Topic: Audiochip Mining (Read 3051 times)

This is way better, you can also play games with this but you can only listen to song on that

I'm surprised this has lasted 2 pages!

Interesting theory on the 'enough chips at low cost + time' idea. Some people have downtime and like to fiddle with things so I think that = free time

Is it like AMD R9 Series?

Hi,

What about the cost of developing such a board? Or are you thinking of having several audio boards on one PC ?

If developing such a board with several audio chips then it won't be worth it (development time and so on make it so).

If using several audio cards on one PC then if you don't have a good MH/s it will also not be worth it....

regards,

P.

this was somewhat interesting

from wikipedia

EMU10K2.5   24-bit   4.0   4,6 million   180 nm, 200 MHz, 424+ MIPS, 64 DirectSound3D sound channels
EMU20K1   24-bit   5.0   51 million   130 nm, 400 MHz, 10,340 MIPS, 128 DirectSound3D sound channels
EMU20K2   24-bit   5.0   ?   Fixes bugs in EMU20K1, PCI Express, embedded RISC processor
Sound Core3D   24-bit   5.0   ?   Integrated analog codec and digital I/O

If the previous generation was 10,000mips .. could this newer model get into the 100,000?
Anyways i see everyone's point its prob impossible, but if there is an old enough/new enough chip that you can get for cheap enough that at a bulk order is actually gets under the current best gh/$ mark AND you could figure out how to get it running together, and efficiently... and the cost of the rest of the hardware doesn't make up for any price advantage...  its obviously technically possible. good luck though.

Its not abut mining with a soundcard.... its about using multiple chips together not a retails audigy etc... ofcourse one signle sound card could do anything

every so often this comes up(as far back as 2011)
https://bitcointalksearch.org/topic/sound-card-mining-12782

mining on a soundcard isn't going to get you very far very fast even if you can get it to work, i'd have better luck with a pentium 3

135W? That doesn't even make for a good space heater! At least my 200W of GPU keeps my feet warm...


A 10 year old Pentium 4 2.4GHz makes 5000 MIPS and about 1 MHash/s. /thread.

No, it's not worth it. You'd lose a ton doing this.

If you were right, buying my ASICMINER blade at $500 would be a steal. After all, it does over 10GH/s. You're suggesting that 90MH/s at $7 is a deal. If so, 10GH/s (over a hundred times more) should be a steal at $500. So why are there ASICMINER blades not selling for even $400?

https://www.bitmit.net/en/item/59882-10x-new-asicminer-blade-10-7ghash-lowest-price-on-bitmit
Look, that's 320 real ASICs, honest to goodness 100% optimized for Bitcoin hashing, more efficient than any general-purpose CPU, GPU, or DSP, fully hooked up and ready to mine for just $3,100 or so.

What you're missing is one very, very simple fact -- only the very latest generation of ASICs are profitable. Anything less than the very best costs more to power than you can make on mining and wouldn't make a profit even if you got it for free.

I have an ASICMINER blade behind me. I just recently shut it off because it costs more to power it than it makes in mining. It has 32 honest-to-goodness ASICs, fully optimized for Bitcoin mining. They're just not the very latest technology, and so even though I've already paid for them and already have them, I can't make money on them. If you're right, put your money where your mouth is. Offer me $400 for my 32-ASIC blade -- it hashes at 13GH/s on a paltry 135W. You can't get anywhere near that with any general purpose hardware.

ok if you keep insisting it means two things.. you are either trolling this topic to get it locked or you are just typing quotes you read in this forum and paste them as you see fit presenting them as your opinnion.. because no offense but the things you say make no sense..


So I make a last try to reason with you...


Most of the things you say are either not true at all or just one side of them is (if you have inmind particular situations and dont see the big picture)


So lets talk with hard numbers... a BFL BitForce Miner that does 50GH/s costs 2500$ if you order it right now 14/10/2013 you will get it around 14/05/2014 if you are lucky..... by then the difficulty will be at a rate of billions so yes this miner is going to bring nothing in terms of profit back to you... so you are right in this particular situation


The big picture now..

Lets assume we live in a perfect world were you could get this particular miner @50GH/s TODAY 14/10/2013 infront of our door.. well then the scenario changes a lot.... because if it starts mining today at current difficulty of 189,281,249 (and scaling that diff to lets say 70% per month that will lead to 4 billion at about 180 days/6 months) you will have summed up about 3000$ networth = 500$ profit (since the miner costs 2500$)

if you had 20 of them that means that you could have 10.000$ profit in just 6 months!




So what stops you doing this? A)AVAILABILITY  or better to say INavailability B)Price of the asics


So if lets say a simple a class microhip (thats used in phones pci cards dvd players etc) ok its NOT a miner asic it cant do GH/s on one chip but it can put out lets say 90 Mh/s on its own and costs lik 10$ or much lower (if you order in big quantities) consider ordering 250 of them (= 2500$ -if we dont count the discount caused by the wholesale type of order- running at ~25GH/s)


that means that after 12 months you will have 3000 usd networth = 500 usd profit

Which by having the same calculations of 70% per month increase of difficulty will lead to a difficulty of 110,279,464,758 at the 12th month.


So this case shows justs that the numbers add up but is it worth it? well if there are not any microchips out there capable of doing more than 90Mh/s and if your capital is only 2500$ then maybe its not worth it..


But if there are some chips that do even better MH rates and/or you capitol is bigger then its totaly worth it.

No, it's because even if they were free they wouldn't make money. I turned the one that I have off. The new generation of ASICs are so much more efficient than the previous generation that the difficulty has risen to the point where first generation ASICs cost more to power than the value of the Bitcoins they mine.

Because that rig would be less efficient, and even the very most efficient rigs of just a year ago are barely profitable, if at all.

The up front cost of the miner is irrelevant it if costs more to power it than it can make in mining.

The key point is this -- the ASICs of a year ago, purpose built to be maximally efficient miners, are already too inefficient to be worth operating if you expect to make a profit. The ship has long ago since sailed on anything not purpose built to be a maximally-efficient Bitcoin miner.

If $7 for 1/3 of an ASIC is a good deal, will you give me $672 for my ASICMINER blade? It has 32 ASICS. 32*3*7 = $672. All you need to do is plug it in, connect it to Ethernet, and it does 13GH/s. I'll even throw in a power supply. Oh, it mines about $2/day worth of Bitcoins and costs about that to power it.

Yea and also you are not good with precentages and proportions... thats all true for AN asic miner for the ONE chip you have..

but have you asked yourself why you just have ONE miner and not like 200 of them?

I am sure you asked yourself that and the answer is easy

1)you do not have money to buy more of them since they are ridiculously expensive
2)even if you have the money to make a big investment by the time you will get the gear you pre-ordered (if you get it at all) it wont do crap or it will have a very considerable decrease of income...



So why not instead of buying like ONE rig that costs 5000 usd and does 2TH/s and will ship Q2 2014 not ordering like 1000 pieces 5$ each of industry available chips reporgram them and have them ready to mine TODAY at lets say 800 GH/s rate... and when difficulty is getting higher with the profits that you accumulate from TODAY since the then buy some more chips that will be ready to ship at virtually the same day!!!

Right, I'm not familiar with technology.

Exactly! And mining with the very latest, very best technology is just *barely* efficient enough to be profitable *if* you are lucky.

Right, now assume those fancy highspeed jets are just *barely* fast enough to get the job done on time. Good luck with the simple planes -- it will *never* work.

I have an ASIC miner that I underclocked about a week ago and finally shut off today because the difficulty has risen to the point that it costs more to power it than the value of the Bitcoins it produces. I've already paid for the miner, I've already waited for it. Those are sunk costs, and it's not even profitable to keep it turned on. This is technology that was designed from the ground up to mine Bitcoins as efficiently as possible about a year ago.

With people using the very latest technology in new designs made from the ground up to mine Bitcoins as efficiently as possible, trying to use anything not purpose built for mining is not even a "knife to a gunfight" idea, it's a "club to a nuclear war" idea.

In about six months, even today's latest and greatest 28nm ASICs designed from the ground up for mining probably won't pay for the power it takes to run them. Even if you got such a miner for free, unless you had a really good way to get cheap power, you would lose money turning it on.

Even your $7 for 1/3 of an ASIC makes no sense. You can get ASICMINER erupter blades for $300 now. They contain 32 ASICs with all the infrastructure needed to connect them. That's less than $10 for an actual ASIC which beats the hell out of $7 for 1/3 of an ASIC. And how can you possibly operate profitably with all the extra stuff not needed for hashing that you have to pay to power when even these ASICs are not, or perhaps just barely, profitable?

This idea is unrealistic by at least one order of magnitude.

Obviously you are not so familiar with technology (ofcourse there is no problem with that) you consider audio chips graphic chips asics FPGAs x86 Processors as to be completely different things..


They all have things in common they do binary isntructions... their difference is their efficiency per chip at some instructions etc...


So a more appropriate analogy would be this:


to transfere 10000 kilos of gold from lets say africa to USA what do you prefere?  to order some fancy highspeed  Jets that will need like 10 hours to do the delivery (which you have to wait a year to get manufactured and be available to you and cost like 10.000.000 usd ) or to use simple propeled one seat planes that will do the job in lets say 20 days but cost in total 5.000.0000 and are available right now.

So why not use popcorn instead of an airplane since popcorn is so much cheaper? Because if you want to fly across the country, you want something that was designed to fly across the country.

Please if you have not the slightest idea about tech then please dont post in this topic... Its like you are saying graphic cards are designed for videogames and not for mining... its needs just for a chip to be capable of doing arithmetics with integers in order to beable to mine... the hard part is to see how to put them all in a board and make a suitable miner program that can use them.

Here is a paper on using FPGA with DSP slices to perform SHA-3 calcualtions.

http://www.ecrypt.eu.org/hash2011/proceedings/hash2011_08.pdf

In this paper, we present results of the comprehensive study devoted to the optimization of FPGA implementations of SHA-2 and ve SHA-3 nalists using embedded FPGA resources, such as Digital Signal Processing (DSP) units and Block Memories. Our methodology involves implementing, characterizing, and compar ing all algorithms with a focus on minimizing the amount of recongurable logic
resources, and achieving a better balance between the use of recongurable and embedded resources. All designs are implemented using four FPGA families, representing major low-cost and high-performance families of Xilinx and Altera


ymmv

fwiw

I had wondered about the same thing -- just never got around to looking for prior work.

The XILIX Zynq 7000 series has a lot of DSP slices.

Digital Signal Processor == DSP == Audio Processor to many people. Sound cards typically have a "low grade" DSP -- i.e. low frequency.

Asic is designed to mine BTC, sound card is designed for sound. I don't think it will work.

Well it could make sense since one BFL asic chip for example costs 70$per piece if for example I could combine 10 audio chip at 7$each doing half or 1/3 the MH/s rate of a single bfl asic then It would be profitable buying like 1000 chips (which will lower the per chip price) spending a few more K for the board design and you have almost instantly mh power whenever you want to upgrade you have just to make a new order that will arrive in a mater of days maybe weeks at worse.. no 6-9+ waiting time.

haha soundcard mining, maybe you could make a proof of work of transmitting soundwaves or something.

Sorry but won't work, it would be too slow and eat too much energy.

If you're serious, this can't possibly work nor could it possibly make any sense. ASICs are made using the very latest technologies to mine as optimally as humanly possible and you need incredible luck to even break even with them. How can you come anywhere close using technology 8 years old that was never meant to hash?

0.07 in some markets!

Agree, I don't really think sound card can do mining much efficiency, unless someone already tried and show proof of it

Umm.  I'll go with, ASICs will be far cheaper and more energy efficient at low speed.  Low-end ASIC chips should be flooding the market. The USB Miner is already down to 0.12BTC or less.  Do you just have piles of chips related to sound cards in your closet?   

A chip that has been manufactured for audio process porpuses like to be on soundcards and stuff

like a CPU,GPU its a APU

Audiochip? what is that.

Hello, I took a look at this: http://emu10k1.sourceforge.net/as10k1-manual/emu_over.html and was interested if there is anyone that has or did heared of someone mine using his sound card or using a miner that uses other kind of chips instead BFL,Avalon asics etc


I mean even if their MH/s rate is low because of their low clocks they are so cheap so maybe they have better MH/$ rate and if their compined into racks of lots of them together then they sound like a good investment.. needless to talk about their imediate availability