Vampire discourse, keep it civil (was: Vampire 1200 V2 waiting times)

[Chucky]

Maybe handle it as a COMPANY and not as several private persons would be a. well. thing? then licensing etc would not be an issue.. but as it seems to be perfectly normal for them to handle all like private persons. this is what you get.. especially on a closed product.

they should learn from the first time. but no. and there will be more "illegal cards" in the future I can bet on that.

[kipper2k]

The V2 cards will naver have any new cores as apollo has abandoned the older a500/a600 vampire boards if you downloaded the latest (working) version before apollo has got rid of them then you have fully working final version before apollo dumped on you!. When i say "working" i mean complete with all the bugs that were never fixed, so the core is flawed and people bought a flawed core which never got fixed and will now never get fixed. But you can probably trade them in to get the V4 bugged core

1 more thing, if you paid for something that has been claimed publically is a perfect working core that was claimed to be 1:1 (as stated by apollo team members in public forums) but obviously not then where do you stand as regards your expectancy of fair usage ?.


Personally i think its time to move on and use other HW, time to put this one to bed

[lmimmfn]

Originally Posted by kipper2k

The V2 cards will naver have any new cores as apollo has abandoned the older a500/a600 vampire boards if you downloaded the latest (working) version before apollo has got rid of them then you have fully working final version before apollo dumped on you!. When i say "working" i mean complete with all the bugs that were never fixed, so the core is flawed and people bought a flawed core which never got fixed and will now never get fixed. But you can probably trade them in to get the V4 bugged core

1 more thing, if you paid for something that has been claimed publically is a perfect working core that was claimed to be 1:1 (as stated by apollo team members in public forums) but obviously not then where do you stand as regards your expectancy of fair usage ?.


Personally i think its time to move on and use other HW, time to put this one to bed

Originally Posted by Chucky

Maybe handle it as a COMPANY and not as several private persons would be a. well. thing? then licensing etc would not be an issue.. but as it seems to be perfectly normal for them to handle all like private persons. this is what you get.. especially on a closed product.

they should learn from the first time. but no. and there will be more "illegal cards" in the future I can bet on that.

This is exactly the issue, they're trying to treat the Vampire as some guy in the basement helping users( like PCBs for the C64 or Amstrad CPC), however they're charging 500-600 euro per card and are a completely unprofessional outfit.

The support channels are : the forum(slow and easily banned for no reason), discord(only help I got on discord was from other users not apollo, Gunnar asked me to meet him on discord to discuss my forum ban but after 3 attempts i gave up), and newly the shop. No support@apollo.com as would be expected from a company.

A normal company(as is done in emulation and FPGA core scenes ) would compile a list of games/demo's with issues(with community help) and provide a roadmap on improvements to reduce those incompatibilities and release the improvements in release notes.

Apollo, while they used to have a roadmap have absolutely 0 roadmap now 0 indication of V2 improvements they've magicked up a new V2 2.6 core(not available for "illegal cards") with no indication of WTF it will do or what it's for.

Their pricing is also becoming a joke, the BOM must be about 150 euro yet they're charging around 600euro for a V4 card while a legit 68060 cars plus TF1260 card can be had for 150euro more and that is a legit card based on hardware available back in the day and not some simulated CPU.

I don't understand their audience anymore, I got interested in the vampire because at the time V2 cards were around 300euro, had a performance boost along with RTG card whilst Blizzard 060 cards were selling for 900-1200 euro,.

That's no longer the case we have the TF1260 card which is mega affordable, so i don't see the point other than maybe standalone V4 cards but I'd prefer a Mister and be able to play arcade games via arcade cores or 8/16 bit machines via their cores and even with the crazy prices for Mister hardware, the FPGA and hardware for arcade cores would still be cheaper than the Vampire and would be far more flexible.

[sean_sk]

Originally Posted by kipper2k

The V2 cards will naver have any new cores as apollo has abandoned the older a500/a600 vampire boards

GOLD 2.16 is being tested now and GOLD 2.17 is also in the works.
Any talk or idea of a fee being charged for these new cores is only for the most recent (last few weeks) batch of V2 cards being sold by Igor on Ebay, not for anything that was sold prior to his departure from the team.

[chucsmom69]

Originally Posted by sean_sk

GOLD 2.16 is being tested now and GOLD 2.17 is also in the works.
Any talk or idea of a fee being charged for these new cores is only for the most recent (last few weeks) batch of V2 cards being sold by Igor on Ebay, not for anything that was sold prior to his departure from the team.

If you believe that your deluded - all apollo care about is the shitty v4 line

[Bruce Abbott]

Originally Posted by lmimmfn

Their pricing is also becoming a joke, the BOM must be about 150 euro yet they're charging around 600euro for a V4 card...

You have no clue.

Relationship between Retail price and BOM cost?

The final retail price is usually 5x the BOM cost. This is for consumer gadgets. There is less markup on electronic components, computers & cheap Chinese gadgets - BrianHG

150 € x 5 = 750 €

And that's for consumer products, where development costs are spread over millions of units. R&D costs for small volume production are much higher.

...while a legit 68060 cars plus TF1260 card can be had for 150euro more and that is a legit card based on hardware available back in the day and not some simulated CPU.

800 € for a TF1260 with MC68060RC50 Rev6 "limited quantity" (for how long?). RTG? HDMI? Enhanced sound and graphics? 512MB? USB? SD Card? Ethernet? Hopefully the Rev 6 chip runs a little cooler than earlier ones, but I bet it's still a scorcher.

I don't understand their audience anymore, I got interested in the vampire because at the time V2 cards were around 300euro, had a performance boost along with RTG card whilst Blizzard 060 cards were selling for 900-1200 euro,.

I guess you would have been more interested in the V2 then, as I was. But I already have a Vampire in my A600 so...

That's no longer the case we have the TF1260 card which is mega affordable

I wouldn't call it that. If you want a real 060 for 'purity' that's fine, but no stock Amiga ever had one nor did Commodore supply it. The 060 has a number of compatibility issues including missing instructions that even a stock A1200 has. If it's speed you want then a 50MHz 060 is about half the speed of the Vampire V2, and it uses more power. IMO the only real advantage of an 060 over the Vampire is the MMU.

don't see the point other than maybe standalone V4 cards but I'd prefer a Mister and be able to play arcade games via arcade cores or 8/16 bit machines via their cores and even with the crazy prices for Mister hardware, the FPGA and hardware for arcade cores would still be cheaper than the Vampire and would be far more flexible.

Again that's fine if it's what you want. Some of us prefer an accelerator that just makes our Amiga better, and if we want to play arcade games we use a software emulator or buy a separate machine.

Just because a Vampire doesn't tick all your boxes doesn't mean it's no good. If I didn't already have a V2 I probably would get a V4, since it is much more capable for not much more money. The only thing I will say against my Vampire is that it's actually too good for me. My A600 has so much power compared to my other machines that it feels like cheating. I want to write software that works on them all including the 1MB A500, so I need to avoid being seduced by the Vampire's power.

But for web browsing or playing RTG games it's awesome - even better than the A3000 with Cyberstorm 060 etc. that I used to have. So who knows? Right now I am saving all my money for November when I retire, when I will buy myself a nice present. Haven't decided what it will be yet...

[Tigerskunk]

Like I said, I really like my Vampire 1200. Works perfectly fine for me and is a nice swiss knife of a card. No need for a 060 here.

It's just not that a lot of new things happened just because of the Vamps existing, and probably won't in future as well.
This, and their treatment of Igor is my only pet peeve.

But your mileage may vary, off course.

[Daedalus]

Originally Posted by Bruce Abbott

Hopefully the Rev 6 chip runs a little cooler than earlier ones, but I bet it's still a scorcher.

At 50MHz it's cooler than the V2 FPGA (I haven't used a V4 so I can't comment there).

no stock Amiga ever had one

The Quikpak A4000T shipped with an '060. While the '060 was only released around the time Commodore went under, the Amiga survived a while longer as a commercial product.

The 060 has a number of compatibility issues including missing instructions that even a stock A1200 has.

Just like the '040. And, as with the '040, they're well known, well documented and catered for by CPU libraries and WHDLoad.

[Promilus]

@Bruce

And that's for consumer products, where development costs are spread over millions of units. R&D costs for small volume production are much higher.

Well that's yes and no. Commercial products have to have everything legit. So using hdmi protocol and connector - royalties. Also - software used to develop commercially available products is usually licensed differently than software for hobbyist usage (despite having similar features) - and yes, that can increase costs. I doubt that's the case with Apollo products though. I'd say it's fairly possible same applies to EDA used by Igor (and whoever made V4) unless it was KiCAD or EasyEDA (or other sw free for commercial usage as well). So while I do agree that consumer electronic costs are that 4-5x BOM cost I think it only proves that it's no hobbyist project to contribute amiga society and get minimal margin but a full fledged commercial device created to make money from it. And it's all good as long as they don't try to advertise it as "made to revive amiga" because that's just a silly claim. Thankfully it didn't penetrate the market deep enough to influence whole community with those silly wars.

[sean_sk]

Originally Posted by chucsmom69

If you believe that your deluded

Nope, not deluded.

Originally Posted by chucsmom69

all apollo care about is the shitty v4 line

And rightly so for a number of reasons:

1. Grond said it well here:

Originally Posted by grond

However, another factor is that Igor took forever to manufacture cards. He wasted a whole year on constructing a pick-and-place-machine because he wanted to turn his home into a PCB solder house (didn't work out). During all this time not a single V2 was built and sold. He resisted the idea of having the V2 PCBs produced by a professional solder house for years. All this resulted in a very low output of cards and many lost customers. The V4-stand-alone was designed to also work in A500/A1000/A2000 computers and this was to be understood as a threat: get your production running or we will produce accelerator cards ourselves. Just ask yourself: when was it publically stated that Igor retired from the team? How long did it take Igor to now get the next batch of V2 cards into the shops and onto Ebay which got Gunnar to react and resulted in this most recent communication from Igor? One year? Two? Eventually time was up and the V4 series were sold.

Don't forget that V2 accelerators were hard to get with waiting lists up to over a year. Then potential customers began venting their frustrations of which Gunnar bore the brunt, while Igor made very few appearances online. Gunnar expressed frustration that V2 accelerators were not readily being made available in sufficient quanitites, so action was taken to produce a unified range of V4 products that people could order on demand without any waiting time. There have been hiccups due to supply chain problems, but overall getting a V4 accelerator now is much easier than trying to get a V2 previously.

2. The Cyclone III FPGA in the V2's is EOL and has been so for a very long time, so why continue to make them? Plus there is virtually no room left on those FPGA's to add any new features. It was Igor's choice to use Cyclone III's in those accelerators, not Gunnar's.

3. The V4 is more feature rich and uses a Cyclone V FPGA which is larger and is still readily available. So it makes sense to focus development on that.

[Promilus]

@sean_sk
@ point #2 - well let's see then... 2015

n V1 we had small FPGA, at the time that was more than enough but after apollo-team started to port Phoenix core to Vampire design soon enough we end up with lack of LE in our FPGA. For emulating MC68000 8K LE FPGA was enough but after we decided to emulate MC68020 instructions each day this FPGA was telling us that we are wasting our time. This time we will try to not create the same mistake. For easier soldering we are going with FPGA build in BGA package.

and then

URGENT UPDATE: March 29, 2015
Quick update to explain new situation. As the Phoenix core getting bigger and more complex (having 3 instructions per clk as we speak) with future possibility to add FPU and get into never seen performance at least in Amiga world I had to change few things regarding design. This will affect the price since I just can't find a way to produce card for 90Eur and satisfy Phoenix core needs in the same time. Inside apollo-team we had 3 day long discussion about this. 10K LE Altera Cyclone III could easily fit in that price range but simply that is not enough for Phoenix. So new redesign is taking place and we will use 40K LE Cyclone III FPGA device instead and that is something I can't get cheap. What is discovered that price needs to go up to 120Eur now to cover everything

it's should be quite obvious te even just slightly intelligent individuals... at that time there was no grand AC68080. Should there be one Gunnar might've gone with DE10 Nano and let someone handle Amiga interface - that one might be expensive but has even more LE than V4 has (and no obvious bugs with signal routing to RAM etc. which could limit performance). But no. There was no real AC68080 at that time. Maybe a concept, nothing more. All those guys had was Phoenix which was growing up to the point it met Gunnar's requirements and became AC. And that growth was stimulated by the only amiga development platform available at that time!. Which is ... yes, Vampire V2 exactly! So Miri's claim Igor "just" designed PCB is one big ugly lie. He did so much more. He created foundation to everything Apollo team has now. So both development platforms for existing members at that time also with that hardware he brought new members as well. With growing owners community he created a certain trademark (Vampire... and yes, despite V4 doesn't use that name it's considered of the same line). One might say his initial design was seriously flawed and without Apollo team it most likely wouldn't achieve even a fraction of what joint-venture did. Yes. And without development platforms widely distributed among amiga users and with their feedback Apollo team wouldn't be able to polish phoenix to get AC68080 level before things like Pistorm appeared. And with host of that niche open PiStorm would eat it's share right away.

[chucsmom69]

Originally Posted by sean_sk

3. The V4 is more feature rich and uses a Cyclone V FPGA which is larger and is still readily available. So it makes sense to focus development on that.

its a product that is never going to be finished - I have been waiting months for a Firebird core update that will fix some of the more silly and simple bugs! instead they mess around with stupid ass 3d features that no one will ever use! My pistorm and TF get more use than my V4 machines for the simple fact they do as they say! no broken promises with them and the lower (much lower) price point! with a full fat 68060 I paid around 450 GBP for my TF!

[Chucky]

well that seems to be their idea. put in functions noone want or will use. instead of fixing compability etc.
while you should FIRST get the damn thing to work as the thing you want to emulate/simulate/mimic. and AFTER that it is working proper, THEN you add the new stuff.. but nah. 3D and weird chipsetformats noone asked for just to lock people into their product. (anyway. THAT is off topic here.- so better stop )

[grond]

Originally Posted by Promilus

it's should be quite obvious te even just slightly intelligent individuals... at that time there was no grand AC68080.

And you are deducting this from two quotes? There has always been a 68k softcore since the Natami days which was called 68050, probably because it lacked some of the 060 features (MMU?). "Phoenix" was the name Gunnar chose for the cut down version of that core that was reduced in size and complexity to fit into the V1 fpga. Igor designed the v2 intending an increase from 8k LE to 10k LE. This, of course, didn't really mean much of an improvement but he wanted to keep the price down to a minimum. I remember him arguing that in his country the average income was 250€ per month and that he considered everything from that point of view. With quite some arguing he decided to make the V2 PCB be able to accept many different sizes of CIII FPGAs. That's why there are some V500 with an 80k LE FPGA that work just as well as the standard variant. That was when for the first time the core could fit into the Vampire series.

Should there be one Gunnar might've gone with DE10 Nano and let someone handle Amiga interface

The DE10 is a board that is only available to developers. You can't base a product on it and not make your own board. It is subsidised by the manufacturer.

There was no real AC68080 at that time. Maybe a concept, nothing more. All those guys had was Phoenix which was growing up to the point it met Gunnar's requirements and became AC.

No, it was a backporting and debugging effort of features one by one from a much larger softcore. That's why the FPU came in such a short time and why the 080 is now only executing two integer instructions per clock cycle when it originally did three. The overall effect on execution speed was too little and clock speed suffered too much compared to the much larger and faster FPGA the softcore was designed for.

And that growth was stimulated by the only amiga development platform available at that time!. Which is ... yes, Vampire V2 exactly! So Miri's claim Igor "just" designed PCB is one big ugly lie. He did so much more. He created foundation to everything Apollo team has now.

How does that earn him any part of the core which is Gunnar's work? Or a perpetual license to the core? Igor can go back to selling Vampires with the tg68 core like he had intended to do at the beginning of the Vampires. Gunnar can choose a different PCB and FPGA for his core (which he did).

You are joining in on the emotional stuff that Igor put forward. All this is a lot like a divorce: it is about money, it is about hurt feelings, it is about revenge. The Vampires are the house the now split couple built, the team members are the children (both parties are supposedly fighting for the best of the children, how much of the money goes to whom...blabla).

[Promilus]

@grond - Natami team didn't show off N68050 softcore to general public even once afaik. There was some groundwork made for it around 2010 and by 2012 there was no N68070 yet (it was basically a blueprint and nothing else). N68050 was "almost finished" at that time but it had single integer pipeline. Now then... your claim is that in an year - between interview for PPA and Igor joining a team there already WAS WORKING AC68080? Synthesised and tested on FPGA? That's your claim? They were supposedly working EXACTLY ON THAT (development platform) but were waiting for 2 things - for natami board to be ready and for FPGAs being powerful enough. At that time there obviously was no FPU yet as well (which AC68080 has).

The DE10 is a board that is only available to developers. You can't base a product on it and not make your own board. It is subsidised by the manufacturer

First - yes you can, that's how Arduino works and that's how Raspberry Pi works. That's also how (early) Mega65 works, UnAmiga works and MiSTer works. Secondly - even if what you claim was true there isn't a SINGLE obstruction for using such development board to just show off full AC68080 just as Cyclone V became available and make a hype, get some kickstarter money and produce custom PCB themselves. Right? Right...

How does that earn him any part of the core which is Gunnar's work? Or a perpetual license to the core

Oh did it sound like I claimed he earned perpetual and free license or something like that? Or was I actually talking about how Miri seemed to diminish impact that Igor's project had on Apollo product development? It gained momentum only due to the development platform readily available and ppl joining the cause due to the affordable turbo card despite it's WIP status. Or maybe due to it.

You are joining in on the emotional stuff that Igor put forward

Not really. I appreciate Igor as a fellow h/w enthusiast and that's that. But I'm pissed off by phrase Miri used to diminish his work and his contribution to the project (and yes, I provided rather valid counterarguments to her statement - or would you like to argue with that as well?).

[grond]

Originally Posted by Promilus

@grond - Natami team didn't show off N68050 softcore to general public even once afaik. There was some groundwork made for it around 2010 and by 2012 there was no N68070 yet (it was basically a blueprint and nothing else). N68050 was "almost finished" at that time but it had single integer pipeline. Now then... your claim is that in an year - between interview for PPA and Igor joining a team there already WAS WORKING AC68080? Synthesised and tested on FPGA? That's your claim?

It is always the same code basis and 050, 070 and 080 were only nametags attached to it. Actually I believe I suggested using 080 seeing how the features of the integer CPU superseded those of the 060 in pretty much all regards. I believe, though, that in the Natami times there was a joint effort for a more feature-conservative softcore and Gunnar's more ambitious own core in parallel.

there isn't a SINGLE obstruction for using such development board to just show off full AC68080 just as Cyclone V became available and make a hype, get some kickstarter money and produce custom PCB themselves.

It was more a decision to focus expertise. Work on the softcore is a lot of work already, somebody else was supposed to do the electronics stuff to give the core a home.

Oh did it sound like I claimed he earned perpetual and free license or something like that?

Well, a lot of people seem to think so. If you don't, then we agree on a lot already.

Or was I actually talking about how Miri seemed to diminish impact that Igor's project had on Apollo product development? It gained momentum only due to the development platform readily available and ppl joining the cause due to the affordable turbo card despite it's WIP status. Or maybe due to it.

Yes, that is true. But the same is also true in reverse. The Vampire project gained a lot of momentum because of the amazing results the softcore showed. Up to then there was only a slow tg68 core that achieved 020 speeds and wasn't bugfree nor feature-complete either. As there was no shortage of 020 and 030 processors, the original project setup was hardly going to win over a lot of potential customers (even though the fundamental idea to use an FPGA and softcore was and still makes a lot of sense).

Not really. I appreciate Igor as a fellow h/w enthusiast and that's that. But I'm pissed off by phrase Miri used to diminish his work and his contribution to the project (and yes, I provided rather valid counterarguments to her statement - or would you like to argue with that as well?).

OK, I can agree with that. But remember that there was a lot of reasons for being p/o-ed by the other side's behaviour. I think the statement "without the Vampire the 080 as we know it wouldn't have happened" isn't any more true than "Igor just designed the PCB". The further conclusions made from these mostly true statements are potentially disrespectful.

[Promilus]

It is always the same code basis and 050, 070 and 080 were only nametags attached to it

No. Memory interface can be the same. 68k compatbile bus can be the same. But due to architectural differences the essence of CPU - execution units - won't be the same. There's significant difference between single integer pipeline and no fpu, dual integer pipeline with pipelined fpu and three integer pipelines without fpu, also it does affect cache subsystem. I don't think they ever before had anything actually working with Amiga hardware and even if they did - they didn't publish such results. As such - only published results of Amiga powered by FPGA-based turbo comes from Phoenix core on v2 which eventually got shaped into AC68080. And it was a long and rough ride. And while I agree that "phoenix core" for V1 might be severly degraded N68050 ... V2 FPGA would hold N68070 (no fpu) just fine so there was no need for a downgrade there - yet it still was phoenix core in development. Right?

[grond]

Originally Posted by Promilus

No. Memory interface can be the same. 68k compatbile bus can be the same. But due to architectural differences the essence of CPU - execution units - won't be the same. There's significant difference between single integer pipeline and no fpu, dual integer pipeline with pipelined fpu and three integer pipelines without fpu, also it does affect cache subsystem.

Well, the claim is that the full-featured Apollo-Core was superscalar, included an FPU and more. From that you can easily cut down to get the lesser variants. I actually was there when the threefold superscalar integer part got reduced to two times superscalar from one core compile to the next. The two times superscalar core we have now allows disabling superscalar execution with a CPU control bit. The execution units are identical anyway, it is just the instruction dispatch unit that changes from one implementation to the other. And, as already said, it is easy to cut down stuff. Of course, the code that was already there needed a lot of debugging which is why at the beginning only a small subset of the features was activated.

I don't think they ever before had anything actually working with Amiga hardware and even if they did - they didn't publish such results.

Interfacing something to a 7 MHz bus isn't really that much of a challenge when compared to the making of the CPU itself. They originally wanted to interface the softcore to the Natami chipset implementation that did not have the slow Amiga bus.

As such - only published results of Amiga powered by FPGA-based turbo comes from Phoenix core on v2 which eventually got shaped into AC68080.

No. They tested the 68k core (whatever nametag it had at that time) in a giant FPGA at IBM where Gunnar and his Natami colleagues worked at the time and benchmarked it against other available softcores including a PowerPC softcore. The results were on the apollo-core website for many years. That was long before there was any Vampire at all.

[Promilus]

If true that actually is even worse. It means they already had working full-featured AC68080 back then and chose to go along with castrated platform to milk users more with "even better than v2". I understand polishing the design and continuous improvement over time but with that information it looks like they held back to earn money twice. One for V2 and then for V4. And that was their intention all along. Smells fishy.

[grond]

Originally Posted by Promilus

If true that actually is even worse. It means they already had working full-featured AC68080 back then

"Working" is a keyword here. Gunnar had a core that could execute code for some benchmarks but that wasn't bugfree at all. It took years of actual use in Amigas to get the core near to bugfree.

and chose to go along with castrated platform to milk users more with "even better than v2". I understand polishing the design and continuous improvement over time but with that information it looks like they held back to earn money twice. One for V2 and then for V4. And that was their intention all along. Smells fishy.

No, it wasn't like that. Gunnar always pushed for using larger and more up-to-date FPGAs while Igor refused and held on to the small FPGAs to achieve a very low price point. Igor can be just as stubborn as Gunnar. The v1200 FPGA is bigger than those of the v600 and v500 due to the ongoing fight between Gunnar and Igor about FPGA size and price. However, it is the v500 and v600 that would need the larger FPGA because they need the AGA reimplementation of the chipset more than the v1200 does. However, the v1200 came later.