@Channel_Z

Quote:

Do not be fooled by the fact that the video shows an OCS game - in fact it was discussed yesterday if the video should not be officially released because of the reactions to this.

There's only guy here that has given this reaction. Be aware that there will always be negative reactions, especially in the Amiga scene. You can't please them all nor should you want that I guess...

_________________
A4000D Cyberstorm PPC 060@50 604@200 128Mb G-Rex Voodoo3 PicassoIV Ariadne Delfina Lite OS3.9 BB2

Online Flash version of BoulderDash: Temporarily offline...

@EmperorLongo

Quote:

Natami is one of the most exciting Amiga projects, IMO.

I agree. HIgh spec classic gear is *expensive* and unreliable. I haven't bought any new gear in quite awhile because it's a gamble as to whether or not I will have to spend a good amount of time tracing down problems and doing repairs. New classic hardware is sorely needed.

_________________

@bitman

Good work!

@Crumb

Quote:

Crumb wrote: If they had taken minimig custom chip emulation Natami would probably be in a more advanced state.

That's exactly the point I wanted to express, Natami is a very ambitious project, if they want to implement their SuperAGA they have to do more than just make a working copy of AGA and add some enhancements. Modern DRAM gets its speed when it is accessed in burst mode, so you need caches for every non burst access. thats one of the parts that slows down the Natami development. I have posted in other threads why I think Natami's SuperAGA is not worth the extra work.

So if there are people looking for a slight advanced new and cheap retro Amiga, they should try to improve the Minimig. set up a bounty system for development of missing parts, 030/040 cards, PCI bus, USB, Ethernet, GFX drivers, OS improvement ... everything is possible.
Speak with Amiga dealer convince them to invets into the production for new Minimigs. Don't wait for a wonder, the community could control the development of a new Amiga hardware.

cu

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@Flashlab

Quote:

Flashlab wrote: @Channel_Z There's only guy here that has given this reaction. Be aware that there will always be negative reactions, especially in the Amiga scene. .

A negative reaction ? Who has deleted that post ? I have never see a negative reaction about Natami in this threat.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@Channel_Z

Quote:

No, two years ago most of the listed features were up and running in the Natami030 prototype and shown publicly at MEKA 2008. The time up till now has been spent by the chief hardware designer designing a standalone board design and adapting those features to make use of DDR2 memory. Just like vidarh says, "implemented" means "implemented on the Natami LX board".

thanks for the clarification. What memory controllers do fpgaarcade or Altera DE2 use?

I saw CloneA chipset running Arte at Breakpoint2006, it was impressive.

Quote:

The board in the video is using the 68060 CPU card. The softcore is being developed by other developers who are IBM employees, and the core is very close to being fully usable, as it is being optimised at the moment.

That details are interesting but as it was not detailed in the first post it was not evident to most of users, thanks.

Quote:

The goals of the Minimig and FPGA Arcade are to recreate the Amiga chipset. The goals of the Natami project is to further develop the Amiga chipset, and that is what the design is having headroom for.

The difference right now is that Minimig and FPGA Arcade evolution and improvements are better published (everyone right now has watched the public videos of the minimig core with 54MB of chipmem running AGA games) and we did not have many evidencies of the existence of Natami (apart from some photographs at amiga.org of a c64-one with a 030 full of wires)

BTW, minimig with 54MB of chipmem is certainly an improvement over standard chipset. Yaqube wrote that he had interest in connecting a real 060 and who knows about further improvements...

SuperAGA sounds nice but it would be nicer if Natami team published videos more often so amigans could see the progress.

Quote:

Do not be fooled by the fact that the video shows an OCS game - in fact it was discussed yesterday if the video should not be officially released because of the reactions to this.

Maybe running some AGA 060 hungry demos on Natami would have been a better idea. "Starstruck" or "we come in peace" would be good examples and would really show if 060ram controller is at least as good as old Cyberstorm ones.

to the Natami Team:
good job! we want more videos!

Last edited by Crumb on 11-May-2010 at 09:32 PM.
Last edited by Crumb on 11-May-2010 at 09:32 PM.

_________________
The only spanish amiga news web page/club: CUAZ

@thread

Natami reached Minimig performance 2 years ago, so during that time lots of enhancments were made to it.

For example, to reduce costs and increase memory performance DDR memory had to be used. The Amiga was not designed with this type of memory so adjustments were made and a new memory controller was created.

All of the Natami motherboard has been designed for high performance and also expandibility.

The Natami is not just an Amiga designe that will have extra features added without consideration for futureproof or performance.

The Natami is designed with most extra features like 3D acceleration in mind at the hardware level to be the most powerful Amiga ever made.

The aim is to have Playstation 2 level performance.

The video shows how compatible the chipset is and is the first step in testing OCS chipset. Once alot of games have been tested with the new memory controller and blitter only then can AGA testing begin.

Baby steps is what is needed and once the testing is done then the high performance blitter and 3D features can be demonstrated in full 24 bit hires screenmodes.

remember the estimated performance for natami is x150 - x200 cpu speed of an A600 and a x50 DMA performance.

Once the new 68050 cpu is finished and software optimized for this we will se outstanding performance.

We will finally have a decent system to make hi quality games and applications on common hardware, fully accessible to bang the hardware if you wish in 24bit glory.

The Natami SuperAGA chipset is more powerful that AAA and the Hombre chipsets by a margin !

As the video shows the hardware is working and compatible and the long development has paid off, because Amiga is special and deserves the best performance we can all be proud of - custom hardware developed from scratch not only to run AmigaOS, but to take advantage, assist and accelerate AmigaOS the way AmigaOS actually functions.

Believe me happy times ahead !!!

Last edited by asymetrix on 12-May-2010 at 12:23 AM.

_________________
Download 499.26 Mbps, 659.94 Mbps Upload :)

@asymetrix

Why are people posting with good news here in amigaworld.net is beyond my understanding It is bound to be attacked with spears and flying axes.

@Foody

Quote:

Why are people posting with good news here in amigaworld.net is beyond my understanding It is bound to be attacked with spears and flying axes.

"You know you got to go through hell before you get to heaven"

"Jet Airliner", Steve Miller Band - 1977



Plaz

@Magic

Quote:

Not being picky but, would be nice to see the machine with it on in the background. All we know that could be an a500 or something running it lol.

I don't think so. If you follow the forums and see the progress being made, this is the real deal. Sad you would think otherwise.

Notice the red bar flickering on the side; Thomas (and the rest of the Natami Team) are working hard on it, still bugs and things to iron out.

Good luck to them, and will definitely buy one when Natami is released.

Q!

"i am home"

@Arko

You seem at pains in this thread and the thread last month to explain why you think the Natami is too ambitious, why a souped-up Minimig is a better idea, why RTG is better than Super-AGA, etc.

You mention difficulties with DDR2 and bursting memory. Well, they've already done that and it can use modern RAM more efficiently. Yes it caused a delay, but it's done.

The Natami team have been completely open with people right from the beginning (and how often is that the case in this community?), posted continuous status updates, screenshots, demoed the previous prototype live and now shown a video of the new prototype.

They've already shown that they have the ability to build this thing by coming so far with it. Difficult things take time.

Sorry to anyone who believes we deserve more from them - we don't.

Chris

Also people forget that the long term goal is to create a board using ASICs not fpgas.

Fpgas are being used for "development". So again, this is where the project differs from Minimig and ReplayArcade.

Also, the "delay" of using DDR2 is irrelevant when compared to the bus speed of the classic hardware and this new hardware. The memory controller is already done and you can clearly see: works. The advantage of DDR2 is requesting 8 bytes is equally as fast as requesting 1 byte from RAM. Once new software is written taking into account the burst transfer speed, then you will see a new level of performance not possible on the Minimig and Replay.

But again, they are for different markets. The Minimig will probably end up as a built-in Amiga-on-a-joystiq. The Replay board is a hardware MAME not dedicated to being an "Amiga" board. Natami is a full 68k PC platform for new developments and a future.

The fact that the Replay board is now using a better fpga that originally planned is simply a matter of economies. Over time, electronics become cheaper. It's great that the Minimig can become faster over time, but no new features are being added outside of faster processing. SAGA will exceed Hombre in performance and features. Again, this is where the projects differ.

Last edited by Lou on 12-May-2010 at 12:59 PM.

@Lou

Quote:

Also, the "delay" of using DDR2 is irrelevant when compared to the bus speed of the classic hardware and this new hardware. The memory controller is already done and you can clearly see: works

@Clebin
Quote:

You mention difficulties with DDR2 and bursting memory. Well, they've already done that and it can use modern RAM more efficiently. Yes it caused a delay, but it's done.

According to the information on the thread, it is not fully done and needed to be reworked, it is not ready.

If the Natami gets available during this year you will have the official right to bring up this thread again and call me a clueless loser.

But until today the Minimg looks more promising to me than any Natami.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@Arko

Quote:

Arko wrote: @Lou Quote: Also, the "delay" of using DDR2 is irrelevant when compared to the bus speed of the classic hardware and this new hardware. The memory controller is already done and you can clearly see: works @Clebin Quote: You mention difficulties with DDR2 and bursting memory. Well, they've already done that and it can use modern RAM more efficiently. Yes it caused a delay, but it's done. According to the information on the thread, it is not fully done and needed to be reworked, it is not ready. If the Natami gets available during this year you will have the official right to bring up this thread again and call me a clueless loser. But until today the Minimg looks more promising to me than any Natami.

Did you not see Giani Sisters running? If the cpu can't access memory, then nothing would work. 1+1=2.

The Natami 'may' be available as a developer's board this year but that doesn't mean SAGA or Robin might be finished this year. I think they could probably use Replay(fpgarcade) to test and finish the '050 design and perhaps test Robin as well. This would allow multiple people access to hardware that they can test on.

If you want something today to replace an A500, get a Minimig. Even the Replay won't be available to a couple of months as MikeJ is looking for component vendors. He is on his way to China to find them and pick up the un-populated boards. Atleast that will give you an A1200 replacement in a few months.

If you want something beyond an A4000, realistically, expect to wait longer than 6 months for a developer board and a year for a consumer NATAMI board. I think that when/if a Replay board is available soon, that it will benefit the NATAMI team with regards to testing there custom chips.

If you can't wait, you know what your options are and there is no point in repeating them again and again.

Quote:

Did you not see Giani Sisters running? If the cpu can't access memory, then nothing would work. 1+1=2.

1+1 != Apples + Pears

You don't need the burst access from the CPU to the RAM to get Giana Sisters running.
You don't need the burst access from the GFX hardware to the RAM to get Giana Sisters running.

According to the information on the thread, it is not fully done and needed to be reworked, it is not ready.

If you want to call the Natami team liars than dio it.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@Arko

As (EDIT) Lou says, you can see with your own eyes that the board is working!

If it was 100% complete and there were no bugs, it would be off to manufacturing already. I don't know the details of every little job they have left, nor would I pester them for such a flimsy reason. But here's post from Gunnar a few months ago explaining the reasoning:

Quote:

Natami DDR2-RAM is programmed for a 4-burst access : correct DDR2 delivers 4 databits per pin: DDR2 does 2 per pin. The 4-burst access accounts for the 2 transfer of DDR2 already. This mean: a 16bit memory interface does 64bit bursts. a 32bit memory interface does 128bit bursts. I hope this helps. But in general you were right. The original AMIGA did do single memory access and did wait for it to fully complete until doing the next single memory access. The Natami originally behaved 100% the same. The new Natami changes the memory access in two ways. a) It always burst. Every read access is done in burst lines. Writes access also burst but the Natami can mask individual bytes - therefore its possible to write single bytes or words as usual. b) The Natami holds several access in flight. To improve memory performance the NATAMI does not do a single memory access as the old AMIGA but can hold up to 8 burst access in flight per chip/fast bank. The concept of keeping multiple burst in flight is crucial for state of the art memory performance. This required some significant enhancements to AGNUS.

Clearly, this is a good thing, and much work has been done.

Also, this from Thomas which shows that Chip-RAM access is already working. I'm sorry if you want to hang your criticisms on one bullet-point about Fast-RAM, but don't you think that's a bit churlish?

Quote:

It was very hard work for the last two weeks. There was a very hard to find bug in the chip-ram memory management which let the DMA read data (in very few circumstances) be unreliable. So I can not show a new picture of something more working - I can only state that chip-ram accesses are stable now. Which is of course a basic feature. Second I figured out that the copper memory interface needs also to be adjusted to new DDR2 memory interface of Natami. It still had the "old" C-one prototype interface. With that an instruction fetch took at least 28 cycles. This might be tolerable for WB but for games and demos it is not. Usually a copperlist is used to wait for some dedicated location on screen and then alter (for example) bitplane pointers. Here the 28 cycle delay per instruction will cause a messed up screen as bitplane pointers need to be altered as a group. So I introduced a short buffer which prefetches instructions. With that the copper does not need to wait for instructions AFTER the wait command. Unfortunately implementing this buffer sounds much easier than implementing it because not only the CPU but also the copper itself can alter the copper PC.

So.... they are fixing bugs and bits are starting to work on the developer board. If there were no bugs, this would be off to manufacturing already. What more can anyone ask?

Quote:

If the Natami gets available during this year you will have the official right to bring up this thread again and call me a clueless loser.

2010 is your timescale - no-one else's. Will Natami be a failure if it is released in 2011? The immense progress over every other piece of classic hardware ever made - will that become null and void because they missed your made-up timescale? This isn't a great attitude to have.

Chris

Last edited by clebin on 12-May-2010 at 03:03 PM.

@Arko

Quote:

If you want to call the Natami team liars than dio it.

Good one. I bet the Natami folks are right behind you.

You're doing valuable work in informing everyone of their folly.

Chris

@Arko

Quote:

Arko wrote: Quote: Did you not see Giani Sisters running? If the cpu can't access memory, then nothing would work. 1+1=2. 1+1 != Apples + Pears You don't need the burst access from the CPU to the RAM to get Giana Sisters running. You don't need the burst access from the GFX hardware to the RAM to get Giana Sisters running.

You would be 100% correct if we were talking about classic Amiga hardware.

Quote:

According to the information on the thread, it is not fully done and needed to be reworked, it is not ready. If you want to call the Natami team liars than dio it.

After what Clebin quoted to you, I can only say: /facepalm

The board is using DDR2 memory exclusively. The memory controller was reworked to do burst mode 100% of the time. Once NEW software is written that will request 8 bytes from 32bit memory at a time, you will see new software do things even faster than the marginal gains displayed on the fpgarcade.

Buy a Minimig today.
Buy a Replay in a few months.
Buy a NATAMI in the future.

Last edited by Lou on 12-May-2010 at 03:31 PM.
Last edited by Lou on 12-May-2010 at 03:29 PM.

@Lou

Quote:

I think that when/if a Replay board is available soon, that it will benefit the NATAMI team with regards to testing there custom chips.

The replay looks nice but its no use for us at all.
The NATAMI chips are to big to fit in the replay.
And also the replay does not match the NATAMI architecture memory architecture.

But this is no problem as we have our own boards to test our chips.

_________________
APOLLO the new 68K : www.apollo-core.com

@BigGun

Quote:

BigGun wrote: @Lou Quote: I think that when/if a Replay board is available soon, that it will benefit the NATAMI team with regards to testing there custom chips. The replay looks nice but its no use for us at all. The NATAMI chips are to big to fit in the replay. And also the replay does not match the NATAMI architecture memory architecture. But this is no problem as we have our own boards to test our chips.

From a thread on Amiga.org, MikeJ said:
Quote:

I've got a good deal on the Xilinx - the boards will contain the larger device (1600e).

Is that still too small? I was only suggesting it for testing the actual cpu core and perhaps Robin functionality for the people working 'overseas' nothing more. :)