However, is it possible that on the very same, first NatAmi, if someone buys the same FPGA that is being used, but it is 60 MHz (or more) faster than the FPGA being used, that it COULD  emit a 1920*1080 HDTV signal?????

Is the NatAmi motherboard built such that this is the only thing keeping HDTV from being available?

For me, even if HDTV crawls, it WOULD be worth having as an output option, and I believe that 2 or 4 colour screens may be even updatable at acceptable speeds.

Not possible.
The FPGA on the Natami is the FASTEST of its family already.
There is no faster FPGA on the whole world that you could put on the board.

Besides the FPGA is soldered to the board. You can not exchange it without high risk to trash your board. The chip many hundreds of pins with less than a milimeter size each. You really don't want to resolder this stuff.

What I don't understand is why you would want to have HDTV output with 2 colors. Maybe you can explain me this?

Okay so a direct replacment isn't a option, since none of the others are faster.(bigger with the same pinout?) :/

Sounds like a challenge to solder, i have done such typ of rework before.
Then again i think i like the idea that when it's broke i could fix it.

Bigger FPGA that have the same number of pins need more voltage supply pins - that means they have less IOS free.
This means upgrading the chip is only possible if the board was designed this way in before hand.

Its really difficult to manually solfer an FPGA.
And the chance that you run your board or the FPGA is very high.
And as bigger FPGA cost a few hundrd $, toasting it on this try is no fun.

Back on topic I think in 2D the FPGA should be capable, only thing i can think of is that the isn't enough bandwidth to fill the screen fast enough in a reasonable time. (22 or less FPS):(
When this happens you'll have a bigger problem in no time.
Namely the audio would be out of sync with the film.

Another problem one might call the lack of a SATA what is needed for a modern Blueray or HDDVD drive.
I am serious thinking that the transfer rate between disk and memory and later to the chipset would bring the system down to a snails pase.
It's not just the Chipset it's also IDE that can't 133Mb/s (can't remember if those where bits or bytes)
Second DDR the clk would be 100MHz what would make the effective working speed 200MHz in burst mode this would give us 32Bit?
then read and writes can't be done at the same time, there is latency (most of which the team could work away, if not all).
okay we have 2 memory busses so we could read one and write the other, this could be done by the DMAC, Blitter and CPU.

Does everyone get the image?
The reason your HD video, hd-dvd, dvd, hdd, blueray player can do with a CPU that isn't so fast is that it's a single purpose machine.
Besides they have specialized DSP's and what else not.

The NatAmi is a multi purpose machine, to let a HD resolution film play on it would require most if not all HW resources the system can give.
BTW over a few years we are talking about terabyte disks since they are adding 2 more dimentions to the DVD standard, that give 5 dimentions of data or 14Tb, I can't understand why ppl want something that would be replaced soon anyhow.

And BTW with a DVI connector nothing stops you from connecting to a HDTV.
On another hand 2 interesting facts on DVI, you can pull I2C off it and Audio (if it's connected to it)

Okay, here goes.

I figure that it's stated, that when an Amiga uses a lower amount of bit planes, it works faster.

So, I thought that there may have been a faster FPGA, maybe $200 more, that could be used instead that was say 80 MHz faster, that could if used do HDTV resolution output.

Someone (more than likely me) would have been willing to buy the faster FPGA to show off to people the potential of NatAmi.

I couldn't care less if HDTV resolution was dog slow, it could at least do a slide show in full colour, were a faster FPGA available.

And, in 2 or 4 colours, it probably would have been fast enough to scroll text and display data. Read email and do word processing at least.

For me $200 or a bit more would have been worth experimenting with to see what might be possible.

I know you're not saying I said that, but just for the record, I have never ever said it could play at that resolution, any amount of colours video, or high frame rate video game. Not once, not here, or anywhere else.

I know that standard DVD (MPEG2) playback should be possible, and then if it's ONLY doing that one task probably, but I never considered achieving H.264(??) with a 150 MHz CPU.

Achronix Speedster FPGAs (1.5 GHz, now sampling at 1.9 Ghz)

EXTERNAL LINK 
bye,
Klaus

Then please DON'T use the term HDTV.

People that use the term HDTV - mean playing HDTV cinema movies with 60fps - and they also think of encoding h264 videos in HDTV resolution.

If you want to run workbench screen in this resolution doing test editing then you talk about something completely different.

To say it clearly HDTV implies watching movies on a HDTV-TVset.

Naïve question:
Is a pure numeric DVI interface will be less expensive (no RAMDAC) and simplify the board ?
A second analog connector can be keeped for NTSC/PAL display only (a "slow" RAMDAC).

From here: EXTERNAL LINK 

Why because those are clock speeds used for memory, and if your bus works 1:1 on memory and the rest of the chipset, it's golden IMHO.
And using and any other setting then 1:1 increases you latency and this would make a system needed to wait while it's running something, that isn't ideal.

Another thing, HW can't go faster then the code that is written for it.
And this goes double for FPGA's, your working with a softcore code and there is a limit to the amount of gates one could use for free.

And then there's software, i don't want to know the amount of nops running on my PC.

Speed isn't everything, Speed sucks if the various timings in a system don't combine smoothly like having one part running 1/3 and another running 1/4 to one component, when will all 3 share a clock?

I fully accept blame for this misunderstanding, however, never had it crossed my mind that that was what people thought when I kept bringing it up in the past. And mainly I didn't think that people thought that because I was usually qualifying my statements with "even though I know it would be slow, we still could do things such as email and word processing".

Gunnar von Boehn wrote:

If you want to run workbench screen in this resolution doing test editing then you talk about something completely different.

This is all I ever had in mind, and even thought that reasonable speed vector graphics games in 8 colours were even possible at 1920*1080 on a slightly faster NatAmi.

I had no idea that the FPGA you went with was the biggest/fastest of it's class. As I said, I thought adventurous persons with deep pockets and faith could go out on a limb and really show off the possible potential of advancing NatAmi further than even you had dared to do.

Someone out there would have been willing to bite.

Gunnar von Boehn wrote:

To say it clearly HDTV implies watching movies on a HDTV-TVset.

Now I know better.

Would you be willing to email them and ask about pricing of 333, 400, and 533 MHz compatible FPGAs that they may have available?

I hope that somebody could inquire if they have something that could benefit our final goal.

Natami production runs will NOT classify as volume production!
This means we should expect at least double or triple of this value.
 
 
Looking at their "data sheets" the speedster looks interesting.
 
But some things are odd. :-/
The company shows zero examples of any real clock rates they achieve with certain types of logic.
ALTERA and XYLINX always proudly provide real examples like "code yxz" runs with "yxz" MHz in Device xyz.
 
That Achronix Speedster claims to be so much faster but does not provide any examples at all is quite odd.
 
Also odd is that they announce the chip and development tools but you can neither download not order any of them on their site.
You can download development tools for Xylinx, ALTERA, and Lattice of course.

And Achronix seemt to be a very tiny company - somewhere on their website they say that they are about 50 people.
ALTERA and XILINX have both several thousand employees.
 
While this Speedster market paper looks nice on the first glance - I'm find the company not fully convinced so far.
 
I would prefer that the Natami is build on top of something reliable -  A proven and fully tested FPGA platform.
I don't want us to be genuine pigs trying out some totally new product.
 
I say lets wait until someone else has a working desing with this chip. If this Speedster is really working and good - then we can always build a next generation based on this next year.

If they're for real, maybe THEY need US, more than we need them (them being a questionable party).

Amiga is a PROVEN property.

THEY need publicity as they don't seem to carry any water at the moment.

Maybe they could send 3 FPGAs for NatAmi to try out??? :-)

I say, contact them.

If it comes to nothing, it's HAIR LOSS. HAH!

Marcel Verdaasdonk wrote:

and loic i have a question to you are your figure bit or byte?

Sorry for that, in france we use octets(meaning group of 8) for Bytes and bits for bits.
So i'm not used to use the CAPITAL B for bytes.
So the number were ~150MiB and ~133MiB if i want to be 100% politicaly correct :-)