Topic: DIY FPGA Mining rig for any algorithm with fast ROI - page 87. (Read 99472 times)

i am repeating because it is quite ridiculous. lol, wat adapt when it is 64%.
Look GPU is 2500 USD for 180mh and E3 can be as low as 800 USD for 180mh. It is somewhat similar to the S9 vs S7 where the S9 is 3 times more efficient. In this case, efficiency is same but cost is far lower. It really depends on E3 pricing. More expensive and less efficient eventually becomes obsolete or cant break even.

There is nothing to gain from scaring anyone haha. If u lot buy into asics, it will increase the network difficulty alot because asics are cheaper on a per hashrate basis. If u lot buy this FPGA instead, it will increase network difficulty far lesser since it is more expensive per hashrate basis. I should be abit happier that more mining money of others will be soaked up for less hashrate. I am not though, doesnt matter to me. I own no GPUs now, juz some asics.

Good luck to u and perhaps track your numbers here too ya.

Well buddy, last time people ride horse. Now, there are cars.
Ok man, will look at your updates then.

I dont have such feelings lol. This is kinda obvious. Bitcoin + mining is a small portion of my assets. Just because it works before doesnt mean it works always. It is human nature to be resistance to change though.

Those GPUs will be there because it is not profitable to use them. Ok then, I guess it is time to put your money where your mouth is.
Perhaps update the cost and hashrate when u buy and program these FPGA then and the profits lol. Lets see.

guts? heh no not really. but i have no problem tossing 6 grand at an experiment that might bomb. that just makes me a somewhat stupid risk taker. at least if i lose this one im still alive, as opposed to some of my other hobbies.

im betting on some coins adopting an anti asic stance. no sense in developing an asic for a coin you know they will change the algo for.

difficulty is a known problem and always has been. its part of the game. we have dealt with it with varying degrees of success before and (hopefully) will continue to do so.

we will see. i will give all data as to profit/loss etc when i am up and running.

OP said these cards take 100w - 160w.. they'll be profitable on something for a while. People aren't going to mine algos with GPUs that lose money lol From the sounds of it these cards will remain profitable long after GPUs are no longer possible. The arguments here are cost vs hashrate vs GPUs.. but you have to remember that the GPUs will consume significantly more power than the equivalent hashrate of a GPU rig. You're also repeating yourself over and over again which honestly seems like you're attempting to just scare people off from the idea of using these cuz 'muh ASICS are coming for everything'. I think everyones aware of what happens to other cryptos when ETH either goes ASIC or Proof of stake.. the insane amount of hashrate pointed at ETH will move. Markets adapt.. at this point I think a lot of anti ASIC cryptos are getting annoyed with these ASIC companies and will likely create a plan to have backup algos slightly modified to switch to prior to an ASIC being released for their current algo. The current ETH ASIC aren't super awesome. At 800 they're decent but add on shipping an a PSU and it pushes the price up.. they can't just flood the market with these or they'll kill the profitability on ETH and they won't sell more of them or possibly annoy ETH devs to the point they really do fork. If anything the current ETH ASIC is likely just going to put a cap on the profitability of ETH.. when ETH increases significantly and GPUs are scarce they'll be able to dump another round of ETH ASICs on the market essentially capping the profit at a certain point to a certain extent.

Those are done but the impact hasnt been felt yet. It will have some impact later and more over time. If u understand, u would sell all your GPUs lol.

Nope, it is a fools play. U still have the thinking as long as I have some non-asic coin to mine, i win. Not at all. When more and more of the GPUs push difficulty of those coins u are mining up, u will make a trickle of profit per month on your fpga or possibly nothing. I guess u will realize why most crypto manufacturers dont take this step from the get go.

Post your purchase and hashrate perhaps and keep this thread update.

No, it looks like you have strong feelings on this subject

Most other posters have a much more pragmatic approach. I've been mining with GPUs and ASICs so I'm quite sure I have a fairly good grasp on how this works. You're dismissing a lot of variables - power consumption, density, forks, exchange rates, etc. BTW the best time to be a GPU miner was when there were cheap GPUs flooding eBay.

you keep harping on the e3, sha256 and other currently asic dominated stuff. those are done deals, no ones gonna use a fpga to mine those (or at least long term) thats not why we (or i) would go with this. there are enough non asic algos to make it profitable. part of the fpga strategy is using its reprogramability  to move among different algos. basically what gpu miner do now.

also i envision we could, if needed, put bounty programs up to get new bitstreams up faster.

Summary.

FPGA is not workable when one asic controls 64% of GPU mining market.
Lets say u use FPGAs and they now control 20% of remaining market via 8 algos.
One asic will appear and easily cut it to 15%. They will take the major coin away again...

Your FPGA is basically an antminer S7 and when new asic appears, it gets "downgraded" to be an S5. More FPGAs/GPUs will also join in the crowd as u earn less. U may not even be able to be profitable once it gets downgraded to S5. Do note that difficulty increase does not lower profit, it lowers REVENUE. So, u may go negative on mining once it "downgrades to an S5". I strongly doubt it will work because the asics are far cheaper and too much more efficient and majority of the market is centered around a few big algos.

This is not yet including the biggest problem which the E3 that can mine eth, which controls 64% of the GPU. It can keep pushing more GPU mining power to the coins u are mining.

No point I keep repeating. I find it conceptually just a horrible idea.

Hope to see transparency in numbers and figures to brave souls those who attempt this project. Good luck people

chineses farms will have all the supply of chips . where are those gpus made? ehehe... will there be gdr6 finally or it's will all be hbm2? will they all have small gram bus? I bet...

it's the apple model... always leave something out, for the "next forced upgrade"...

and subsidiary will directX 13 be only avaiable against payement or the telemetry (aka full spying on your public computer) is enough from microsoft to make money?


why sell proftable fpga? take a loan and build more for yourself .

I fully agree with prospect of FPGA vs ASIC but the point is that you are not that dependent of one manufacturer. There is no reason to wait for E3 for example over 3 months right, since its basically a nice packed GPU rig. and by the time you get it everyone gets it. Imagine that you can mine now CNv7 with 1 card @ 64Khs. Also the flexibility and consumption vs GPU is pretty good. Agree that the pricing is out of range for most miners....

It does not matter if new algo is safe for 6-8 months. Difficulty is CONNECTED.

An OLD ALGO gets taken over by asic and all the GPU/FPGA miners on that old algo have to go to your new algo coin. Difficulty of new algo go up. Now, the new algo u are mining makes very little, if any. Lets say the old algo has 5-10% of total GPU mining power. The new smaller algo only have 2-3% of old mining power. Wat do u think happens when huge mining power (2-3 times more) moves to the new smaller algo? The new algo coin becomes worthless to mine. The old algo coin that gets taken over will be a bigger coin with more mining power. Asics always target the bigger coin


That is why GPUs have been losing market share and thus, are now easier to buy than before. What u are doing with FPGA is just making another GPU lol. A few algos control the GPU market. Thus, a few asics control the gpu mining market.


I guess u people just have too much feelings or something. I bid u god luck anyways. All the best.

In my opinion, interest to FPGA can be short-term, when you need to mine a certain number of coins before the release of ASIC. Like a Groestl, Nist5 and others.
But a more fundamental interest may be to offer a useful proof of work function  which can also be ASIC-resistant.
By setting new standards for the usefulness of the POW function, we will strengthen mining on FPGA in the long term.
In the end, someone uses mega-FPGA cards in real life. For what? Amazon AWS earns with them, why not do the same?

It seems u dont understand how eth controls most of the GPU mining difficulty and how mining difficulty for GPU is linked.

Why dont u guys just go do it and waste your money then. U will not be able to compete on price or efficiency. U will be at the whim of the E3 miner price and u will have very little to mine and that amount will keep decreasing.

Also, wat u are doing is nothing proprietary since everyone can just go in and do it. As it is, just a lowered E3 price would mean alot of GPUs add to other coin mining difficulty.

Perhaps state wat u spend and wat the FPGA can do here and track your returns. And when u lose money, u will realize why crypto manufactures dont make the FPGA that u are doing. They can do it at a larger scale, at a cheaper rate and easily beat wat u are doing But they wont because it isnt workable or is just too small to bother with.

Look, it doesnt make sense. It will cost too much more, have very little to mine and be at the whim of the antminer E3 pricing basically lol.

Most important question. WHY just WHY do u wanna make an FPGA when one asic already controls 64% of GPU mining? A few asics control most of the GPU mining already.

And this FPGA u are doing is just another GPU. Why? Because anyone can do it. U make a slightly stronger GPU known as FPGA. All u need is abit of dumb money going in there and u make no profits due to high difficulty.

@yrk
Bitmain lowered their L3+ price from 2000 to 1500 and to 450. It is now 700+
The D3 was once 5k and then 2.5k and 1.45k and now 311 USD. Yes, 311 USD.

They have huge margins on these machines. Wat makes u think u can compete with them on an expensive FPGA?

Most important question. WHY just WHY do u wanna make an FPGA when one asic already controls 64% of GPU mining? A few asics control most of the GPU mining

That's not true. They can release a new ASIC every month only because they started developing them last year. A new algo is safe from ASIC for 6-8 months. Some coins (Monero IIRC) have plans to fork every 6 months. Just the threat of that is enough to deter ASIC development for that algo.

FPGAs haven’t been able to compete with GPUs for mining in any meaningful way since their application to SHA256. The main issues being their high costs and very long development cycles.

On the other hand, they are great for low latency, low memory bandwidth, and real-time applications while being very energy efficient. You can say that they are a great complement to GPUs, which are ASICs, by the way, for graphic processing.

So we will have to wait and see if they can be economical for mining in 2018. I am open minded since I have four years of experience working with them.