Interlaken etc..
Could be something to drool over in a minimig2..
165 USD in 1000 qty (or hit digikey..).
No way could anyone justify a run of a 1000 FPGA boards for the Amiga community... Also, if the FPGA costs that much, by the time we get down to the consumer level, the board will cost well over $300...
-
bloodline wrote:
No way could anyone justify a run of a 1000 FPGA boards for the Amiga community... Also, if the FPGA costs that much, by the time we get down to the consumer level, the board will cost well over $300...
OEM's may not have the capital to make 1000 boards, but if they do not believe they could sell at least 1000 boards, then why are they making any?
How many minimigs have been sold?
The people behind natami will not be happy if they only sell 200 boards.
OEM's should bite the bullet and produce them in batches of 1000.
-
A6000 wrote:
OEM's should bite the bullet and produce them in batches of 1000.
Why should the OEM take that kind of risk? There is no way they could sell 1000 boards at $300 each... I'm having a hard time justifying the actually quite reasonable cost of a MiniMig...
-
bloodline wrote:
Why should the OEM take that kind of risk? There is no way they could sell 1000 boards at $300 each... I'm having a hard time justifying the actually quite reasonable cost of a MiniMig...
They want to make a reasonable to large profit, they cannot do that unless they sell a lot of boards, batches of 1000 gets lower component costs and enough boards to make their profit.
The high cost of a minimig is because their production quantities are too small.
-
A6000 wrote:
bloodline wrote:
Why should the OEM take that kind of risk? There is no way they could sell 1000 boards at $300 each... I'm having a hard time justifying the actually quite reasonable cost of a MiniMig...
They want to make a reasonable to large profit, they cannot do that unless they sell a lot of boards, batches of 1000 gets lower component costs and enough boards to make their profit.
The high cost of a minimig is because their production quantities are too small.
But if they can't sell 1000, and to be realistic, they can't... then the will lose more money than they could ever make.
Acube got it right with the MiniMig.
-
@bloodline:
"No way could anyone justify a run of a 1000 FPGA boards for the Amiga community..."
You'r proberbly right. That's why one can make use of standard developer boards. Having an Amiga specific board isn't absolutely necessary.
This also why TobiFlex m68k cpu is so important. It's the key component besides Dennis chipset. That completly eliminates the need for a custom made board.
-
There are three solutions: Have users pre-pay or develop a line of credit. Sell other products alongside of it and make a profit. This all depends as to whether you are developing a successful produced product.
Produce a reference manual opening it up to the next generation of programmers so they can write their own LOS (Language Operating System) for the unit.
-
freqmax wrote:
@bloodline:
"No way could anyone justify a run of a 1000 FPGA boards for the Amiga community..."
You'r proberbly right. That's why one can make use of standard developer boards. Having an Amiga specific board isn't absolutely necessary.
This also why TobiFlex m68k cpu is so important. It's the key component besides Dennis chipset. That completly eliminates the need for a custom made board.
I disagree that 1000 Minimig2 boards could not be sold. Too many here assume that the market is limited to only a part of the remaining Amiga community. If packaged and marketed correctly to the general computer public, specially those feeling nostalgic at this time, I believe that many more than 1000 complete game machines (not just the bare boards or chip populated boards) could be sold. Shink it as far as is practical, fit it into a joystick with multiple output ports, including composite NTSC & PAL switchable, VGA, stereo audio RCA jacks, and if practical and possible, HDMI. Add a second joystick port, keyboard and mouse ports and expansion capability to possibly add ethernet and/or wifi if it can not be included in the original Minimig2 design.
I mean it could be a mainstream product if done right. I know it is not going to compete directly against XBox or PS3 or Wii, but it is a different niche machine and can be successful if marketed to the right segment of society. Yeah, not hundreds of thousands, but certainly a few thousands instead of a couple of hundred.
My 2 cents
-
A product has to have sustainability so that is why it should have some compatibility with things you can get today. The problem is that if you have disks laying around and no way to get them onto a compact flash card because your Amiga is down then you really have to answer those questions for a potential buyer.
It needs a programmer's reference guide for the FPGA's if you want people to write programs for it. You need to package it and make it more simple for people like formatting the first card or two for people. Developing more tools for people like me to use with it and this adds to your marketing ability and profitability.
I'm looking to start up my own company in the future but there is a lack of "openings" if you will. I'm looking for sustainability, people who I can trust and cooperation. If the system could be opened up so that other platforms could run on the Minimig then maybe other people's products would sell the Minimig.
-
After looking back at my original post, I had rather buy a Soft-Core CPU than a hard version. The reason is that the soft version can be used in a FPGA that has more pins and eventualy in a package with other soft-cores. I guess there could be a minimum on the number of pins used but not maximum.
edit:
Putting MiniMig and the 68k core on the same chip reminds me of commodore DTV. It would not bother me to buy a twin joystick with a keyboard in the middle and a mouse to the right or left of the outer stick.
Putting the two on one chip and then making an ASIC out it, if necessary, to make it smaller and cheaper would be nice. I feel some one mentioned the up front cost. I guess the manufacturer would have to weigh the benefit of doing so.
I hope my point is clear.
What's next hard-core cpu porn? :-D
okay that was cheesey.