Intel unveils BonanzaMine, a Bitcoin accelerator ASIC | Hacker News Reader

Intel unveils BonanzaMine, a Bitcoin accelerator ASIC

fuse.wikichip.org

59 pointsrbanffy129 comments

Huh.

Here we have the largest, most important chipmaker in the US making and selling chips specifically for Bitcoin transaction processing (AKA "mining"). I was not expecting to see something like this, from a company like Intel, anytime soon.

This sort of thing makes me think Bitcoin is on track to become a permanent component of the world's financial infrastructure, regardless of what I or anyone else thinks about it, because any efforts to outlaw it or regulate it out of existence will now hurt, not just tiny startups and offshore operators, but also leading companies like Intel, which is of strategic importance to the US.

Most politicians in the US will be very reluctant to mess with a fast-growing, potentially large market for US technology products and financial services.

>Bitcoin transaction processing (AKA "mining")

Unfortunately, mining isn't scaled with processing transactions.

Bitcoin has a hard cap of 7 transactions per second, regardless of whether 5 miners are running or 5 million.

The amount of resources that are put into mining scale with the price of BTC, not transactions.

Also Nvidia has made mining-only silicon for several years now. So Intel isn't the first major company doing it.

> I was not expecting to see something like this

On one hand, it shows Intel believes mining is here to stay.

On the other hand, I suspect it makes up a very small part of their business, and any regulation won't hurt much.

Think of it this way: Intel just validated the market for every Fortune 500 corporation.

I can't believe we keep on sending tons of CO2 in the atmosphere just to find special numbers...

You can use these as electric heaters. Since there is no mechanical work, all energy input is radiated as heat. In this scenario, capex aside, the hashes come out for free and can offset a little of the electricity bill. If they can be encased in passive radiators and run at 60+ Celsius, that'd be awesome for a home boiler. Lower temperatures would be really nice for keeping heated pools hot.

It's actually very inefficient use of energy to use miners as electric heaters.

Now that we care about energy efficiency and atmospheric pollution, hopefully more people will learn that efficient electric heating takes the form of a heat pump, where you end up with more heat energy in your space than just the electricity put into the device.

A miner cannot match the heating efficiency of a heat pump because it dissipates the negative-entropy into computation instead of pumping.

Kind of true, but is it widely used form of producing heat? From where I live, the heater burn gas from PG&E. So it is still doing the inefficient kind of turning energy into pure heat.

I’m still undecided on the subject (whether crypto-currency is intrinsically bad to the environment or not.) Eg what if we regulate and only allows renewable energy to perform this. The basic argument is to intercept energy which would have been turned into heat anyway (solar, geothermal, etc.) and instead directly turned into heat, perform mining first then to heat.

For others who may misunderstand, they are talking about thermodynamics. Ie the amount of heat produced can be higher than the energy you put it, because it also get some from the environment. A random article found on the subject: https://courses.lumenlearning.com/physics/chapter/15-5-appli...

Indeed. The only use case for resistive heaters is when air-sourced heat pumps are much less efficient - in the extreme cold (significantly below freezing).

In any case, the miner fans are loud.

Or when you don't have central heating/air.

Or when you can't afford a retrofit of your system with a heat pump.

Or when you need to target dry a small room or a pair of wet shoes.

They have plenty of other uses, the only caveat I guess is that they aren't used as the main source of heat.

> In any case, the miner fans are loud.

That's why I proposed encasing them in a heat transfer shell and keep the thing submerged in water. They could keep the water in the boiler at around 60 Celsius.

At what temperature would a heat pump be less efficient than an electric resistance?

A proper heat pump is never less efficient than an electric resistance.

Soil, at a depth of only around a few meters, rarely exceeds 10C, even in summer. That gives you a 45C differential for your boiler in the worst of worst cases with a ground heat exchanger, which is more than enough for best-case efficiency, year long.

This will be less expensive than a crypto in material cost very rapidly.

> This will be less expensive than a crypto in material cost very rapidly.

That I never doubted.

> At what temperature would a heat pump be less efficient than an electric resistance?

There is no such temperature, with a good enough design of heat pump. A good heat pump is always more efficient at heating than electrical resistance.

The improvment is a multiplication factor, which can be quite large; it's not just a few percent. The factor is why I'd describe mining - or resistive heating - as a very inefficient use of electricity for heating.

Since we are theorising about the merits of miners, we may as well use the theoretical limit for the alternative. Wikipedia explains: "If the low-temperature reservoir is at a temperature of 270 K (−3 °C) and the interior of the building is at 280 K (7 °C) [...] only 1 joule of work [...] 28 joules of heat are added to the building interior".

In other words, a heat pump can be 28 times more efficient than running a miner or other resistive electrical heater in those conditions. But that's a high number, a theoretical limit in conditions chosen to put heat pumps in their best light.

More realistic, based on real devices: "In Europe, the standard test conditions for ground source heat pump units use 35 °C (95 °F) [...] and 0 °C (32 °F) [...] the maximum theoretical COPs would be [...] 8.8. Test results of the best systems are around 4.5. When measuring installed units over a whole season and accounting for the energy needed to pump water through the piping systems, seasonal COP's for heating are around 3.5 or less. This indicates room for further improvement."

So 3.5-4.5 times more efficient than running a miner for heating in real conditions with current heat pumps, which can be improved. That translates directly to reduced atmospheric emissions from electricity generation.

The efficiency of the heat pump reduces when the outside temperature reduces, and reduces when the inside temperature increases. Conversely, if you are heating water, the time when the heat pump is most efficient is in the summer.

Cost is distinct from efficiency. The cost of a unit of energy as electricity is higher than the same unit of energy as gas, for example, so if you are comparing an electrical heat pump to gas heating, at a low enough outdoor temperature the gas heating is cheaper. This is why some places switch between heat pump and gas modes. The same argument does not apply when comparing with electrical heating like miners, or with an electrical water boiler.

> reduced atmospheric emissions from electricity generation

That's a key point. If your power comes from renewables, this discussion is a lot less interesting. I'm all in for efficient use of energy (and I don't think crypto mining is a good use of it), but I also know a heat pump is a lot more complicated than an electric resistance (or a hot PCB board).

> It's actually very inefficient use of energy to use miners as electric heaters.

Not sure I get where are you going here. There is no mechanical output of the miners, only heat.

> A miner cannot match the heating efficiency of a heat pump

My hot water comes from the heat pump in the winter, when heating kicks in, but the boiler relies on an electric resistance in summer, when the house is mostly in equilibrium with the outside.

> the boiler relies on an electric resistance in summer, when the house is mostly in equilibrium with the outside.

That's the situation where a heat pump is most efficient. I think you could reduce your hot water electricity cost by a factor of 3 by using heat-pumped hot water in the summer, instead of resistive heating. I guess the reason your system doesn't do that in summer is that it would require separate equipment.

With the assumption of outdoor temperature 30°C and water 60°C, the theoretical limit is a factor of about 10.

With so many people experiencing a cost-of-energy crisis at the moment in the UK, it's disappointing that these methods are rarely seen here. I've never seen a heat pump in a home in the UK.

If you were to invest the amount of money that a good crypto setup would cost, a heat-pump with an underground loop would be miles more efficient in the summer than an electric resistance.

This is the only justification for mining I’ve ever been able to come up with.

It's not even correct. It's much more efficient to use a heat pump.

Heat pumps may be more efficient, but they can't be employed everywhere an electric resistance is used. To be fair, miners can't either, but are much easier to deploy.

We don't need to think solely of miners - crypto-mining is just one form of monetizing computing. If we could have something like a reverse Amazon Mechanical Turk, we could put computing resources available to receive and process packaged workloads cheaper than, say, an EC2 Spot Instance could, provided the client can cope with the added latency. As for incorrect answers, it could use some consensus where you send the payload to two independent parties and, if they don't match, you send it to additional parties until your consensus is ready.

There are no applications I can think of where crypto-heating works, but a heat pump doesn't. Miners are a lot more difficult to deploy than a heat pump - they need an internet connection, they can break down, they will eventually need to be upgraded, etc..., whereas a heat pump is pretty much once-and-done, and you'll maintain it once every few years.

As for your second paragraph, will it ever be cheaper to do a computation twice independently than once in a trusted party? I'm sure there are a few situations where that's the case, but I struggle to think of them.

Here are two scenarios when a heat pump is harder to deploy: 1) Heat pumps need access to the outside cold reservoir, for example they don't work entirely inside a closed room, where an space electric heater would be fine. 2) If the temperature gradient is too large, heat pumps are less efficient than a regular heater. So many heat pumps fall back to a regular heater when the outside gets too cold. If it is very cold all the time, a heat pump would not make any sense at all. This is why they are deployed mostly in places with mild winters.

This is flat out wrong. Heat pumps are only practically efficient within a range of temperature differentials.

You already need the heat, might as well help launder some money ;-)

More or less. I guess there’s incentive issues too. Maybe now people keep their houses unnecessarily warm to mine more —- but again, this is the closest thing to any kind of justification I’ve ever heard.

Now all we need is a way to make that boiler lined up with top-process silicon costing thousands of dollars per square inch cheaper than the one that uses an electric resistance.

And do that for the environment!

With all the discussion around UBI, housing shortages, and crypto, I think it would be interesting to build tiny houses with solar arrays that would be funded and heated by the crypto miners running in them.

They don't need to be tiny (unless we want to maximize external surface and, with that, the heating requirements).

Unfortunately I find it very easy to believe, just exceptionally disappointing.

Can you come up with a better solution?

Political ones have failed, since with national currencies we end up with inflation and spending the resources of the future (i.e. debt).

The decision to "suspend temporarily the convertibility of the dollar into gold" has lasted indefinitely. This means the government can print its own money with no accountability, until it acquires everything and everyone.

Have you looked into proof of stake? It is a pretty good contender, but is vulnerable to Sybil attacks in case of a network partition. With no objective way to tell legitimacy (like "work" that was wasted in proof-of-work), a network partition can not be resolved.

> since with national currencies we end up with inflation and spending the resources of the future

One of the important distinctions between Austrians and Keynsians is that the latter have noticed that money isn't a resource. It's a claim on resources.

The resources of the future that we are spending are the real resources: fuel, clean water, energy, crops, minerals, human effort, the natural environment, and the biosphere services themselves.

You can print money, but you can't print oil. If you print money faster than you exploit real resources, then yes you get inflation. However, the converse is not true. The ultimate demand of anti-inflationists is that your salary remain $X - forever, the cost of a car remains $Y - forever, and the cost of a loaf of bread remains $Z - forever. You can't achieve price stability with technological hacks.

The reason why bitcoinism is so offensive to many people is the way it takes real resources - in this case silicon wafers - and turns them into nothing but money. It's a prayer wheel to Mammon, like Arthur C Clarke's Nine Billion Names Of God.

I apologize for not being so well-read. I have not read books on Austrian or Keynesian economics. But I will tell you what my demands are, and I will appreciate any recommendations.

> The ultimate demand of anti-inflationists is that your salary remain $X - forever, the cost of a car remains $Y - forever, and the cost of a loaf of bread remains $Z - forever.

I don't want prices to stay still; that is unrealistic (and inadvisable) in a dynamic, adaptable system.

I want prices to trend downwards more than salaries. I know this is possible because it happens to computer hardware, adjusting for performance. Why does it not happen to, say, bread? fuel?

Is it not because central banks are targeting inflation? And because currency created by central banks go to private banks and the government first and not wages at large?

I would appreciate a book answering these questions.

> I want prices to trend downwards more than salaries. I know this is possible because it happens to computer hardware, adjusting for performance. Why does it not happen to, say, bread? fuel?

> Is it not because central banks are targeting inflation? And because currency created by central banks go to private banks and the government first and not wages at large?

These are good questions to ask! And the answers are kind of littered around orthodox macro. As well as the real economy. Computer hardware benefits from (astonishingly huge) productivity improvements. Bread does as well, but to a much lesser extent. Fuel, now that goes all over the place, but the increasing capital cost of extracting it as the easier oil has already been extracted plays a part. As does cartelization - OPEC were largely responsible for kicking off US inflation in the 1970s by restricting supply. That's the point at which wages and profits start to diverge on that graph that everyone always circulates.

> I want prices to trend downwards more than salaries.

There have been a lot of policies aimed at keeping wages from rising; both the maintenance of a minimum level of unemployment ("NAIRU" https://www.investopedia.com/terms/n/non-accelerating-rate-u...) and the attack on unions and collective bargaining as means of increasing wages. It's policy to make sure there's always somebody ready to underbid you.

Thank you. I will look for good "orthodox macro" books.

> One of the important distinctions between Austrians and Keynsians is that the latter have noticed that money isn't a resource. It's a claim on resources.

Have you read Mises? Because based on this facepalm comment I can only assume you haven't, and that your "knowledge" of Austrian economics is derived from Paul Krugman screeds.

> The decision to "suspend temporarily the convertibility of the dollar into gold" has lasted indefinitely.

Followed by the single most prosperous period in Human history.

> This means the government can print its own money with no accountability, until it acquires everything and everyone.

Ah yes, which is why Venezuela and Zimbabwe rule the world.

Economic progress was actually far better under the gold standard. That was the time of the industrial revolution. That was the first time great masses of humanity were lifted from poverty.

> Economic progress was actually far better under the gold standard.

Would love for you to point me in the direction of the supporting evidence.

Millions were lifted out of poverty in the 19th century, whereas billions have been lifted out in the 20th and 21st centuries on the back of enormous credit expansion.

I'm not arguing the fiat currency and credit cannot be incredibly dangerous...they can be. But it's absurd to think that money supply should be fixed to the arbitrary quantity of a random element that we can dig out of the ground.

> Would love for you to point me in the direction of the supporting evidence.

There is scholarly work arguing for this. George Selgin comes to mind as someone who I believe has written about this. There's probably also scholarly work arguing the other side. I can't take the time to dig this up for you right now.

> But it's absurd to think that money supply should be fixed to the arbitrary quantity of a random element that we can dig out of the ground.

It certainly isn't absurd. There are lots of people who think this is a good idea, with good reason. Plus, it is historically precedented, and for good reason.

Should it be fixed to an arbitrary quantity of random numbers we burn electricity to find, either?

"Absurd" is the new normal in post-modern society.

If I had to pick which better exemplified post modern society and absurdity, I'd pick unlimited government money printing by the Fed, over bitcoin.

Weimar Germany pretty much defined what it means to be "post-modern," and they printed themselves into extinction (admittedly, that's not the whole story).

Good god did someone drink the kool-aid... please explain succinctly what "problem" is being solved by a handful of wealthy individuals using all of the world's GPU power to mine numbers for mostly unusable currency...

> mostly unusable currency

I suppose that's why it has grown in price in spite of its supply slowly increasing. Because it's unusable! (sarcasm)

> "problem" being solved

In essence, it takes scarcity to make scarcity. So, to get digital scarcity, you have to fork over scarce resources.

But I agree with you that the computing cost is excessive, and the reward function was not well calibrated.

On the other hand, the rate of the explosion in price could not have been foreseen.

>I suppose that's why it has grown in price in spite of its supply slowly increasing. Because it's unusable!

The fact that the price is consistently increasing is what makes it an unusable currency. A deflationary currency is a useless medium of exchange as any rational actor will not spend it due to the potential future increase in price. I hold multiple crypto assets but I think you have to be kidding yourself if you think that Bitcoin and the like are usable currencies.

Computer hardware has gotten exponentially cheaper in the past half-century. But people keep buying it.

I am looking to buy a flat, so I will spend some of my Bitcoin.

All of the same problems that the "Internet" as we agree upon it, via protocols, is solving today,

just slightly worse (because its brand new and you keep speaking against it instead of improving it, for some reason)

and also with better security, and portability.

Rhetorical, almost flabbergasted questions for you to consider:

What problems does "money I can store on a Ti-83" solve?!

What problems does the Ethereum Virtual Machine solve, as far as turning decentralized network code into a server/computer?

> and also with better security, and portability.

What security?

> What problems does the Ethereum Virtual Machine solve, as far as turning decentralized network code into a server/computer?

Given that the eth VM uses a fraction of a single percentage of its processing power to run smart contracts (like, 0.00000001%), and the rest in redundantly verifying those transactions, it's not solving any problem of decentralised processing in a useful way.

The "vanity receipts for badly drawn apes problem"

This article is about making custom purpose silicon only usable for sha256 based proof of work mining, and has nothing to do with GPUs. GPUs have not been profitable to use in bitcoin mining for many years now.

> Can you come up with a better solution?

Yes!!!! It's called central banks and monetary policies. Seriously. We need to get back our democratic control over it, and make it work for us again. We have to stop wanting to fix our societal issues with technology: we can't fix our money with crypto-currency any more than we can't fix mass surveillance with encryption.

If anything, crypto currencies are a concentration of what is worse with our existing financial system: even less democratic control, no regulations and therefore all sorts of scams, and we can't tune it to make our economy work.

In a world ruled by crypto, our governments would have the perfect alibi for failing: it's not their fault, their hands are tied, they can't get the money for the projects that are needed. This has happened plenty of times in the past when we were using gold and such. And that's just the credit crunch problem, there are many more.

>Political ones have failed, since with national currencies we end up with inflation and spending the resources of the future (i.e. debt).

I never touched crypto so I am I wrong that it's value is even more unpredictable then putting money in the bank? Like I can buy something now and tomorrow some news about a hack or some internet drama and I just lost everything. As a dude that is not into economy theory my conclusion is that crypto would have negligible value if it wasn't used to speculate it will grow or used in illegal stuff, I won't be surprised if behind all of it is a group of obscenely rich and greedy individuals

Bitcoin is like owning some cool trading cards, they are valuable now, they might be more valuable tomorrow and some internet drama about the show/authors will make it's value negative day after tomorrow.

>Have you looked into proof of stake

PoS is like fusion power in that it will be here soon but unlike fusion power those working PoS don't actually want it to be used it seems to me.

> Political ones have failed

In what sense has the current monetary system failed in your opinion? The gold standard certainly failed, that's why it was abandoned by every single country in the world, but the current monetary system seems to be doing just fine.

Just fine?

To be able to save for retirement, you have to assume a lot of risk (stocks) or accept a negative real (inflation-adjusted) return (bonds). This is because inflation acts as a tax on top of income tax and social security etc.

The gold standard failed and inflation is higher because governments like to spend more than they earned. This is unsustainable - any asset holding its value will be preferable to funding a government that is constantly losing money (i.e. bonds).

The economy of the future doesn't owe you anything, that's not how it works.

Yes it does. I work for money, and I expect to get something in return when I spend that money.

And what I get in return should not be exponentially robbed by tyrants to pay for bread and circus.

Well, it doesn't work like this. Producers of goods and services have no obligation to sell a fixed amount of goods and services for a certain price in the future. You have a choice to consume now or to save and consume later, and there's an inherent risk to saving because the future is uncertain. And this is true regardless of the currency that you use, or the inflation rate. A crypto-currency is not going to make the future any less uncertain, sorry. You have been sold a lie. In fact, if you put your savings in a crypto-currency, saving is going to get a lot more risky for you personally, not less.

I'm not sure you realize how bad deflation is.

I'm not sure you realize how leaky the abstraction of "deflation" is, or for that matter all of the abstractions thrown about in the laughably oversimplified version of economics that has made it into the lexicon of pundits and bureaucrats.

The IT hardware industry has been deflating for, what, half a century now (Moore's law).? Yet nothing bad happened.

This has nothing to do with inflation or deflation, which are very specific concepts.

That is inflation. Transistors have gotten less valuable.

It is transistor-inflation, where we print up tons of transistors. It is the opposite of currency-inflation, where we print up tons of currency, driving prices UP.

Inflation is not necessarily a bug or something you “need to solve”.

Inflation is a feature, not a bug. It discourages hoarding of money -- and encourages putting it to use, as opposed to letting is slowly lose its value. Too much inflation, of course, is bad, but a small amount offers the right incentives with very few downsides.

Putting it to use where? Stocks are at ridiculous valuations, and bonds have negative real yield. That's not really encouraging investment.

Not even classical one-sided economic schools like "Keynesian" are as absolute as you pose it here.

Inflation, nor deflation are bad. Too much of any is. The common consensus is that the scales, on average, should lean a little towards inflation. But only a little. And not always.

>Inflation, nor deflation are bad. Too much of any is. The common consensus is that the scales, on average, should lean a little towards inflation. But only a little. And not always.

I think the common consensus is that any deflation is too much. What's contentious is just how much inflation is acceptable and how much is counterproductive to economic growth.

> I think the common consensus is that any deflation is too much.

This depends on how broad you define it. We've seen deflation for the past decade now. In the sense that the buying power of your savings on a traditional savings account have shrunk, while official inflation was below zero at points. We've now bounced back to "high" inflation.

None of this is bad, as we can see clearly by how well tech has benefitted from this (cheap loans, beneficial DCF). This is even policy.

>None of this is bad, as we can see clearly by how well tech has benefitted from this (cheap loans, beneficial DCF). This is even policy.

This led to a lot of economic problems with zombie businesses that were able to continue to trade only due to cheap loans and continued asset price growth sans proportionate economic growth. Any this was before we saw any significant periods of real deflation.

Inflation encourages consumption, in the same way that returns on capital encourage investment. We need both, and sometimes we need more of one or the other. Anyone who thinks moderate inflation is bad does not understand deflation.

We talking about sudoku apps?,

Dora the Explorer broadcasts?,

or are you referring to some other waste (y'know, to someone) of electricity/CO²?

"All things use electricity!" isn't a great critique of "Bitcoin is an inefficient use of electricity". Bitcoin is designed to be inefficient and get worse over time.

You'd better start working on a more efficient Layer 2 if you truly care about this issue,

because this shit isn't going back in the box.

Governments are piling in, nobody meaningful is going to ban it, because they cannot.

It's not as if Bitcoin does something inefficiently that is accomplished in some other, more efficient way. "Commodity" money which has no counterparty is going to be valued at the marginal cost of production, period. It's not Visa, it's not FedWire, it's hard money.

The critique is actually spot on, because when you peel back the non sequiturs about efficiency the underlying criticism is that Bitcoin itself is pointless. Maybe so, maybe not. But that's a different argument.

> Bitcoin is designed to be inefficient and get worse over time.

OTOH, we may be able to use it to heat Mars into a tropical paradise.

Please spend a few minutes reading more on this topic, as you may change your view after learning where Bitcoin’s electricity usage fits into the broader picture. Some good starting points may be:

https://www.newsweek.com/bitcoin-mining-americas-most-misund...

https://www.coindesk.com/business/2021/03/05/the-frustrating...

The articles you cited could do a better job putting the real numbers front and center and present the real, and anticipated (if Bitcoin takes off) ecological impact on the planet. The coindesk article in particular cites really alarming claims about energy usage but doesn't do any real math to disprove it, and only vaguely hints at possible mitigating arguments. Appeals to emotion, citing ideal visions of Satoshi and using "domestic tumble dryers" (newsweek) as comparisons for energy usage further weaken the point these articles should be making.

Even if this is not cutting egde, it will sell and more so with all those small ASIC chips upon their top node, will help fill out the wafers to allow the unused jaggy edges to be filled with these smaller dies in production.

Now, even without that possibility, Intel is ramping up its fabrication and offerings and if they can maintain that momentum outside their core buisness as they seem to be going, then the more products they can keep the fabs busy and turn some profit is good. Even if they broke even upon this, they would still have this and other avenues to pad out production runs and their lab running costs.

Depending upon how this goes, may see this fused into some GPU offering down the line for that market, certainly they have the ability to do that if the demand is there and also help open up there GPU plans more.

So on the face of it, sure it's another option for crypo miners, but deeper, it too me seems like Intel leveridging some low hanging fruit to give there fab offerings more momentum and shifting towards competing with the TSMC's being an end goal perhaps, certainly lots of catchup. Though the aspect of production wafers and having lots of smaller chips you can pad out a wafer, sure does seem like a capital move and be interesting if they go down that path. For me, it seems like mixed wafer production using smaller dies to flesh out the unused space left by larger dies upon a round wafer as a great way to reduce waste and turn that into profit.

"... the BonanzaMines System is shown achieving 55 J/TH whereas the Bitfury Clarke achieves 40 THash/s at 56 J/TH while the Canaan Avalon A9 does 30 TH/s at 58 J/TH."

55 J/TH versus 56 and 58 J/TH from competing vendors. Small margins of course matter at scale, but this (preliminary) improvement in energy efficiency, less than 2%, isn't really impressive. Nor does it seem like an amount that can make a dent in the energy consumption problem of Bitcoin mining. Are they aiming at selling their product cheaper than the competitors' offerings? Doesn't sound like Intel. Am I missing something?

> Nor does it seem like an amount that can make a dent in the energy consumption problem of Bitcoin mining.

If you make a dent in the energy consumption problem, doesn't the network automatically adjust to pop the dents out (increase the difficulty)?

It does if miners persistently aim at their maximum power allotment and keep investing in an equal amount of new hardware and extra hashrate. The pivot point where operational costs outweigh the revenue seems to sit just beyond the power allotment for most operations, at least until something radical happens with the power efficiency and costs of mining equipment.

Yes it does.

Indeed. The price of a bitcoin and the price of energy are the only two variables in the function that determines the energy consumption.

The depreciation of mining equipment is a third variable, and it is rather significant at this point in time.

As an example, the Antminer S19 Pro retails for $8-10k and is rated at 3250 W. Running it 24/7/365 would consume 113.88 MWh of electricity over a four-year lifespan. At average marginal hydro power rates ($0.05) this would cost $5,694 while at typical residential rates it would roughly double to $11,388.

The hashrate is an input into the difficulty calibration but not the energy consumed. Energy consumption is not an input that could be considered by the network because it's not observed globally (or even locally, really).

EDIT: the equilibrium determining whether and how much capital to invest in hashing equipment is not related to 'difficulty' -- it's a human economic factor that considers difficulty as an input.

> The hashrate is an input into the difficulty calibration but not the energy consumed

my understanding is that this is correct - others are introducing "cost of electricity" in an economics statement, not the math of difficulty itself

Any increase in power efficiency will lead to lower operating costs, which will lead to a higher hashrate, which will lead to an adjustment of the mining difficulty.

Energy consumption determines the cost, whereas the benefit is determined by the value of a bitcoin, and the demand for on chain transactions. Cost will always approach benefit.

They are also comparing against some older generation ASICs...the S19 from Bitmain[1] which 'mature' tech today gets 34J/TH, 95TH/s which is like... 40% better?? It's a super weird comparison point that they chose. Or maybe I am reading something wrong here...

[1] https://support.bitmain.com/hc/en-us/articles/900000253583-S...

The S19 Pro is down to 29 J/TH system efficiency, not far off beating the Intel design by 50%.

Indeed, it has a ring of "Hello, fellow youngsters" to it that Intel is even in this space. Maybe AMD and ARM has them thinking they're behind in other areas

I'm surprised that Logitech isn't already shilling Bitcoin Mining Heated Gaming Chairs and Ergonomic Crypto Mice and Mechanical Hashing Keyboards and Bluetooth NFT Whoopee Cushions (every time you sit on it, it mints a fresh NFT for each farting sound it emits, and automatically posts it to your twitter feed).

> Bluetooth NFT Whoopee Cushions (every time you sit on it, it mints a fresh NFT for each farting sound it emits, and automatically posts it to your twitter feed).

Do I still have to pay gas fees to mint a fake fart? How about I rip a real one and we call it even?

The Beats Whoopie Pill functions both as an bluetooth microphone input and speaker output device, with an ultra-low-frequency-high-power woofer. But it's a real pain in the ass to install and uninstall. Fortunately, Kyle Machulis has written an Emacs mode that supports it.

https://www.youtube.com/watch?v=D1sXuHnf_lo

https://imgur.com/gallery/4yEX49T

This product does not even need to be sold in order to boost interest in Intel stock. They can bring it up every quarter and keep the hype going.

The idea that more efficient mining chips will reduce Bitcoin's energy consumption is just wrong. Miners will just buy and use that new chip for as long as the value of the block reward exceeds the energy cost for mining it.

The mining network should always converge to an equilibrium where X$ worth of electricity is consumed for producing (a little more than) X$ worth of BTC, no matter how efficient mining chips are.

Personally I feel that Bitcoin miners are being unfairly targeted whereas the real issue is unclean energy production. Bitcoin is an easy target because a lot of people believe it is overvalued or worthless which gives rise to arguments such as "Bitcoin mining is unnecessarily wasteful".

> Nor does it seem like an amount that can make a dent in the energy consumption problem of Bitcoin mining.

Jevon's paradox indicates that they will make it much much much much worse.

Why sell it at all? They can just mine directly.

I believe this is the modus operandi of most manufacturers; selling new models in refurbished condition after a few months of earning revenue on the hardware. Former Swedish Bitcoin miner brand KnC accidentally admitted this, then backtracked and changed their stance to "extensive quality control".

Such a waste, but the die voltage stacking is interesting. I wonder if GPGPU devices could use the same thing if the cores were sufficiently lockstep.

> That allows an input voltage of 8.875 V to be divided into 355 mV across each ganged ASICs group

It's not novel, it's been used since some of the very first mining ASICS. The reason you don't see it anywhere else is that the matching needs to be very good, down to giving chips busy work if they are idle otherwise you get interesting results.

https://i.imgur.com/hAss6mY.jpeg

They are presumably ganging them at each voltage level to try to reduce the candescence.

Humans are so smart and so deeply dumb.

Bitmain announced the S19XP [1] coming out in the 2nd half of this year with an efficiency of 21.5 J/TH, so Intel still has a lot of catching up to do.

Another news report [2] mentions

> After news of the company's efforts came to light, Intel finally officially acknowledged its blockchain/Bitcoin silicon program, divulging that it already has several large customers for the second-gen chips. That includes BLOCK (helmed by CEO Jack Dorsey of Twitter fame), Argo Blockchain, and GRIID Infrastructure.

> Intel hasn't shared details of the second-gen chips and systems yet, but we do know they are derivatives of the BMZ1 ASICs shown below.

but I'd be surprised if this second generation has bridged the nearly 2.5x efficiency gap...

[1] https://shop.bitmain.com/product/detail?pid=0002021120711304...

[2] https://www.tomshardware.com/news/intel-details-its-bitcoin-...

Great so can you stop buying all the RTX now so casual gamers can maybe get one?

Etherium is mined most commonly on GPUs, and the mining was designed to optimise badly for ASICs.

How long until it moves to proof of stake?

Ethereum has been 6 months away from proof of stake for the last 5 years.

Q2 2022 is most likely. Afaik, one more dev testnet and one or two public ones to go. So far the biggest issue appears to be to achieve sufficient client diversity.

I don't think anyone's mining bitcoin with GPUs.

It doesn't matter that much, people often use auto-switching algorithms that mine whatever is currently most profitable and then swap it for what they wish to hold. It's not that difficult to mine ETH and get BTC for it (there are services but I don't want to promote anything).

That would be mining other cryptos with GPUs and selling those proceeds for btc.

One of the ways people were getting around the LHR/"low hash rate" cards was mining a few types of cryptocurrency simultaneously, which probably fits right in with the auto switching.

Probably not. The era of affordable high-end GPUs is over and I don't think it is coming back even after the entire cryptocurrency-nonsense bubble pops. The expected price of a high-end GPU is now fixed in the >$1k range.

I would take any claims of improvement with a grain of salt until independent verification has been done. This is Intel we're talking about, they have a very long history of overpromising and underperforming when it comes to power/compute efficiency.

What's described here is deeply noncompetitive, even if it was released today. From the pictures of the PCBs in that article I suspect it is the combination of both being incredibly expensive to produce and miserably inefficient.

I wonder whether it'd be effective to build a boiler lined with these so, while you help to launder money you'll, at least, get some hot water.

Co-generation :) I'm sure it's been done.

It wouldn't be hot enough for more than a Stirling engine, but it may be good enough for a bathtub.

Stirling engines... don't get me started. I spent the better part of two years trying to design one that would run on solar power (flat mirror concentrator) to generate power cheaper than solar panels. Conclusion: too marginal to make up for the (then) expected decrease in costs of panels, which came true. Fun to play with though, extremely elegant principle of operation.

Bitcoin is quite unsafe. For example if you send money to a cause you care about, but government disagrees with it, they can disallowlist your wallet and you probably won't be able to touch any coins derived from it ever. If Intel was serious, they should have looked into supporting Monero or other truly anonymous currencies.

Thankfully -- for now -- most merchants do not seem to respect any such blacklist/greylist. However, for exchanges, KYC regulation may end up having some similar effects.

Intel's no dummies. They know that a lot of investors have heard of Bitcoin and not a lot of investors have heard of monero. They didn't get into this business because they wanted to contribute to a decentralized currency. They just want to make money.

Considering that Bitcoin and other cryptocurrencies intend(ed) to operate in a decentralized manner, your point probably matters only for those who want to deal with centralized exchanges that follow such allow or deny lists. There are services and ways to connect with others to trade (though they may come with some more difficulty compared to a centralized exchange).

Isn't this why you have multiple wallets?

If you use one wallet for all your transactions, you're the fool.

You don't carry your checking account in your back pocket,

why would you give a small charitable fund (or anyone!) knowledge of how much you have, and where?

Using multiple wallets is just minor obfuscation, not protection. The funds came from somewhere and that somewhere is also knowable & observable.

Especially with transaction fees being what they are, shuffling money between wallets is just wasting money. And a lot of it.

So take your payments into different wallets,

avoid the "shuffling".

You can also use mixers, or friends, and do "one shuffle", and have the same result.

Point is, the only thing that can be improved here is app UX.

A lot's changed in the 24 years since "Deep Crack".

https://en.wikipedia.org/wiki/EFF_DES_cracker

Maybe somebody interested in retro-crypto-computing could write a Deep Crack emulator for modern GPUs! All the specs and sources for the hardware and software are published in the free book "Cracking DES: Secrets of Encryption Research, Wiretap Politics, and Chip Design". I wonder how fast it could run now -- it would be cool to run it in browsers with WebGPU!

https://news.ycombinator.com/item?id=19916938

In 1998, the EFF and John Gilmore published the book about "Deep Crack" called "Cracking DES: Secrets of Encryption Research, Wiretap Politics, and Chip Design". But at the time, it would have been illegal to publish the code on a web site, or include a CDROM with the book publishing the "Deep Crack" DES cracker source code and VHDL in digital form. https://en.wikipedia.org/wiki/EFF_DES_cracker

https://dl.packetstormsecurity.net/cracked/des/cracking-des....

>"We would like to publish this book in the same form, but we can't yet, until our court case succeeds in having this research censorship law overturned. Publishing a paper book's exact same information electronically is seriously illegal in the United States, if it contains cryptographic software. Even communicating it privately to a friend or colleague, who happens to not live in the United States, is considered by the government to be illegal in electronic form."

So to get around the export control laws that prohibited international distribution of DES source code on digital media like CDROMS, but not in written books (thanks to the First Amendment and the Paper Publishing Exception), they developed a system for printing the code and data on paper with checksums, with scripts for scanning, calibrating, validating and correcting the text.

The book had the call to action "Scan this book!" on the cover (undoubtedly a reference to Abby Hoffman's "Steal This book").

https://en.wikipedia.org/wiki/Steal_This_Book

[...]

aasasd on May 15, 2019 | prev | next [–]

Afaik Phil Zimmermann was one of the first to do it, in '95 through MIT Press—when his PGP circulated a bit too widely for the export regulations. However, the question of him being protected under the 1st wasn't decided in the court.

Reply on news.ycombinator.com