JimDrew wrote:You can download the source for the various Replay cores from the svn.
JimDrew wrote:I worked on the Amiga driver for the Replay's RTG (Amiga core), so what do you need to know? The RTG driver allocates a 64K aligned 4MB block of memory (from the memory pool) during the driver initialization by Picasso96. The memory base is passed in along with the video modes. There is a blitter engine that Mike made that is pretty slick, but it's really not used much for anything at the moment. I need to fill in the rest of the OS replacement blitter functions to make it worth the effort Mike put into it. The RTG is pretty simple. It's just a frame buffer of a particular resolution and depth, defined by the RTG driver. There are only a hand full of actual functions that need to be supported to make it work. The Picasso96 does all of the magic of making RTG possible under the Amiga OS. There has never been a Rpi module being used for the RTG.
JimDrew wrote:Your argument about price is actually the motivation for making a much more robust setup. With the DE-10 at $130 retail, a user can afford to put a hundred dollars into add-ons and still have a premiere FPGA setup that is much cheaper than a Replay or Vampire board. So, a memory chip that costs 4 times as much is well worth it. Cost is just not a factor when it's this cheap to begin with. I have had a lot of people asking me to make a new IO board that is a full length board (so that all 4 screws will mount it to the DE-10), include 128MB of SDRAM, and several other features. I think the reason that some cores don't support more memory is because more memory has never been available.
JimDrew wrote:If you can use the 64MB chip in place of the 32MB chip, does that mean that the upper address bit (upper/lower 32MB) is already there and usable?
Sorgelig wrote:Just start to do what you "offer" and we will see what we can get in the final. If it will fit the MiSTer philosophy then it will me merged into. If not, then you are free to make a fork. It's possible to have a special Amiga fork with extra/special HW and improved Minimig if you are ready to support it.
Sorgelig wrote:You want to make a very complex SDRAM and even whole I/O with integrated SDRAM with size longer than 100mm (bye bye cheap PCB from Chinese mfg) with probably 4 or more layers.
Sorgelig wrote:Who will be able to do it as DIY? Most likely no one. You are here today to offer such board, but tomorrow will disappear and no body else will be able to make the same board. It will be several times more expensive PCB with expensive parts. You already gave up to produce I/O board while talking about more advanced board.. Weird, isn't?
Sorgelig wrote:JimDrew wrote:If you can use the 64MB chip in place of the 32MB chip, does that mean that the upper address bit (upper/lower 32MB) is already there and usable?
Some your posts makes me think that you are out of discussion topic. Don't you know how SDR SDRAM address pins are organized? You are going to add more chips but didn't explore the topic.. SDR SDRAM address is split into 2 parts called RAW and COLUMN. These parts share the common address bus with 13 pins. Totally it provides 25 (minus precharge signal) bits for address, where only part is used on 32MB chip. 64MB chip adds one more column bit and still have many unused bits in column. So, 64MB chips can be used as 32MB chips without any changes in schematics. 128MB chip also could be compatible, but i never heard about 128MB chips in this package.
I didn't explore the possibility of adding an extra chips without using additional pins. Probably it's possible by using a simple logic to hook on column part of address.
JimDrew wrote:However, if there is no assurance of supporting more memory in cores, then it would be a waste of my time. I wouldn't just go making a new IO board without it being something that people have been asking for, so not everyone shares your opinion of making this as cheap as possible.
JimDrew wrote:I don't work that way. I am not going to spend thousands of dollars on designing and producing a new IO board w/SDRAM if it is not going to be supported.
Sorgelig wrote:You didn't even start to create but already want guarantees of success.
JimDrew wrote:Hardware design is what I have done for decades. I have created more than 300 commercial products (millions of circuit boards), spanning many different industries and I have a long history (1979 to present) in the Commodore/Amiga industry.
JimDrew wrote: I would hate to make a board with more RAM that is higher priced but can't sell because there is no support (chicken and egg thing).
JimDrew wrote:No matter how the RAM works out, it will be on the IO board itself, not as a separate daughter board. I would put the RTC on the IO board as if there were I2C pins available from the I/O expander.
JimDrew wrote:What SRAM is 10ns?
JimDrew wrote:Why do you actually need SRAM vs. SDRAM?
Users browsing this forum: leejsmith and 4 guests