Vampire 500: Joining two worlds, Amiga and Bitcoin...
Exactly 4 years ago I got an offer to make Bitcoin miner using FPGA and just for fun I did it using Vampire 600 V1 card and Amiga 600 as a base, showing pictures that it can work. For some strange reasons I decided that complete Bitcoin story wasn't for me at least as a person who is not regular user but developer. Lack of time, because I was digging deep into my Amiga projects was one of them. I was never prepared to leave my Amiga dreams and search for something else. Soon I have realized that my Amiga projects have no future because constant lack of money for research. Despite donations from various sides it was never enough and after lot of thinking in 2014. I made crucial decision to continue my Amiga journey as a developer but support Bitcoin idea as a regular user. With money earned from selling several Vampire 600 V1 cards I have got 9 Sapphire 280X cards and started to mine. Plan was simple, make economy for my Amiga projects who will be independent from banks, loans or donations and with recent events about Bitcoin complete Vampire development, production or anything related is financed like that. In this moment when I m writing this article just one Bitcoin cost over 15 000USD. That statement bring us back to the beginning of this story and makes me wonder how much I could earn by participating into Bitcoin story as a developer...
What is Bitcoin exactly? I won't bother explaining much just simple facts we need right now. Bitcoin is a cryptocurrency based on SHA-256 hashing algorithm. It is decentralized digital currency without bank to control it. That says that middle man(bank) don't exist anymore. To simplify things, without middle man world as we know it can't survive, without taxes there is no country who could see another day. Who could blame Bitcoin users, well people finally have opportunity to OWN their money.
Another aspect of this story is that every single Bitcoin user have encrypted data of any other user in peer-to-peer network like true decentralized system. Who could stop that?
Terminology to investigate before you continue to read: Blockchain, Bitcoin wallet, Miners, Blocks, Block reward, Difficulty adjustment, Proof-of-Work, Mining pools, Mining farms...
Yesterday I was interested to calculate how much money I could earn mining Bitcoins using one Vampire 500 V2 card so I have ported opensourced FPGA bitcoin mining core to it and started to investigate.
Final result was poor as expected. I could mine 0.00000577 Bitcoins per year or in today's value exactly 0.08552 USD, so not even 10 cents per year. In other words I would need 173 310 years to mine one Bitcoin. Let's get back once more, in 2013 with same equipment making 100MH/s I could mine 2.78 Bitcoins per year, or in today's value exactly 43,372.82 USD. Funny story says that in 2013. I had about 200 FPGA chips with working FPGA Bitcoin core, but let's not make further calculations it would force me to drink some funny pills to calm down. You get the picture. So what changed is that more miners with strongest hardware increased difficulty to mine. Right question now is who can mine Bitcoins now. Simple answer is no one. Without serious equipment, ASIC based miners, loads of them, you can't do anything. Bitcoin mining is now reserved only for mining farms who costs millions and millions of dollars. Back in 2014. I was unable to directly mine Bitcoins with my GPU equipment it was that hard. Where is the trick you may wonder.
Bitcoin isn't only cryptocurrency there is more than 10 000 of others and by mining them you can earn lot more. All of them have their own developers, supporters, communities, goals and usability and most important own economy. Some of them are used just for paying for games, others are created for renting cars, rooms, buying food... Mark my words, soon every single serious company will have their own cryptocurrency based on theirs name and used as a payment method for their products. Why, because there is no middle man, because it is secured, safe and damn fast! By speaking about other cryptocurrencys we can also say one important thing they are not based on same hashing algorithm as Bitcoin(SHA-256) there are the others like scrypt, blake, equihash, keccak and many many more. With this I m able to get Bitcoins not directly but by mining other coins because tiny portion of Bitcoins mined by mining farms is always invested in other coins who have promising future and innovative ideas. Economy of others is also ensured by direct investments in fiat money. Mining other coins is much easier because difficulty is much lower than on Bitcoin. That says that just by using one Vampire card for mining I can earn lot more by mining some other coin. Especially in the terms where only several people mine one coin. It can happen that in just one day with modest power you can do very good.
Ok so then why don't we make our own coin who will establish our own economy and who will be used for reviving Amiga.
In other words I could make my own coin and offer to sell Vampire cards only and only for that coin. Same time buying those coins on exchanges would ensure stable economy and bring investors from outside of Amiga land. Trick is in one thing called opensource.
Almost every crypto coin that exist is opensourced so everyone can see how is designed. This is very important because it gains trust in potential miners or investors. Opensourcing is good and bad thing. By presenting code to the masses they can understand hashing algorithm you used in your code and using high power devices like ASIC or GPU miners they can instantly destroy any potential value of the coin by creating many of them in short period of time and selling them over exchanges purely for instant profit. But again there is a trick, ASIC or GPU miners can't handle all algorithms yet so as long you use brand new algorithm or modified old one or even combination of several algorithms you are safe. By creating semi-opensource showing everything regarding coin parameters but hashing algorithm you can be protected. You can mine it with single CPU for years without much power.
MC68000 and crypto
As said not even modern processors can handle huge difficulty on Bitcoin and other popular coins. As tested on some retro hardware old processors don't stand a chance. But think it like this, coin who can be only mined with MC68X series of CPU. Again, power is not important because others can't enter. With minimum processor power you can mine blocks and get reward in coins who gain value because they are rare. They are rare because only people who owns retro hardware can join. They are the only ones who could participate and eventually create profit. Others will be forced to just buy those coins on the exchange. Yes, they could use emulators on their PC but that could be also limited with CPU detection routines. Limited but not removed as option because using emulators you will again gain interest for retro computers. Same time those who buy them are your investors. Why should anyone buy them? As my experience says, what is rare, new, innovative and with purpose is very interesting for investors in this crypto world. By creating new coin with new algorithm you are forcing mining farms out and that's exactly what new coin developers doing right now so why not do the same with MC68000. Bitcoin remains mainstream for big investors, we don't need to mess with it directly.
Just to mention there was something called Amigacoin but only connection with Amiga was in its name because it needed to be mine on X86 architecture. Imagine coin who could be mined only on Amiga.
Our position is much different, a lot better. In the situation where we can actually build specific instructions and create our own rules Apollo-core could be next best thing outperforming MC68K in our specific coin. Joining retro and crypto coins world could become reality with millions of potential users. Popularity of Amiga and everything that follows could rise because they are lot of newcomers who are always exploring. By creating specific coin for us we could ensure two way story, buying core from end user for personal profit. Right now we have more than 1000 users of Vampire accelerators and good starting point for this to happen. If we start talking about standalone system, Apollo-core based we could open new markets spreading power of Workbench towards mining community who is getting bigger and bigger each day.
In not so distant future everything that single person possess now regarding material stuff will worth 0.01 Bitcoins, those who have 0.10 will be rich. Other coins will eventually have same value Bitcoin have now so why not create our own and continue promoting Amiga in this new futuristic world.
Posted by majsta on Saturday, December 09 2017 @ 15:55:47 CET (20 reads)
(comments? | Score: 0)
Vampire 500: Inside the beast
Apollo-team member Claude took Vampire 500 V2+ card to the X-RAY. Here are some pictures you may find interesting.
Posted by majsta on Wednesday, October 11 2017 @ 16:39:52 CEST (97 reads)
(comments? | Score: 1.66)
Vampire 500: 4th batch
After holidays our suppliers and manufacturers had we resumed production of Vampire 500 V2+ but this time in much bigger quantities so there will be plenty cards for re-sellers and for those who placed orders over apollo-accelerators.com website. Same time I m already ordering parts for next batch so from now on there won't be unwanted delays between batches. I will do my best to provide constant flow. Waiting time is now reduced from 3 month to 7 days. So expect to be contacted any minute :)
Someone spend incredible amount of time to report my E-bay sales of Vampire cards. He reported me for various reasons, that I m not using genuine pictures, that I have stolen them from someone, that product is fake, that I m misleading buyers, that product is not listed in proper category and so on. Yes, imagine Vampire card can't be listed in Vintage computers category. Then he reported me that I m advertising article as Amiga product, hard to explain but he reported me because I wrote "Amiga accelerator" instead "Accelerator for Amiga". It is really hard to count how many times that person reported me and for what reasons and this is something that goes on and on for more than year. Each and every time as a result card was removed from the auction. After that people who were bidding contacted me asking what is going on and asking questions about removed listing. First few times I didn't even bother to investigate what is going on thinking that this was purely E-bay's action. Then I got one more mail from them who contained this sentence "Another eBay member notified us that your listing used their copyrighted image or text without permission.". That sentence proved everything. It is just shame that E-bay don't investigate things properly. Don't get me wrong I m doing my best to respect all the rules but just search for anything retro computers related over E-bay and you will see that anything goes for everyone but for me. Well as a result, my account on E-bay is suspended yesterday, consequences are huge because this harms my reputation there and rights to buy certain components we need for cards production.
IMHO complete purpose of those attacks is to slowdown project development and for that I see various reasons.
Either someone don't like complete concept and somehow he is afraid that Amiga development will have different roadmap than he was thinking. It is perfectly fine to think like this because Amiga community is divided by various of reasons. There are groups of people who don't like PPC, FPGA, 060 or even any kind of upgrades and they think that Amiga should remain stock and used like that. Someone of them get sick when you mention AGA running on A500. Even those are divided, some thinking that AGA is good in A500 but not in FPGA environment. Some of them supports 68K only, again some gets angry when you mention PCI in Amiga and consider that as a true sacrilege. How about mentioning x86 inside Amiga or God forbid mentioning Atari emulation on Amiga. There are many examples...
Those who likes cats don't understand those who likes dogs but all of them will say that it is good thing that we have cats and dogs and possibility to chose what we like the most. I wish for more Amiga related projects, more choices, only that could move us forward. However, I think that those reports are not related to some person who don't like our roadmap. Once more I see money behind all of this.
I see much bigger frustration here and see some things connected. If you look for every article on this website you will see that someone put huge effort to vote for each article hundred of times trying to lower the score to 1. I m picturing person who don't have any real and honest friend, who don't go out, who is scared in the room full of people. Yes, I m picturing person who don't have life. I m picturing it even now when he is reading this, smiling, thinking how wrong I am, but am I wrong. Think how many days of your life you lost wasting time reading E-bay rules and regulations just to stop Vampire sales. Let me tell you that so far I have send 464 Vampire 500 V2+ cards who are sold outside of E-bay. At the end you did what exactly? How about your lost time?
Again one more report and again my account on E-bay in danger, just read what E-bay said to me. This is awesome people, simply awesome!
Please understand that you stated "Vampire 500 V2+ accelerator" in the title of the listing while the description states "You are bidding only for accelerator card" which is misleading. Misleading titles confuse buyers by making it unclear what's actually for sale in a listing. Buyers should be able to determine what's being sold based solely on the title and subtitle. Confusing or misleading information in titles and subtitles clutters the search results and creates the potential for bad buying experiences. Kindly revise / relist the item with the appropriate title of the item that is being offered in the listing.
This batch consisted of 200 cards and it was biggest batch so far. All of the cards were sent in just 20 days according to mentioned plan where highest priority had our pre-orders list then re-sellers. Next batch will be build soon and it will be even bigger than this one.
After several years of negotiations Apollo-team made agreement with AmigaKit for distributing Vampire 500 cards. For me personally this is some kind of dream come true and I m so proud for this to happen. Amigakit is the big name on retro scene and I can freely say that because those kind of shops retro scene survived. Anyway, I noticed that because of that lot of people made bad comments on various forums. just this morning I have received more than 30 mails to explain situation. Problem appeared because people from our personal pre-order list feel betrayed because they think that AmigaKit got cards before them after so much waiting. True story is that in last batch 200 cards were produced and few of those went to AmigaKit so they can test the cards and see are they interested to resell them. They tested them and said that they are interested in future cooperation. Those cards were just preparation for future sales and they are sold within 15 minutes of publishing news. Right now there are no cards left to send to anyone before next batch is done. I only have one for myself used for debugging. Our pre-order list have highest priority and will be cleared according to timeline. No one will be left out or betrayed. From previous post you can see how much cards we have done and that simply can't keep up with the demand. Again, all become insane. And again, people please have some more patience because once you get the card you will forgot about all of those problems and long waiting and you will say that it was worthwhile. After all you are waiting for something like this for 20 years not two months. Remember AmigaKit is not just re seller... Soon plan will be revealed and you will say that this was smartest move ever.
May 30, 2017.:::UPDATE:
Packages including parts for next batch arrived today. Created misunderstanding will be history soon.
You can't satisfy all but people from the Apollo-team and from complete Amiga community was right in lot of things regarding this project. First and most important thing is that new members joined the team just because they got the cards by some pure luck. In just few days they helped a lot on various areas same time saving us lot of time. That makes me think who knows how many good coders and hardware designers are out there who could help in our development but they can't because we are delivering cards slower than any developer/manufacturer in the history. In fact we failed in most important thing. We were able to do everything but to actually deliver cards. That caused lot of bad reputation and that was the reason why price manipulation happened at all. My attempt to solder cards by hands and using my Pick and Place station failed because massive demand. We still have about 2200 cards to deliver, imagine that! I only wish everyone could imagine how hard it was for me to spend so much hours each day soldering and not seeing an end, not seeing a day when I ll be free. Same time looking at comments of certain people who couldn't survive one single day with the tasks I m doing for past 4 years. No one knows that several times I ended up in emergency room after 15 hours of constant soldering. No one knows that I prepared project files and what's left of investment money to forward it to some team members to continue if something happens to me. Well who cares...
When others developers succeed with modest products we failed. They were just able to deliver number of cards in short period of time. After all we can conclude it is not so important that you have best product, all is about marketing and organization. For those who advised me what to do and how to do it all looked simple then I should just say. You have the hands, you have the brains, money and whatever is needed to do it yourself if you think that this is so simple. When I worked in electronics company we had army of people doing certain tasks, here all of those tasks fell on few people who become lost. After all everyone blamed me but I planned my moves and soon all will become clear. Next step, hiring professional company to do the job.
Thanks to IDVORSKY LABORATORIES from Belgrade, Serbia and our team member CRVENI we were able to enter one of the latest stages of this project. EMC testings on Vampire cards are done in 150 kHz - 30 MHz - 2.7 GHz range in this specialized laboratory funded by EU. Complete test records are available upon authority request.
More pictures in Read More, Pictures, Files... section of this article.
Again, have no idea how many times I have redesigned Vampire 500, now I m done. Am I happy with latest design, well NO. Don't get me wrong design is nearly perfect, that nearly is what's bothering me. What I could do better, let's say lot of things but let's face it I ll be rude once more and say that this is best accelerator that is done so far for Amiga 500. What I wanted is to leave SDRAM as something we use and go for Mobile DDR. That design is done last year but never released, again have no idea why. To tell you the truth I didn't know what I was doing in last 3 years with Vampire 500 design. It was never clear to me what to do. Should I leave SDRAM, add Ethernet or WiFI, build card who don't connects directly on top of old CPU, use different FPGA. In fact, as we progressed with the apollo-core I have realized that my oldest design was good enough for the job. Hey, more than 100MIPS rock solid accelerator and I was thinking about what could be done better. For about one year I had bunch of Vampire 500 PCB laying around, not thinking at all that someone will be interested in those cards. As you know my pre-order lists for other accelerators were smal,l then interest from the community started to rise at the huge rate so I got scared asking myself only one question. If there is so much interest for V600 V2 what will happen to the version for Amiga 500. Constantly I was receiving lot of mails where people asked me about other models and that was definite prove that demand will be huge. Since I said that I won't run pre-orders anymore all will be handled much differently. When I solder enough cards I ll run one batch, send it to the people then run another. Waiting time will be reduced and we could cover the market in a fair way and I won't be in heavy pressure like before.
Strange thing is that lot of people lately tried to advice me how to handle stuff. Don't you think that I m capable of ordering 1000 cards pre-assembled in China on my own? Don't you think that I can pay the man who will send them or open a company, hire nice secretary to answer phone calls. Those who knows me, knows well that I m not shy or incapable of talking to the people and doing business. Right word at right place, business. I won't run it, this will remain underground project, one man in the basement doing hand soldering forever. If you could understand the feeling when you create every single card using your hands and bring it to the life. Each one is special, each one is connected to me, no matter where is it now in the world. It is piece of me forever. Each one have it's own story. For each one I burned my finger at different place. I would lost that feeling if I build them any other way. Simply it won't be the same for me. As you know all huge project were born in basements created by stubborn persons. Me and my Pick and Place machine working as one last few days, cards looks better than ever and everyone will get them in awesome package. When? Keep watching at amibay!
Vampire 500 V2+ differences over V2 version: 1. Rework on internal layers that will probably enable higher clocks.
2. Modification on headers <->DIL to support much better connectors who will ensure stable connection to the Amiga motherboard
3. Tiny improvements on video-out again regarding internal layers
4. I/O expansion header with support for WiFi module(AP, client mode) (WiFi module drivers not developed yet, but in future maybe they will be:))
5. Logo update
Vampire 500 V2+ is 100% compatible with Vampire 500 V2 version.
More pictures in Read More, Pictures, Files... section of this article.
As much I don't like to write about myself and publish any pictures this time I had to. To this day this video had more than 10K views on youtube so I think that is OK to post more information's now. Yeah, you may say 10K views is nothing, but this is different crowd so number is huge.
I went to Novi Sad, Serbia on September 10, 2016 to visit BalCCon2k16, Balkan Computer Congress organized by LUGoNS, oldest Hackers community in Serbia. I was invited by Once Upon a Byte and 8bitchip retro communities and this is the best way to thank them for invitation, nice pictures they took and awesome video they have done. In the video itself I m talking nonsense most of the time, living in a dream, or am I :)
Plan was to talk more about Apollo-team and Vampire 500 accelerator. You may ask why video isn't on English, because when I started to talk with some man from USA I just froze, couldn't say any single word on English. Why, I have no idea. That's the main reason why there are only few people in the room because most of the visitors were from English speaking territories and complete plan was that everyone who present something there should talk on English. Simply I couldn't. Anyway, people who invited me there were very friendly and we talked about lot of subjects. What I liked the most, hm don't get me wrong but girl at the bar was awesome, sorry I didn't took any pictures but she is still in my mind :P
More pictures in Read More, Pictures, Files... section of this article.
My only concern about this complete project was how to provide enough current for this type of accelerator to work. In worst case scenario I had to think about someone who have 2.5A Amiga power supply. What if our FPGA core at some point become so complex and FPGA needs to sink 3A to work properly, stable... External power supply was out of the question so I had to find another way. To solve this problem I had to enter the territory who is not known to me, understanding how switching regulators works. All the time in my previous designs I have used linear regulators with fix output values and that was easy to do. Let's try to compare linear and switching regulators first.
Only steps down (buck) so input voltage must be greater than output voltage
Step up (boost), step down (buck), inverts
Low, about 25% regarding efficiency VS load current
High, we are about 85% regarding efficiency VS load current
Low, few capacitors needs to be used
High, requiring inductor, filter caps, resistors
Large, part itself is not big, but together with other components needed takes lot of space on PCB
High, mostly because external components used
Medium, needs to be fixed in PCB design itself by creating AGND planes...
*Table used from article "Understanding the Advantages and Disadvantages of Linear Regulators" By Steven Keeping, simplified for our needs.
As you can see from the table above there are lot of advantages we got this way, but to simplify all let's say that we don't have heat problems anymore and that our current will be more stable and in values we need, so in short from 1.5A we have left after Amiga motheboard takes what's needed we create 3A needed for the accelerator to work properly.
Vampire 500 voltage regulation simplified
On this specific accelerator few voltage regulations are needed and 5V we get from Amiga motherboard we can use directly only for HDMI, rest of the included devices works on different voltages. To suit their needs we need to create lower voltages and enough current. First switching regulator who is connected to 5V will create 3.3V for voltage translators and dedicated FPGA I/O banks. From there we use linear regulator to create 2.5V for internal FPGA PLL. We use linear regulator there since PLL don't consumes much current and we are using it from 3.3V since voltage difference is smallest there so we will not create much heat. Last two switching regulators are for creating 1.2V needed for powering FPGA core and 1.8V for mobile DDR memory used, dedicated FPGA I/O banks.
Posted by majsta on Thursday, March 26 2015 @ 15:41:11 CET (10008 reads)
(comments? | Score: 5)
Vampire 500: IDE controller integrated
Posted by majsta on Friday, September 26 2014 @ 13:15:51 CEST (1013 reads)
(comments? | Score: 5)
Vampire 500: Revisited
Last year I started Vampire 500 project, FPGA accelerator for Amiga 500 but without keyboard, floppy disks or let's say complete Amiga 500 I needed to stop that project. Goal was to have fast accelerator with emulated core and support for IDE with Gayle emulation. Few days ago I received Amiga 500 keyboard from ShK and bunch of Amiga floppy disks so now I have everything to continue this project. SysInfo results and Gayle detection are shown in Read More, Pictures, Files... section of this article.
Posted by majsta on Friday, September 19 2014 @ 16:06:21 CEST (2888 reads)
(comments? | Score: 4.87)
Vampire 500: Core testings
As you may noticed some very interesting things are happened regarding this project but news about that I didn't publish on this site. One Vampire 500 was sent to robinson5 from retroramblings.net and it was the one that I used for my testings. This is maybe most important thing happened to complete Vampire FPGA project because we are now in situation that we have opportunity to exchange codes and to point each other what might be a problem. We have already exchanged few mails and codes and it is very interesting to me that we solved few problems but using different approach. It is normal that two people can think different and coding in VHDL is something that is easier to create than to understand what someone wanted to do with specific part of the code. Another great discovery from robinson5 was bad solder joint on A16 trace of Vampire 500 and that was the reason why my code didn't worked. Earlier I have discovered that there was problem with one resistor placed on 7Mhz clk trace so after solving those hardware barriers everything regarding code should not be such problem.
From the start of this project people on various forums are again against my projects talking that Amiga 500 uses MC68K to the max and that I will never be able to create something that is done for Amiga 600. What can I say about that, only few lines of code were needed to be implemented in old code and everything worked. Now we have 3 versions of the code and all of them are working fine and main differences are in solving accesses to MC8600 peripherals.
After about 8 hours work on my regular job I come home and I was so bored and decided to start another project. So at 10:30PM VAMPIRE 500 project is started and after just half an hour I have done most of the work :) Yes it was so easy to adapt hardware part but in software part I will have to create some additional signals but that should not be a problem. Few minutes ago I ordered Amiga 500 motherboard because I have the privilege to see Amiga 500 just once around 1999 for about 15 minutes. Why I started another project and didn't finish VAMPIRE 600? In the process of waiting for new PCBs to arrive I have nothing to do. Another reason why I started this is because I m tired of waiting of some people to finish their project constantly changing designs and again trying to make a living from Amiga name. Those days I m moving to much bigger room and there I could get all needed space for my equipment. So now it is the time to push myself to the max.
UPDATE: April 17, 2013 After publishing this new project I received lot of mails to include IDE on this board. I don't understand this demand for such thing because I was thinking to replace it with MicroSD. Then again what could I know what is important for Amiga scene after all I m new here. So since I have never looked at IDE schematics it took me two days to understand complete thing so I implement it but also decided to keep MicroSD socket. For now it is purely in theory that IDE will work. So implementing this delayed my production but It will take only few hours more to verify PCB. Orders regarding components are placed and most of them are on my table so at the end of this week PCB production will start.
UPDATE: April 18, 2013 PCB in production :)