Please Apollo team, a Mini PCIE slot is a MUST have for future Vampire

[Codetsu]

please do not add any new stuff !
just do that damn thing for any who need one !

[matthey]

Originally Posted by wawa

i think you are wholesale leaving the human factor as entity out of the equation. im actually surprised and very pleased by the continuity of effort gunnars and majsta projects are providing. hardly anything in amiga land can concure with this. i wouldnt be able to fulfill such a duty. i admire that!

I'm not blaming Gunnar or Majsta but rather the people holding the Amiga hostage. There are some 50 billion ARM processors produced every year. The Raspberry Pi has sold over 10 million units (I want a piece of that pie). A single Natami thread had over .75 million views (Amiga interest with no advertising!). A-EON has managed to sell low thousands of Amigas while claiming to preserve the Amiga? No, this is just a slow and guaranteed Amiga Death as they regurgitate the same failed ideas while they block everyone else. If they won't take the risk to mass produce halfway affordable Amigas which Amiga users want then they should pass on the technology to someone who will instead of waiting until they are bankrupt to pass it on (if they really cared like they pretend to). We know Amiga Inc. is a few con-men and don't expect anything from them but A-EON and Cloanto are a bunch of lamers too .

Originally Posted by wawa

pretty sure amiga users out there are online, one way or the other. these who are not, dont count anyway in today terms. its not a question of having means being online (with amiga hardware), but to engage with the common cause, which is a different matter to discuss.

I expect 90% use a "PC" for the internet.

Originally Posted by grelbfarlk

Well how about this, same idea basically but make it Mini-PCI slot to use as a PCI-to-PCI bridge for future Amigas.

If or when the big box version of Vampire comes out, you could use the Mini-PCI slot to plug into the bridge slot of a Mediator. Then we get the benefits of full PCI speed to the Mediator slots instead of the slow PCI->Zorro way it is now. Then we get fast-ish access to Voodoo or Radeon RAM or PCI HDD controllers or you know any of what we thought would be fast on Mediator but isn't really at all, fast that is.

This is a good idea. I wonder if a cable could connect a mini-PCI slot on an FPGA accelerator card to an existing PCI slot in Mediators? It would probably have to support bus mastering and I'd be a little weary of the cabling. The old mini-PCI shouldn't need a more expensive FPGA with SerDes and has several times the bandwidth of Zorro II/III.

Originally Posted by grond

I used to think similarly but you really get used to the RTG capability and online-access to aminet or media and code storage on your PC in no time at all...

Yep. I would probably give up my Amiga without relatively fast CPU, fast gfx, fast drives and the internet. As much as I love the Amiga, I'm not going back to 68000, ECS, floppies and a 1084 monitor.

Originally Posted by Juz400

Again there is a wealth of Mini-PCI laptop network cards and other bits that are landfill nowadays.
The reason I signed up with the forum(After lurking many YEARS) is that the technology is already to go with some FPGA and that the PHYSICAL size or `Real Estate` the Mini-PCIE slot will consume on the PCB is as small as possible with the MAXIMUM potential.

Even a single lane PCIe slot has huge bandwidth compared to any classic Amiga hardware (it is even adequate for most slow AmigaOS 4 hardware sadly). PCIe would still need those SerDes from a more expensive FPGA but that is likely affordable for big box and 1200 accelerators.

Originally Posted by Juz400

The Trend is to build smaller with the same or more power.

Yep. Back to doing more with less. This is what the 68k Amiga was good at and what most Amiga users have been doing since the '90s.

Originally Posted by Juz400

The Amiga wasnt going to be quite so Amiga when they make the move to the Hombre chipset being RISC not 68K but something improved.

Did you read the whole thread I linked? C= was planning a non-Amiga Hombre 64 bit game box but they also had plans to use the Hombre chip set for 3D gfx (card or integrated). They were planning to have it compatible with the Amiga and they were also planning an Amiga SoC (68k+custom chips). C= would have liked to continue both the Amiga and 68k but the 68k was more difficult as Motorola was sabotaging the 68k line as they tried to convert everyone to PPC and they were not known for making friendly licensing deals of their technology. C= seemed to aim too low in CPU specs also but some of that may have been Motorola's marketing and premiums they demanded on their chips. While it was unlikely that C= would have ended up with a 68060+Amiga SoC, it could have been competitive in performance/watt for over a decade for mid then low end retro gaming and embedded applications (set top boxes, kiosks, etc.). The 68060 had similar performance to the Pentium while using ~42% less power (5.5W vs 9.5W) and 32% less transistors (2.5mil vs 3.3mil). The 68060 could get away with 1/2 the memory of RISC for a further energy efficiency and cost advantage. A somewhat similar to the Pentium in order CISC Atom with the energy efficiency advantage the 68060 had compared to the Pentium would be outperforming the best Cortex ARM processors in integer performance/watt/core today.

[Pat the Cat]

CBM's take on "compatibiility" was gibberish.

I disagree with you on CD64 Vs Playstation 1 technicality, but I do agree with you that, with CBM trying to "manage" it, Hombre was screwed anyway.

HP did do SOC with their first SIMD capable systems but with CBM's reputation, there was zero chance of it being an Amiga SOC.

[Juz400]

Originally Posted by matthey

Even a single lane PCIe slot has huge bandwidth compared to any classic Amiga hardware (it is even adequate for most slow AmigaOS 4 hardware sadly). PCIe would still need those SerDes from a more expensive FPGA but that is likely affordable for big box and 1200 accelerators.

Looking at a GTX1070 benchmark via one of these devices shows that at x1 speeds you can still push an awful lot of data through PCI-E a x16 wont be fully maxed out for years yet.

Moving the topic along a bit, I have had a look at this doc from the AMD Developer Guides

http://amd-dev.wpengine.netdna-cdn.c...ation_v1.5.pdf

Its quite interesting reading from page 18 onwards.

You can either have the CPU push the data into the cards CP(Command Processor) or point it to a memory area and tell it to cycle through(pull) the data until you tell it to stop

You kind of set up a `Copper List` of instructions for it to follow, I wonder where they got that idea?

There is a list of Comand Processor registers from page 147 onwards.

Would a SerDes have access to the memory onboard the Vampire card?
So you could say have an interupt signal in your code that instructs the cards CP to use pull mode and grab a list of instructions from memory then wait until your next interupt
If this is the case, then it is not imposible to get the card to do something.

The programming of the shaders and vertex units looks very complicated and getting a `graphics card driver` up and running would be quite the labour of love.

But if you look at the demo scene, they get a picture and sound out of almost anything you can solder together, program and wire up to an AV system.
Hardware banging the card just to see what happens could be quite fun for the more adventurous programmers out there
It could be 1987 all over again with a new copperlist and blitter registers to figure out!

[matthey]

Originally Posted by Juz400

Looking at a GTX1070 benchmark via one of these devices shows that at x1 speeds you can still push an awful lot of data through PCI-E a x16 wont be fully maxed out for years yet.

The slowest PCIe 1x is 250MB/s where the fastest common 32 bit PCI@66MHz is 266MB/s (There is 64 bit PCI but it is rare). Each additional PCIe lane doubles the throughput and newer spec versions are significantly faster as well.

Originally Posted by Juz400

Would a SerDes have access to the memory onboard the Vampire card?
So you could say have an interrupt signal in your code that instructs the cards CP to use pull mode and grab a list of instructions from memory then wait until your next interrupt
If this is the case, then it is not impossible to get the card to do something.

An FPGA is how it is wired (programmed) internally. Lower end FPGAs do not have a memory controller so they must be created in the FPGA. The Altera Cyclone V (as I recall) has a memory controller so I expect most mid to high end FPGAs do. SerDes are only in mid-high end FPGAs so I expect these FPGAs have a memory controller also. I expect a memory controller would simplify transfers to/from the SerDes or at least save complex logic in the FPGA which would do it. The SerDes would be severely handicapped without access to memory although most FPGAs have SRAM (perhaps used as a buffer for the SerDes?). I read a little about the SerDes in an FPGA once but this in not my area as you can see. If I understand what you want to do correctly then I expect it is possible with proper FPGA programming and an FPGA with SerDes probably has tools to make it easier.

Originally Posted by Juz400

The programming of the shaders and vertex units looks very complicated and getting a `graphics card driver` up and running would be quite the labour of love.

But if you look at the demo scene, they get a picture and sound out of almost anything you can solder together, program and wire up to an AV system. Hardware banging the card just to see what happens could be quite fun for the more adventurous programmers out there. It could be 1987 all over again with a new copperlist and blitter registers to figure out!

I modified the Voodoo 3-4 Warp3D drivers for my Mediator (mainly to fix bugs and horrible optimization at first). This card has the best gfx hardware documentation available (still not for beginner programmers) and is fairly easy to program. More modern cards are generally much more difficult to program. Yes, programmers could program various gfx cards at the bare metal (educational) but I think this is the wrong philosophy. IMO, the Amiga would benefit from standardization with one 3D hardware (upgradeable but similar design) and programmers should choose to use the OS and drivers provided which would better utilize multitasking, SMP and GPU parallelism. This would also allow the gfx to be integrated on the motherboard instead of transferring data across relatively slow buses. This would make the Amiga more like a console with efficient use of hardware to make up for affordable hardware with less performance. Modern gfx cards are a big hindrance to small designs because of the heat they produce so it is important to do more with less like the Amiga used to be good at. Sure, it would be nice to have optimal and bug free drivers for hundreds of gfx cards but I believe this is a less feasible and less practical solution for the Amiga.

[pandy71]

RPi3 can be used as graphic card for Amiga - USB interface with SPI and MIPI should be sufficiently fast for Amiga (flickerfixer trough MIPI), adding to this most complete (for now) documentation for GPU (video encoder/decoder still not documented), moderate power consumption, price - it will be very hard to beat in FPGA... Side to this Amiga can be beneficial from RPi community work (GPGPU).

[plasmab]

Like seriously... dont beg for features from the board designers. we already try to cram on as much as we can for the money.

[Overflow]

If the dreamers go too far, Im sure BigGun just do the cpt Picard facepalm move and then laugh.

I dont see the problem if people going off the deep end on forums, since developers can just ignore it, or simply say "NO, not going to happen!".

[plasmab]

Originally Posted by Overflow

If the dreamers go too far, Im sure BigGun just do the cpt Picard facepalm move and then laugh.

I dont see the problem if people going off the deep end on forums, since developers can just ignore it, or simply say "NO, not going to happen!".

Because it makes us rage quit eventually. Thats why.

[eXeler0]

Originally Posted by plasmab

Because it makes us rage quit eventually. Thats why.

Most people mean well and are enthusiastic, but the will to "participate" and add their own wishes or opinions will always be there in internet forums even when no one asked for it ;-)
As with life in general, you draw strength from the positives and ignore the negatives.

Cheers

[wawa]

Originally Posted by plasmab

Because it makes us rage quit eventually. Thats why.

if igor and gunnar were so easy to upset, that would have happened already. luckily not. that said, all these sensless proposals, usually insisted on by people who dont intend to bend a finger, actually are an annoyance. good thing they have an own forum.

[grelbfarlk]

I just think having a somewhat fast, expandable, cheap bus would be nice for big box Amigas. Right now we have the option of the sort of fast, very expensive and rare Grex or the slow (to motherboard and accelerator slot), reasonably priced and common Mediator.

Or maybe it's just too hard, whether it's from SPI to a PCI bridge or a PCI compactPCI controller on the Vampire with some ribbon cables. Don't blame me it's a wish fulfilling thread.

[matthey]

Originally Posted by plasmab

Because it makes us rage quit eventually. Thats why.

Get a punching bag. Seriously though, you are blessed that Amiga people are interested in what you do. I'm lucky if I get any input on my projects and I'm usually ignored.

Originally Posted by grelbfarlk

I just think having a somewhat fast, expandable, cheap bus would be nice for big box Amigas. Right now we have the option of the sort of fast, very expensive and rare Grex or the slow (to motherboard and accelerator slot), reasonably priced and common Mediator.

You have a good and reasonable idea. A fast bus with a gfx card would be many times faster than an FPGA 64 bit SIMD without floating point for 3D. The added cost would likely be less than double (and still affordable for a big box accelerator card). An FPGA with SerDes could be used for SATA also. I don't see any big dreaming here. Actually, it seems more reasonable than creating and using a CPU in an FPGA for a "personal computer".

[kolla]

USB would be much more useful.

[Locutus]

Would prefer Fibrechannel

[michaelz]

I would like thunderbolt support.

[Samurai_Crow]

The developments of the Vampire team have mostly shifted from the CPU development to SAGA core at this point. Vampire supplies an SoC design like the RasPi3 thus a second one is unnecessary. SPI is available on the V500+. The stand alone units will not initially be card slot boxes AFAIK.

[Anakirob]

Correct me if I'm wrong but my understanding is:
SAGA = 24-bit modern HDMI display
PC/HW = 24-bit modern HDMI display

Why would you need two sets of display hardware that do the same thing?

Also
SAGA = relatively Amiga compatible, uses current GFX card drivers
PC GPU = not really Amiga compatible would need drivers for each different card model.

Why would Apollo team even bother with modern PC display hardware adaptors when their current method is functional and effective where an adaptor to allow an old PC GFX card is not?

[matthey]

Originally Posted by Anakirob

Why would Apollo team even bother with modern PC display hardware adaptors when their current method is functional and effective where an adaptor to allow an old PC GFX card is not?

Reasons for an adaptor.

1) GPU performance in an FPGA can not match the GPU performance of even a semi-modern gfx card.

2) A gfx card offers more parallel processing power.

3) SAGA currently has no support for 3D.

Reasons against an adapter.

1) Cost.

2) Transferring data through a bus is slower and increases latency.

3) Standardized integrated Amiga 3D gfx may be preferable to add in a FPGA SoC later, especially if planning an ASIC.

[Anakirob]

Originally Posted by matthey

Reasons for an adaptor.

1) GPU performance in an FPGA can not match the GPU performance of even a semi-modern gfx card.

2) A gfx card offers more parallel processing power.

3) SAGA currently has no support for 3D.

Reasons against an adapter.

1) Cost.

2) Transferring data through a bus is slower and increases latency.

3) Standardized integrated Amiga 3D gfx may be preferable to add in a FPGA SoC later, especially if planning an ASIC.

Thanks, that explains it pretty well. There are a lot of long and elaborate posts here and the answers were generally not easy to read. One question...

Could a modern GPU somehow be added for 3D acceleration and other features without actually using it as a display output device? I think that's what you mean by "parallel processing power", but I'm not 100% sure.

I can see how that might be a desirable feature.

and one more question...
But then how many Amiga titles actually make use of hardware 3D acceleration anyway?