Packman1444 (OP)
Newbie
Offline
Activity: 4
Merit: 0
|
|
June 30, 2013, 10:00:40 PM |
|
Out of curiosity, does anybody know how powerful the Bitcoin Network is? I am not sure if it can be give in FLOPS, but if it can, I would like to know.
|
|
|
|
|
|
"This isn't the kind of software where we can leave so many unresolved bugs that we need a tracker for them." -- Satoshi
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
June 30, 2013, 10:09:48 PM Last edit: June 30, 2013, 10:27:23 PM by DeathAndTaxes |
|
The network right not has a combined computing power of ~170 TH/s (terrahashes per second) or 170 trillion double SHA-256 hashes attempted each second. Sometimes people like to provide a FLOPs equivalent but the reality is that FLOPS is a measure of floating point math and the Bitcoin network combined executes exactly 0 FLOPS a second.
If you need something tangible a high end GPU (HD 7970) does about 700 MH/s so the network has computing power equivalent to ~250,000 high end GPUs (170,000,000 / 700 ). Obviously the whole network isn't made up of only 7970s but it provides a reference point.
|
|
|
|
J35st3r
|
|
June 30, 2013, 10:23:26 PM |
|
Right now, 186THash/sec http://bitcoincharts.com/bitcoin/Unfortunately FLOPS are not a good metric for bitcoin hashing as its intrinsically an integer operation, rather than floating-point. Perhaps the hardware comparison list is a good place to start https://en.bitcoin.it/wiki/Mining_Hardware_ComparisonLook at the hash rate for a typical CPU and scale it up to the full network hash rate. So a top range intel processor give 5MHash/s per core. which equates to 37 million cores (or 9 million quad-core processors) for the full 186 THash/sec. Of course nobody mines with CPUs these days as GPUs. FPGAs, ASICs all give (in order of magnitude) far greater performance per dollar/watt/joule metric. If you want to compare with the top 500 supercomputers (20 PetaFlop leader), its difficult as the Petaflops tell you nothing about the integer performance. Hmm, getting out of my depth here, I guess the integer performance figures are out there somewhere, and then its a matter of scaling up vs a known CPU hash rate. Oh well, maybe experts will chip in, they do haunt newbie land from time to time. And since I've made the effort to write all this, I'm going to post it anyway. Push.
|
1Jest66T6Jw1gSVpvYpYLXR6qgnch6QYU1 NumberOfTheBeast ... go on, give it a try
|
|
|
Packman1444 (OP)
Newbie
Offline
Activity: 4
Merit: 0
|
|
June 30, 2013, 10:29:39 PM |
|
Right now, 186THash/sec http://bitcoincharts.com/bitcoin/
Unfortunately FLOPS are not a good metric for bitcoin hashing as its intrinsically an integer operation, rather than floating-point. Perhaps the hardware comparison list is a good place to start https://en.bitcoin.it/wiki/Mining_Hardware_ComparisonLook at the hash rate for a typical CPU and scale it up to the full network hash rate. So a top range intel processor give 5MHash/s per core. which equates to 37 million cores (or 9 million quad-core processors) for the full 186 THash/sec. Of course nobody mines with CPUs these days as GPUs. FPGAs, ASICs all give (in order of magnitude) far greater performance per dollar/watt/joule metric. If you want to compare with the top 500 supercomputers (20 PetaFlop leader), its difficult as the Petaflops tell you nothing about the integer performance. Hmm, getting out of my depth here, I guess the integer performance figures are out there somewhere, and then its a matter of scaling up vs a known CPU hash rate. Oh well, maybe experts will chip in, they do haunt newbie land from time to time. Ans since I've made the effort to write all this, I#m going to post it anyway. Push. This link provided the hash rate in petaFLOPS, unless it is not what I am thinking. I was curious because I wanted to see where the network would rank in the top 500 supercomputers lol.
|
|
|
|
J35st3r
|
|
June 30, 2013, 10:37:01 PM |
|
This link provided the hash rate in petaFLOPS, unless it is not what I am thinking. I was curious because I wanted to see where the network would rank in the top 500 supercomputers lol.
You are right, I didn't notice it there at the bottom ... Network Hashrate PetaFLOPS 1970.61 They must be using some sort of equivalence factor (because as DeathAndTaxes said, bitcoin SHA256 hash makes no use of floating point operations). But that's your answer anyway ... roughly 100 of the top of the range supercomputers (unfortunately there only tends to be one of those available at any one time )
|
1Jest66T6Jw1gSVpvYpYLXR6qgnch6QYU1 NumberOfTheBeast ... go on, give it a try
|
|
|
BurtW
Legendary
Offline
Activity: 2646
Merit: 1131
All paid signature campaigns should be banned.
|
|
June 30, 2013, 10:40:36 PM |
|
Yes, it is a direct mathematical conversion of the network hash rate to a controversial FLOPS number. This tries to estimate the equivalent rate. Like I said the conversion is controversial.
|
Our family was terrorized by Homeland Security. Read all about it here: http://www.jmwagner.com/ and http://www.burtw.com/ Any donations to help us recover from the $300,000 in legal fees and forced donations to the Federal Asset Forfeiture slush fund are greatly appreciated!
|
|
|
dree12
Legendary
Offline
Activity: 1246
Merit: 1077
|
|
June 30, 2013, 10:47:13 PM |
|
Yes, it is a direct mathematical conversion of the network hash rate to a controversial FLOPS number. This tries to estimate the equivalent rate. Like I said the conversion is controversial.
It's probably a gross overestimate. ASICs are incapable of producing FLOs at all, so generate 0 FLOPS. Considering that they make up more than half of the network, the FLOPS equivalent of Bitcoin is likely much lower than estimated.
|
|
|
|
Foxpup
Legendary
Offline
Activity: 4354
Merit: 3042
Vile Vixen and Miss Bitcointalk 2021-2023
|
|
July 01, 2013, 05:23:42 AM |
|
The FLOPS conversion is useful in the sense that it provides an accurate indication of how powerful a general-purpose computer would have to be in order to carry out a 51% attack. The fact that the network is neither entirely composed of general-purpose computers nor even executes any floating-point operations at all is irrelevant to that question.
|
Will pretend to do unspeakable things (while actually eating a taco) for bitcoins: 1K6d1EviQKX3SVKjPYmJGyWBb1avbmCFM4I am not on the scammers' paradise known as Telegram! Do not believe anyone claiming to be me off-forum without a signed message from the above address! Accept no excuses and make no exceptions!
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
July 01, 2013, 05:32:02 AM |
|
The FLOPS conversion is useful in the sense that it provides an accurate indication of how powerful a general-purpose computer would have to be in order to carry out a 51% attack. The fact that the network is neither entirely composed of general-purpose computers nor even executes any floating-point operations at all is irrelevant to that question.
The issue is there is no standardized IOPS to FLOPS conversion ratio. Any such ratio is merely speculation. The ratio between integer and floating point performance varies dramatically even when just looking at CPUs from a single manufacturer.
|
|
|
|
|