Bitcoin Forum

Hello everyone,

my name is Rosa, I'm studying International Business in Germany and am currently writing my Bachelor thesis 'The disruptive potential of Blockchain on economy'. I would like to ask a clarification question concerning the current computing power of the bitcoin network (resp. the network hashrate).
Please correct me if I'm wrong with the following:

1. The Bitcoin network hashrate is 45,988,115.27 petaflops ( http://www.bitcoinwatch.com/ ).
    45988115.27 petaflop = 45988.11527 exaflop ( http://endmemo.com/convert/computer%20speed_s.php )
2. The fastest super computer works with 93,014.6 petaflops ( https://www.top500.org/resources/top-systems/sunway-taihulight-national-supercomputing-center-i/ ) Cost for the system (excluding running costs: US-$ 270 million ) ( http://www.netlib.org/utk/people/JackDongarra/PAPERS/sunway-report-2016.pdf )
3. Hence: All nodes of the bitcoin network combined are 494 x more powerful than the super computer? ( 45,988,115.27/93,014.6=494.422)

It just seems to be so overwhelming that it makes me doubt my own calculations.

I'm not an expert on technical details (though I'm trying to become one;), so I would be really grateful for your feedback:)
Thank you for your time in advance. I appreciate your effort to help me :).

Rosa

Actually your calculations are quite true. According to bitcoinwatch's 2013 report, the Bitcoin Network is more powerful than the world's top 500 supercomputers combined, and that's because Bitcoin is a massive computing network.

I do agree with that since even by using simple logic if you compute all of the bitcoin nodes when their numbers combined we can get a result much more powerful and stronger than the supercomputers. But if you increase the numbers of supercomputers or the CIA will launch their Quantum computers which is a thousand times more powerful than supercomputers then probably it will defeat all the combined nodes not only in number but in power.

Dear Xester,
thank you for your inspiration with the CIA supercomputers, I have not thought of this massive computer power source. Maybe I'll be able to find some figures to compare. I appreciate your effort!


Dear FlamingFingers,
I totally agree with your2013 report, I just wanted to upgrade those old calculations to 2017. Thank you for your time! This helped me a lot.

 A couple of points:

1) Nodes are not miners.  Miners do the hashing (ie most of the FLOPs)

 Your statement  "Hence: All the nodes of the Bitcoin..." is false.

2) Sunway is 93,014.6 TFlop/s not PFlops (this becomes 93.0146 PFlops)

3) Why are you converting PetaFlops to ExaFlops?!

4) 45,988,115.27 PFlops divided by 93.0146 PFlops =  494418.24477017586486422561619359

 So it would seem that All the Bitcoin miners combined have 494,418 x the power of the fastest supercomputer.
 
4b) Your number are now correct but keep in mind that the mining hardware on the Bitcoin network is specifically designed to do only one thing - the SHA256 hash function - using ASIC chips (Application Specific Integrated Circuit).  Your iPhone would be much more powerful than the Bitcoin network for finding the square root of 93,014.6... infinitely more powerful I bet!


 Hope that helps you to achieve technical prowess.

It's not an Apples to Apples comparison.


Supercomputer performance is measured in FLOPs, which is an abbreviation for floating point operations per second. But the Bitcoin mining rate is measured in hashes per second.


The Bitcoin hashing algorithm for calculating the next block header (which miners invoke to produce the overall network hashrate) is not a floating point operation, it is integer based arithmetic, not floating point arithmetic.


So to make a comparison, assumptions have to be made about equating 1 FLOP to 2 SHA-2 hashes (a single Bitcoin block header involves 2 SHA-2 hashes). The assumptions one makes alter the equation, so it's not really a true empirical measurement.

Just reiterating what Carlton Banks wrote...

The Bitcoin network hash rate is not 45,988,115.27 petaflops. It is 3638726.32 terahashes/second. FLOPS (floating point operations per seconds) and hashes/second are different and are not directly comparable, especially since computing a hash involves no floating point operations.

if i'm not mistaken, one 'hash' here means a complete blockheader (SHA-256) hash, which consists of 64 'turns' or 'rounds', each of which is made of about 2 dozen actual operations.

At least at the Bitcoinwatch web site they already did the conversion THash/sec to PetaFLOPs.

I don't know if they did it correctly, but then you must blame the website admin, not the OP.
 

isn't this kind of a pointless comparison? supercomputers can do anything they're tasked to do. an asic is a useless brick for anything other than calculating bitcoin. it's like comparing a sprinter to a marathon runner.

The reason for that is because Bitcoin miners are hugely profit motivated.  We're talking about a Bitcoin being worth over $1000, and about 144 blocks with 12.5 Bitcoin plus transaction fees being mined every day we're talking about around two million dollars per day of profit motivation which can be spread out among thousands of people easily and dedicated to one simple purpose.  Supercomputers will never have to be as fast as the whole Bitcoin network - that would be insane.

In which case, you mean 128 rounds, not 64. Bitcoin SHA-2 ASICs perform x2 SHA-2 256 hashes for each block header calculation, not x1.

supercomputer can't mine bitcoin if that is the point of your post, they are far less efficient and slow compared to asic, also they would require huge resource just to mine few satoshi not worth it

basically right now bitcoin mining ecosystem is a whole giant supercomputer that do only one task, which is generate more bitcoin, not talking about node, which only support the network

maybe a better comparison could be made between a supercomputer and all the computer that run a full nodes

OP, why are you calling your thesis " 'The disruptive potential of Blockchain on economy'" instead of " 'The disruptive potential of Bitcoin on economy'?

Hello everyone,

thank you for your time. I made notes to everything you said and try to make conclusions for my thesis.

To sum up what my intention is: I want to make a comparison (e.g. supercomputer, bitcoin mining network) to demonstrate the potential of Bitcoin/Blockchain.

I'm happy if anyone gives me information that I can use (and understand;) to make a persuasive analogy.
_______________________


Dear xhomerx10,

I appreciate your input:
1. True. I mixed up the network and the network of miners. Thank you for your feedback.
2. You're correct, Sunway is 93 Pflop/s.
3. Idk. Seems like this was pointless. Probably bc I'm not good in converting.
4. I would love to use this "494,418 x times the power" but as Carlton Banks stated, I can not make this comparison.
4b. Good point, I'll keep in mind the SHA256 hash.
Well, I'm happy if I'll manage to understand the very basic technical background:)
______________________

Dear Carlton Banks,

what you mentioned is an important reminder (although I can not really make sense out of it, due to my lack of understanding the difference of a floating point operation and an integer based arithmetic) How do I compare 1 Flop. Can you give me an idea of how those two can be compared to bring up a powerful conclusion? I want to make a statement pointing out how powerful the Bitcoin network is, in comparison to supercomputers.

If I understood your assumption to alter the equation correctly, then...

... 1.45,730,934 (Network hashrate in Petaflops) / 2 (bc a single block header involves 2 SHA-2 hashes) = 22,865,467 (headers).
2. 9.3e+16 /22,865,467 = 4,067,268,776 (what ???)

_________________________

Dear odolvlobo, dear jonald_fyookl,

I bet it's important what you mentioned. I can see that there must be a difference between petaflops and terahashes/second and 64 'turns' in one header. I can just not make sense out of it, I'm sorry, that I can not follow you.

___________________________


Dear European Central Bank,

thank you for your question. The comparison is not pointless. Let me explain to you why.
I bet to you and everyone familiar with encryption, computing language and technical knowledge (I guess most of you guys), the comparison is useless, because you know the details and program and do things I have no clue about ( I'm bad in maths, but ok in macroeconomics)

But there are many people out there that should understand what Bitcoin/Blockchain is about. They should be able to feel the impact the open ledger can have, not just for finance but so many other industries that we can't even imagine now. I need the comparison between the supercomputers and the bitcoin network to delight the minds of non-experts. And even we can, to some point, imagine if a US-$ 270 million supercomputer loses against Bitcoin, that must be wicked stuff. I'm writing this thesis to give economists some fundamental information about Blockchain.

With your name (eCENTRALb, an institution that's not amused about a peer-to-peer network invented to substitute them) you should see where I'm going.

_______________

Dear Qartada,

thank you for your contribution. Mining itself is interesting, unfortunately, I can not focus on all the different points of interest of Bitcoin.

_____________________

Dear Amph,

I understand your point of view. The bitcoin mining ecosystem is a whole giant supercomputer. This is, why I was trying to make a comparison in the first place.

I agree, that a comparison with all the full nodes would work too. I'm not sure whether I'll find figures for all full nodes combined though.

______________________

Dear AGD,

The title of the thesis is focusing on Blockchain technology instead of bitcoin because I want to show the impact for economy with all the industries and not just finance. Finance is, by now, the most discussed industry which will be impacted by Bitcoin (or already is). I admire Nakamoto for having this outstanding, positive change idea of a public ledger. The idea behind bitcoin is what I'm concentrating on.

Humans prefer to waste resources for pursue of money than pursue to end sufferings.

@Rosa_Weinhold This is such an interesting bachelor thesis in the field of international business, Rosa. We learnt lots of things due to your topic. Good luck with this.

As a guy said above, bitcoin or altcoin miners cannot do any task given. They're only programmed to make coin realted calculations. Yet a supercomputer can do anyhing (task) given.

Your thesis is very good, it contains all the necessary information and accuracy. I see the investment of your time, you have thoroughly explored the information about it, so, you have a complete project. If I'm the one who reviews and censors for your scheme, I'll give you a good result, which is perfectly worth it for you. I do not see any mistakes. Wish you have a good result with it.

go search wiki on SHA-256 hash

Dear Rosa.
"Blockchain Technology" is just one fragment of the "Bitcoin Technology", where "Blockchain" is mostly refering to the decentralized open ledger system of Bitcoin. The industry has no interest to let everybody else know, what is written in "their" Blockchain. I.E. a bank would have no interest to establish an open decentralized ledger for obvious reasons. A regular company wouldn't release customer data in an open blockchain for the same obious reasons.
Now in the end, when most people (esp. bankers) refer to "Blockchain technology" instead of saying "Bitcoin", they do it because they can't control Bitcoin and they would like to do the same as Bitcoin, but NOT decentralized and NOT open source, which makes no sense. Bitcoin works because it is open source, decentralized, (pseudo-)anonymous and doesn't need a third party, like a bank for a transaction.

edit: I forgot the most important factor for the early rise of Bitcoin: Bitcoin in the early days made people earn money just by running a miner on a PC.

This is the best comparison I can do.

The top supercomputer, the Sunway TaihuLight, uses SW26010 manycore RISC processors... I dont even fully understand what that means but Id guess its some proprietary CPU and isnt easy to compare. The #2 supercomputer is the Tianhe-2 which uses Intel Xeon E5-2692 CPUs. This we can compare so thats what my math is based on.

I couldnt find any Bitcoin specs on the E5-2692 specifically but I did see that based on this website, a very optimistic number for Bitcoin mining would be 50 MH/s per CPU. More realistic might be 40 MH/s but lets go with 50 MH/s for our math.

https://en.bitcoin.it/wiki/Non-specialized_hardware_comparison#Intel

The Tianhe-2 uses 32,000 CPUs to operate. https://en.wikipedia.org/wiki/Tianhe-2
At 50MH/s each, thats 1,600,000MH/s or 1,600GH/s or 1.6TH/s of Bitcoin hashing power.

This supercomputer consumes 17.6MW of power, 24MW if you include the cooling. (from the wiki)

The Bitmain S9 Bitcoin ASIC produces 13.5TH/s and uses 1300 watts of power. https://shop.bitmain.com/productDetail.htm?pid=00020170405095325800pA1af37a0667


If you calculate the supercomputer based on the 40MH/s hashing speed per CPU then you get a combined hashing speed of 1.2TH/s.

This shows that the S9 is between 8 and 12 times faster at Bitcoin mining than the second fastest supercomputer in the world. Its also over 18,000 times more efficient.



*Someone check my math.*

Hello Haladay,

Many thanks for the wishes.

_____________________

Hello naughty1,

so far the thesis isn't anything but almost half-done. I appreciate your wishes though. The grade is not that important either, it's just exactly what I want to do right now.

______________________

Dear jonald_fyookl,

I'm aware of the crucial importance of SHA-256 already, I just thought the '64 turns' were miner relevant to my question. I didn't meant to be rude :)

_______________________

Dear AGD,

thank you for your time. I welcome your thoughts.
Maybe we are talking about the same thing but just use different words. "Bitcoin" is one (and the first) application of "Blockchain Technology". I agree with you, that the decentralized nature of the open ledger is important.

Your statement "the industry has no interest to let everybody else know, what is written in "their" blockchain." is correct. And it's an argument why we shouldn't wait to let the "the industry" (e.g.banks, insurances, google) run blockchain and take advantage of a technology that was on purpose build to replace those intermediaries. We all can only assume what Nakamoto's actual intentions were, but there are some clear pieces of evidence, that he(or she;) intended to, not only criticizes, but make those third parties redundant.

I couldn't agree more with your second statement "they can't control Bitcoin"
The financial sector would kill to grab Blockchain and reshape it for their profit.

I'm trying to give you an idea, why I'm writing this paper. I would like to do my share and spread the word about the Blockchain/Bitcoin/Etherum. I really believe that the more people do understand the less fear of change they'll have. The less power banks would have. The more Nakamoto's idea of an economy/currency without central authorities would be realized.
It's not an option for me to accept the status quo of what a regular company would/wouldn't do. Once people do know what Blockchain can do, they will embrace it.

Well, and lastly, I do not know much about mining business. Those people (maybe you are one yourself?) that were running miners on a PC, were they really doing it for money cashout, which wasn't that much back then? I can only speculate on the intentions of miners nowadays and what/why they are doing it (pay off? likelihood to be the creator and earn US-$ 13,000 )

How do you feel about companies adapting Blockchain for their business? You're obviously aware of the reservation of "old" companies for new change possibilities. I'm curious what your future outlook is...

Dear AmDD,

Wow! I'm impressed. Thank you for your comprehensive calculation.
Yet I'm not an expert, I can not really judge whether your estimation is correct.
The comparison is interesting. I wasn't quite familiar how effective a specialized mining computer is. It was not my question, how good at hashing a supercomputer would be. But I wrote down your figures, I'm not finished with the chapter about mining, I'm thinking about integrating your analogy.

Unfortunately, my question from the beginning is not answered by your post.
I want to make a comparison to show how effective the entire Bitcoin mining network is (e.g. by contrast with supercomputers, or anything John Doe can relate to). It needs to be a comparison that non-experts (like me) can easily relate to ::)

_____________________________

To everyone still not exhausted about my stubborn focus on this comparison: Thank you for your time. I wouldn't know where/who else I could ask this questions, as my Professors are non-experts like me.

Regarding FLOPS vs. hashes/second, here is an analogy:

The Fletton brick company can produce 150 million bricks per year.
Gary Works has a capacity of 7.5 million tons of steel per year.

Both bricks and steel are building materials, so how do you compare the building material capacity of the two? In bricks or tons of steel or something else?


In computational capacity terms, how do you compare floating-point capacity to hash capacity?  They both measure computational capacity, but are not directly comparable.

1 "hash" is 64 rounds of (roughly) 20 operations.  20 x 64 = 1,280 but these are bitwise operations (I believe) vs floating point.

This article https://lwn.net/Articles/286233/ says we get about 1000 bitwise per flop... so lets call this 1.28

now multiply by the approximate 4M terahashes a second the network does... thats 5.1M terraflops.... or 5100 petaflops..

the fastest supercomputer has a peak speed of 93 petaflops. https://en.wikipedia.org/wiki/Supercomputer

To compare to the entire Bitcoin network, it would like this:

total Bitcoin network hashing power is about 3,800,000 TH/s or over 281,000 Bitmain S9 machines. It would take 2.3-3.1 MILLION of the Tianhe-2 super computers combined just to match the power of the bitcoin network.

In terms of cost, the S9's would cost $365 million while the supercomputers cost... a lot. 2.3 million X $390 million = $897,000,000,000,000. 3.1 million T2's would cost $1,209,000,000,000,000!!!

Im not sure I know any other way to compare the power of this global network other that in terms of cost or number of "normal CPUs".
I hope this is of some use to you.

.....and hopefully you can now see my original point more clearly.


There is more than one way to compare FLOPs with distributed SHA-2 hashrates, and none of them are definitive. All comparisons between the two are arbitrary to some extent, and yield different factors of difference between supercomputer performance and Satoshi's design for a distributed trust network.

 Ah, nobody understands money; that's why we have Bitcoin. ;)

I actually like where you're going with this! So I was able to find some information that should help to get more exact numbers.

 Specs on the E5-2690 (https://en.bitcoin.it/wiki/Non-specialized_hardware_comparison) say 66 MH/s for a dual processor setup.
Here is a link (couldn't post it as a link due to the square braces in the url)
to a comparison between the E5-2690 and the E5-2692.  Image is below but if you need to cut and paste specs, use the link.

 https://i.imgur.com/ThCur1T.png?1

 I think it would be fair to do a real comparison between the Tianhe-2 and the Bitcoin network for this one specific application of securing the Bitcoin Blockchain.  The only thing is, we also need to find out about the 48,000 Xeon Phi 31S1P.  I know they were used in a PCIe card for a co-processor of sorts but I'm having trouble finding it.  Does anyone know what I'm referring to?
 Never mind! I found it -

https://i.imgur.com/fFNi6Fh.jpg?2

 Now I know somebody tried to mine Bitcoin with one of these puppies!  Anyone have some specs?

 Never mind again - found it!  (thread was locked so I couldn't quote)

https://i.imgur.com/6nIWHHI.png?1

 Now we can compare hashes to hashes - instead of Flops to ~Flops

We know the hash rate of the E5-2690 from real-world experiments to be 66 MH/s, similarly we know the hash rate of the Phi to be 140 MH/s
Using the ratio of the benchmarks for the E5-2690 and the E5-2692 and the hash rate of the E5-2960 we can calculate the hash rate of one E5-2692

to be (16018/19362)*66MH/s/2 = 27.300589

 We know there are 32,000 E%-2692 processors
 
  32,000 * 27.300589 = 873,619 MH/s
  
and there are also 48,000 Xeon Phi 31S1P @ 140 MH/s

48,000 * 140 MH/s = 6,720,000 MH/s  for a total of

 7,593,619 MH/s or 7.593619 TH/s
 
The bitcoin network (http://bitcoincharts.com/) is presently running at 4,136,474.291 TH/s

4136474.291 TH/s / 7.593619 TH/s = 544,730

 So the Bitcoin network is 544,730 times faster than the world's 2nd fastest super computer, Tianhe-2, at securing the Bitcoin Blockchain based on real-world measurements.

Rosa.

Every company uses FIAT banks to administrate their financial stuff. Their taxes are paid in FIAT and almost all of the suppliers/farmers etc for the industry are in need to use FIAT to pay their bills/tax. Now when banks (and their partner governments) are the declared enemy of Bitcoin (written in the Genesis Block)  I would not be surprised, when we see a similar prohibition wave, like te one we saw at the almost world wide prohibition of marihuana.

If you google ECB and Bitcoin, you will find some interesting articles on how Europe might treat the "danger of Bitcoin":

http://www.reuters.com/article/us-ecb-bitcoin-eu-idUSKCN12I1HC


Interesting how they try to point the attention away from "Bitcoin" to "digital currencies such as bitcoin". They even avoid to write Bitcoin in capital letters hrhrhr
Also interesting is, that all the regulations that can be applied, are limited to the point where Bitcoin is exchanged into FIAT (exception would be to shut down the internet to get rid of this evil internet money). If the complete supply chain of the industry would accept Bitcoin, there would be no need for FIAT (and banks) and the governments would have big problems to collect their tax. People could simply hide their wealth from their governments and nobody could force people to pay their tax (besides torturing to extract the private keys, of course).

Another interesting fact is, that the descibed dangers of Bitcoin for the government (crime and terror financing, money laundering and tax evasion) can exactly be applied to FIAT, which would be no reason to prohibit FIAT ... of course.

So, most banks atm are avoiding customers, who are making their business with Bitcoin and meanwhile bankers are making jibberish about "How the Blockchain concept will revolutionize the financial sector", without any clue about, what "The Blockchain Concept" really is.

 Hello odolvlobo,

I know that it is difficult to compare two different measuring systems. That's why I was looking out for someone with technical knowledge to help me out. Fortunately, xhomerx10 managed to do so  :) Nevertheless, thank you for your participation in the discussion :)

__________________

Dear jonald_fyookball,

Thank you for your help! And that you explained it in a way I could follow it so easily! That’s great and very useful for me! I appreciate your time.

__________________

Dear AmDD,

Your calculations (and you posted several ones) are really very helpful to me! Thank you for your support so much! The comparison in terms of costs and CPU is relatable even for non-experts like me. Exactly what I was hoping to find. Thank you for your time

I’ll express my gratitude by buying you a beer via your BTC tip jar ( as soon as I figured out how to actually send Bitcoins instead just writing about it, I’ll pm you;)

______________________
Dear Carlton Banks,

I think I’m able to relate to what you said, now that I have some additional information. The comparison is not ideal. But it does work (I know I’m repeating myself over and over) for Blockchain newbies.

_______________________
Dear xhomerx10,

Finally, you came up with the number I was desperately looking for: 544,730! I’m lucky that you adopted AmDD’s calculation on a level that I’m able to follow. Thank you very much for taking your time to empathize with a non-expert like me  I’m happy to buy you a beer, too, just to express my gratitude somehow

___________________
Dear AGD,

I can follow your chain of thoughts on our fiat money based system. I don’t want to weaken the strength of your arguments, stressing how powerless consumers are in this system are. I don’t see a solution for this problem in the system (i.e. capitalism). I’m more interested in the question how Bitcoin can support moving to a new economic system without capitalistic pillars.
We already do see prohibitions on Bitcoin exchange in the UK. The ECB is completely behind with an appropriate reaction (which leads to hilarious statements as you quoted). But Estonia and Philippines are already welcoming the use of digital currencies and implement regulations. Which shows that it’s not an issue for a state to deal with Bitcoin, if it wants to.

Collecting taxes would work for states using this three way bookkeeping system. Probably even better than it does now. From my personal point of view paying taxes is an appropriate way to give back to society, but I don’t want to argue about the deficits of tax distribution key in this discussion.

The question is: Can you, AGD, imagine a different world? Right now! If the answer is ‘yes’ change is achievable. If it is a ‘Theoretically yes, but, No, but what if’ no change is realistic. It begins in our heads. This is where the fences begin. We could argue about the dangers (crime, terror financing, money laundering) for years. Or we can decide to make it work. It’s a glass half full/half empty question broken down. So please understand, that I don’t want to get into more detail with the different topics you brought up (prohibition wave, money supply, legislative bodies).

I totally agree with you, that all those dangers apply to FIAT. And I see this as a proof, that people are just afraid of Bitcoin, which they are scared to confess so they come up with crazy arguments. Fear of fear or fear of change is within all of us, it’s our challenge to face it.

But allow me to come back to one sentence “People could simply hide their wealth”. People are already hiding their wealth. That doesn’t stop them from becoming President of the United States. But IF a government decided to set up regulations for Bitcoin, decided to analyze, build trust, have faith in a decentralized system, the downsides for the state could be reduced.

AGD, I’m confident that you are a Bitcoin expert and I’m not. I’m not trying to reduce your arguments because I’m pretty sure, you and I are both on the Pro-Bitcoin block. I’m trying to give you some other points of view, some pro arguments that you maybe haven’t thought of by now.


___________________________________________________

Eventually I would like to thank everyone involved in this discussion. You helped me a lot, with your calculations, pointing out my mistakes and of course by answering the questions, which –without you- I couldn’t have answered. So thank you  #blockchange

P.S.: I’m happy to share my bachelor thesis with anyone interested. If so, please let me know, because I’m not sure whether it would be a matter of interest to post in a new thread solely for this purpose.

im sure the community would enjoy reading your thesis when its ready

So on today figures

Bitcoin Hashrate: 7,200,000 TH/s
Antminer S9 can do13.5TH/s
To produce the network hashrate you need 533,333 Bitmain S9 machines, which would cost you 675million and about 4million in electricity each day

While Bitcoin marketcap $73.5 billion what deter attackers from performing 51% attack?

if not on Bitcoin but on a smaller network (crack the numbers for litecoin and you find it cost you about 35-40million/day for 51% attack on cryptocurrency with 2.8B marketcap!)