DIY FPGA Mining rig for any algorithm with fast ROI

[whitefire990]

@whitefire990,

- will you be coordinating the group buy? How soon will those  Bittware FPGAs  become available? In 2-3 weeks or in September?

- Since you are mining with VCU1525 could you recomment any specific RAM brands/model  that would work for sure with BCU1525?   I have bought a few FPGAs from Allmine and Squirrels but no-RAM version.
Thanks

The XUPVV4 is available very soon (a few weeks in small quantity), but the 300A 'mega' crypto board from Bittware will be October.  They are trying to get someone to organize a group buy, details available soon.

None of the developers are currently using the RAM in the VCU1525/BCU1525.  I would avoid buying any until a developer specifically mentions it.  In my case I extracted the RAM from all my VCU1525's to use in my PC's.

[wevieNL]

At multiple locations I read requirements of getting the FPGA's being programmed with cryptographic keys (efuses being blown) to support encrypted bitstreams. Is this also a requirement for the bitstreams from http://zetheron.com/? Which providers of bitstreams do have this requirement?

[whitefire990]

At multiple locations I read requirements of getting the FPGA's being programmed with cryptographic keys (efuses being blown) to support encrypted bitstreams. Is this also a requirement for the bitstreams from http://zetheron.com/? Which providers of bitstreams do have this requirement?

Paranoid developers are more likely to require the efuse method for anti-piracy.  I am mostly doing what I am doing to support the community and prevent ASIC centralization.  By having as many possible ultra-fast FPGA's in everyone's hands, ASIC's become non-profitable to develop, even if they are still faster than FPGA's (they are just not enough faster to make the millions of dollars of investment worthwhile).

In that sense I am okay with less strict piracy protection.  None of my stuff requires e-fuses, all of my stuff will run on any VU9P FPGA board that is supported on the website.

I know DigitalCruncher's bitstreams also do not require e-fuses.

I would like to think that everyone wants crypto to succeed, and everyone wants to fight against ASIC centralization, so hopefully the community will also try to support the developers rather than screw them over.  Maybe my hope is naive, but I like to believe in the best in people.

[Luisjc69]

Hello guys,

sorry, but I dont know so much about this "new" FPGA!

My questions....

I have a rig with 7 cards RX580, can I use the same Board for the "new" FPGA cards?

When not, what I neaded exactly for run one system? (FPGA card, mainboard..etc)

Thanks a lot for the help

[Manojdaran]

Im really a noob please bear with me if it looks silly. I looked at this model "XUPVV4" Does this require any additional hardware to set up or just like a GPU, plug and play? And this one "BCU1525" - Xilinx has a sales office in my country, can I buy it from them directly? Does this require any additional stuffs like the former one? Please guide me guys. Thanks

[Weidley]

Hello. I have some doubts, let`s see if some one can help me.
According this link: http://zetheron.com/index.php/fpga-performance-profit/
There is a limit number of FPGA that can be turned on before the profit became negative, about 25k FPGAs. OK.
My questions:

1. Well, it`s a small number for a big farm. Why exactly this farm don`t monopolize this limit ?

2. In real world, when something rare came to public it`s because it already has been offered to big players and they refused. If so, What they know that we don`t ?

[whitefire990]

Hello. I have some doubts, let`s see if some one can help me.
According this link: http://zetheron.com/index.php/fpga-performance-profit/
There is a limit number of FPGA that can be turned on before the profit became negative, about 25k FPGAs. OK.
My questions:

1. Well, it`s a small number for a big farm. Why exactly this farm don`t monopolize this limit ?

2. In real world, when something rare came to public it`s because it already has been offered to big players and they refused. If so, What they know that we don`t ?

Right now the number of these FPGA's (in the world) that are actually available is about 500 (taking into account high-current boards only).  So even if someone wanted to 'monopolize' this technology they could not.

Later this year, 5000-9000 boards will enter the market.  It will be late 2019 before 25,000 boards are even manufactured.  By then, the crypto market will probably have doubled or tripled, which means the number of supported FPGA's also doubles or triples.  So it is not clear that the FPGA's can ever catch up to the market.

[bitwookie]

Is anyone here mining Denarius DNR with any FPGAs? If so how many do you have? How much DNR are you getting per card per day on average in the last 24 hours? Is anyone finding it hard to mine?

Im thinking about getting a few myself and want to know results.

Cheers

Wook

[whitefire990]

Is anyone here mining Denarius DNR with any FPGAs? If so how many do you have? How much DNR are you getting per card per day on average in the last 24 hours? Is anyone finding it hard to mine?

Im thinking about getting a few myself and want to know results.

Cheers

Wook

Digital Cruncher released Tribus 1.95GH/s for Denarius.  However I don't know how much it generates because the two mining calculators give drastically different projections.  One says $1/day the other says $18/day.

[bitwookie]

put it this way DNR had 50-100GH nethash on there until FPGA's came on board. Now nethash is peaking at 2 TH

[DigitalCruncher]

Is anyone here mining Denarius DNR with any FPGAs? If so how many do you have? How much DNR are you getting per card per day on average in the last 24 hours? Is anyone finding it hard to mine?

Im thinking about getting a few myself and want to know results.

Cheers

Wook

1.95 GH/s Tribus will generate about 35 DNR/day, which is equal to 0.0018/day at current prices.

This calculator produces valid estimation:
https://whattomine.com/coins/187-dnr-tribus?utf8=%E2%9C%93&hr=1950.0&p=250&fee=0.0&cost=0.1&hcost=0.0&commit=Calculate

[roc-noc]

Hello guys,

sorry, but I dont know so much about this "new" FPGA!

My questions....

I have a rig with 7 cards RX580, can I use the same Board for the "new" FPGA cards?

When not, what I neaded exactly for run one system? (FPGA card, mainboard..etc)

Thanks a lot for the help

Here is where I found the best hardware description.

http://www.zetheron.com/Downloads/VCU1525_Instructions.pdf

Also you need one USB port per Xilinx card.  I plan to use the same hardware that I have in my GPU miners but with a much faster CPU and more RAM.  And also a bigger power supply.

[Manojdaran]

@whitefire990, Once I get the cards, do I need to program for each card or just connect all and program once?

[RivAngE]

I wonder if we have any estimation about how many FPGAs have been sold and how many will become available on August.
If the few sellers don't put a rule like "2 cards per address" there's no way I'll be able to secure any, but I hope that they'll do their best for the sake of keeping the networks decentralised.

[Weidley]

Hello. I have some doubts, let`s see if some one can help me.
According this link: http://zetheron.com/index.php/fpga-performance-profit/
There is a limit number of FPGA that can be turned on before the profit became negative, about 25k FPGAs. OK.
My questions:

1. Well, it`s a small number for a big farm. Why exactly this farm don`t monopolize this limit ?

2. In real world, when something rare came to public it`s because it already has been offered to big players and they refused. If so, What they know that we don`t ?

Right now the number of these FPGA's (in the world) that are actually available is about 500 (taking into account high-current boards only).  So even if someone wanted to 'monopolize' this technology they could not.

Later this year, 5000-9000 boards will enter the market.  It will be late 2019 before 25,000 boards are even manufactured.  By then, the crypto market will probably have doubled or tripled, which means the number of supported FPGA's also doubles or triples.  So it is not clear that the FPGA's can ever catch up to the market.

I see. Ok but this don't explain why someone don't try to take for him this production. Except if they are intend to manufacture his own fpga. If this is correct the manufacture of this fpga can increase significantly next year.

Otherwise I think what can save the limit production is if there are only a feel factory that can manufacture this fpga or if there is a limit on feedstock per year. But the same way don`t explain why big players aren`t try to take for him the production.

[R0land]

Hello. I have some doubts, let`s see if some one can help me.
According this link: http://zetheron.com/index.php/fpga-performance-profit/
There is a limit number of FPGA that can be turned on before the profit became negative, about 25k FPGAs. OK.
My questions:

1. Well, it`s a small number for a big farm. Why exactly this farm don`t monopolize this limit ?

2. In real world, when something rare came to public it`s because it already has been offered to big players and they refused. If so, What they know that we don`t ?

Right now the number of these FPGA's (in the world) that are actually available is about 500 (taking into account high-current boards only).  So even if someone wanted to 'monopolize' this technology they could not.

Later this year, 5000-9000 boards will enter the market.  It will be late 2019 before 25,000 boards are even manufactured.  By then, the crypto market will probably have doubled or tripled, which means the number of supported FPGA's also doubles or triples.  So it is not clear that the FPGA's can ever catch up to the market.

I see. Ok but this don't explain why someone don't try to take for him this production. Except if they are intend to manufacture his own fpga. If this is correct the manufacture of this fpga can increase significantly next year.

Otherwise I think what can save the limit production is if there are only a feel factory that can manufacture this fpga or if there is a limit on feedstock per year. But the same way don`t explain why big players aren`t try to take for him the production.

[paranoia]
There are some organisations (secret service,...), who need a lot of this fpga´s to build big clusters (like the copacobana). The production is limited and they have higher priority.  
[/paranoia]

[mike_la_jolla]

Right now the number of these FPGA's (in the world) that are actually available is about 500 (taking into account high-current boards only).  So even if someone wanted to 'monopolize' this technology they could not.

Later this year, 5000-9000 boards will enter the market.  It will be late 2019 before 25,000 boards are even manufactured.  By then, the crypto market will probably have doubled or tripled, which means the number of supported FPGA's also doubles or triples.  So it is not clear that the FPGA's can ever catch up to the market.

Your '500' is very low.  5000, and possibly more, of the 1525 variants will enter some sort of operation within the next few weeks.  This is gonna be a race until the rigs become unprofitable. 

I would add that the secondary market for these boards, once rendered unprofitable for altcoin mining, is nil.

PS -- The '300A' Bittware card is a single FPGA?  VU13P maybe?  Seems like overkill.

[fr4nkthetank]

I believe its 5000 mineority/squirrel/fpgaland group buy.  All sold out and wont be in stock for 6 months I think.
Plus all the bittware cards coming out, numbers unknown.  I think those numbers can be large because I asked for a price for 50 liquid cooled v and they didnt get back to me too quickly

[gigabyted]

I believe its 5000 mineority/squirrel/fpgaland group buy.  All sold out and wont be in stock for 6 months I think.
Plus all the bittware cards coming out, numbers unknown.  I think those numbers can be large because I asked for a price for 50 liquid cooled v and they didnt get back to me too quickly

those 5k bcu1525 are being shipped in august...

[whitefire990]

Right now the number of these FPGA's (in the world) that are actually available is about 500 (taking into account high-current boards only).  So even if someone wanted to 'monopolize' this technology they could not.

Later this year, 5000-9000 boards will enter the market.  It will be late 2019 before 25,000 boards are even manufactured.  By then, the crypto market will probably have doubled or tripled, which means the number of supported FPGA's also doubles or triples.  So it is not clear that the FPGA's can ever catch up to the market.

Your '500' is very low.  5000, and possibly more, of the 1525 variants will enter some sort of operation within the next few weeks.  This is gonna be a race until the rigs become unprofitable. 

I would add that the secondary market for these boards, once rendered unprofitable for altcoin mining, is nil.

PS -- The '300A' Bittware card is a single FPGA?  VU13P maybe?  Seems like overkill.

The secondary market for these boards is huge.  Xilinx sells 250 boards per month to non-crypto developers.  The boards can also be used for AI-mining and other distributed cloud-supercomputing applications.  In my opinion once these boards are everywhere, very few people will be mining with them, as there will be more profit to be made in other fashions.  Just look at Amazon.  They rent out these FPGA's for $30 per day to non-miners.  The cell phone giant Huawei has also launched a cloud-FPGA platform to rent out FPGA's and compete with amazon.  Crypto is a tiny segment for the applications of these boards.

The Bittware CVP-9 is a 300A VU9, and the CVP-13 is a 300A VU13.  Not overkill; even at these currents we are still power limited in our applications, by a significant margin.  We would need 500-700A to run at full voltage, full logic, full clock speed.