Bitcoin algorithm change - page 2. | Bitcointalksearch.org

Gabi

legendary

Activity: 1148

Merit: 1008

If you want to walk on water, get out of the boat

Quote from: sippsnapp on October 25, 2012, 03:13:54 PM

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.
This way you ban botnets and make asic manufacturers/buyers wet eyes.
I still would like to know for what else beside mining an asic or its technology is useful/applicable, maybe it would be smart to invent a technology if this is not the case that would be useful for other computing operations, with multi cpu this would be the case i guess.

"what else beside mining an asic or its technology is useful/applicable"

Uuuhhh do you know that every hardware piece is an "asic"? Your cpu, your graphic card etcetcetc

And do you know that if you make a ram intensive algo with cpu and gpu you can make an asic optimized for it without problems, exactly like now?

Seriously guys at least learn what asic means and how computers work before "suggesting" things Cheesy

ralree

hero member

Activity: 518

Merit: 500

Manateeeeeeees

Quote from: sippsnapp on October 25, 2012, 05:04:34 PM

Quote from: MysteryMiner on October 25, 2012, 04:59:18 PM

Quote

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.

ASIC always will be more effective at this! That's why they are called Application Specific Integrated Circuit.

Quote

only multi cpu/gpu can solve effectively.
This way you ban botnets

I did not know that computers in botnets have no CPU's or GPU's. The bots probably run on vacuum valves and punch card printers.

Yeah sure, my idea was, almost no conventional pc has 32 or even 64 gb of ram, it was just a thought that would allow to mine cpu/gpu like devices without the so feared botnets.
Thats why i mentioned to eventually make the ram speed a factor, but hell ,yeah, i just dont have enough technical insight to provide really useful contributions here, all i can do is guesswork too.

So if this really is a topic that could be considered people with a technical insight should take over here, otherwise i see no real reason to continue this thread.

So instead of spending money on ASICs I have to buy a bunch of friggin RAM? How does that lower the barrier to entry you're so angry with?

Littleshop

legendary

Activity: 1386

Merit: 1004

Quote from: MysteryMiner on October 25, 2012, 08:23:33 PM

Quote from: Littleshop on October 25, 2012, 08:01:52 PM

Quote from: MysteryMiner on October 25, 2012, 04:59:18 PM

Quote

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.

ASIC always will be more effective at this! That's why they are called Application Specific Integrated Circuit.

Quote

only multi cpu/gpu can solve effectively.
This way you ban botnets

I did not know that computers in botnets have no CPU's or GPU's. The bots probably run on vacuum valves and punch card printers.

While it is true that ASICs can be better at many given tasks, the changing the protocol and VOIDING the ASICs is the attack against them. They adapt (which takes months or even half a year) then you change it again killing them financially. I am NOT FOR THIS, just pointing it out.

And why do that? To have no ASICs in distributed miners possession but to allow government to make the ASICs and attack Bitcoin with 51% attack? Think something like Bombe that cracked Enigma codes in WW2.

I am not for it as said above. Just pointing out the strategy that some may be in favor of. I believe having ASICs made by two or more companies makes bitcoin stronger not weaker then having no ASICs. I also believe in keeping the protocol the same unless there is a problem with it. Having the winners and losers change via the free market is NOT A PROBLEM.

MysteryMiner

legendary

Activity: 1512

Merit: 1049

Death to enemies!

Quote from: Littleshop on October 25, 2012, 08:01:52 PM

Quote from: MysteryMiner on October 25, 2012, 04:59:18 PM

Quote

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.

ASIC always will be more effective at this! That's why they are called Application Specific Integrated Circuit.

Quote

only multi cpu/gpu can solve effectively.
This way you ban botnets

I did not know that computers in botnets have no CPU's or GPU's. The bots probably run on vacuum valves and punch card printers.

While it is true that ASICs can be better at many given tasks, the changing the protocol and VOIDING the ASICs is the attack against them. They adapt (which takes months or even half a year) then you change it again killing them financially. I am NOT FOR THIS, just pointing it out.

And why do that? To have no ASICs in distributed miners possession but to allow government to make the ASICs and attack Bitcoin with 51% attack? Think something like Bombe that cracked Enigma codes in WW2.

Littleshop

legendary

Activity: 1386

Merit: 1004

Quote from: MysteryMiner on October 25, 2012, 04:59:18 PM

Quote

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.

ASIC always will be more effective at this! That's why they are called Application Specific Integrated Circuit.

Quote

only multi cpu/gpu can solve effectively.
This way you ban botnets

I did not know that computers in botnets have no CPU's or GPU's. The bots probably run on vacuum valves and punch card printers.

While it is true that ASICs can be better at many given tasks, the changing the protocol and VOIDING the ASICs is the attack against them. They adapt (which takes months or even half a year) then you change it again killing them financially. I am NOT FOR THIS, just pointing it out.

MysteryMiner

legendary

Activity: 1512

Merit: 1049

Death to enemies!

Quote from: mimarob on October 25, 2012, 07:32:48 PM

Who is actually in charge of such a descision btw?

The majority of users who run Bitcoin nodes.

There was no such decision made previously. Satoshi decided on the initial hashing method and we use it today. Probably if such change is needed then much hated Bitcoin Foundation and Gavin Andersen as a lead developer will start discussion about it.

mimarob

full member

Activity: 354

Merit: 103

Who is actually in charge of such a descision btw?

sippsnapp

sr. member

Activity: 322

Merit: 250

I never had a problem with botnets mining bitcoin, why should i, its just a sideeffect of either asics or higher required ram or whatever that botnets can not participate anymore.

MysteryMiner

legendary

Activity: 1512

Merit: 1049

Death to enemies!

Quote

Yeah sure, my idea was, almost no conventional pc has 32 or even 64 gb of ram

In next 2 to 3 years your average gaming rig might have this much RAM. And for high-end servers this is a norm already. Change Bitcoin algo again then? And it is two different things RAM size and RAM speed. And the largest RAM is often the slowest with increased latencies. The CPU cache is faster than any RAM.

Quote

without the so feared botnets

Why are you afraid from botnets? Are they attacking you? Or are you a part of botnet? Try to rent some you might actually start to love them! Wink

Computing is computing. As long as it plays by bitcoin rules it helps Bitcoin to survive!

sippsnapp

sr. member

Activity: 322

Merit: 250

Quote from: MysteryMiner on October 25, 2012, 04:59:18 PM

Quote

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.

ASIC always will be more effective at this! That's why they are called Application Specific Integrated Circuit.

Quote

only multi cpu/gpu can solve effectively.
This way you ban botnets

I did not know that computers in botnets have no CPU's or GPU's. The bots probably run on vacuum valves and punch card printers.

Yeah sure, my idea was, almost no conventional pc has 32 or even 64 gb of ram, it was just a thought that would allow to mine cpu/gpu like devices without the so feared botnets.
Thats why i mentioned to eventually make the ram speed a factor, but hell ,yeah, i just dont have enough technical insight to provide really useful contributions here, all i can do is guesswork too.

So if this really is a topic that could be considered people with a technical insight should take over here, otherwise i see no real reason to continue this thread.

MysteryMiner

legendary

Activity: 1512

Merit: 1049

Death to enemies!

Quote

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.

ASIC always will be more effective at this! That's why they are called Application Specific Integrated Circuit.

Quote

only multi cpu/gpu can solve effectively.
This way you ban botnets

I did not know that computers in botnets have no CPU's or GPU's. The bots probably run on vacuum valves and punch card printers.

ralree

hero member

Activity: 518

Merit: 500

Manateeeeeeees

Quote from: FreeMoney on October 24, 2012, 03:34:17 AM

It's like a beautiful castle was built and everyone wants a job as a janitor.

You aren't forced to provide hashes to help secure the network, if you don't want to do it in the most efficient way, then do it at a loss or don't do it.

+1

sippsnapp

sr. member

Activity: 322

Merit: 250

I still like the idea of a ram intensive algo (eventually ram speed could play a role) maybe combined with something only multi cpu/gpu can solve effectively.
This way you ban botnets and make asic manufacturers/buyers wet eyes.
I still would like to know for what else beside mining an asic or its technology is useful/applicable, maybe it would be smart to invent a technology if this is not the case that would be useful for other computing operations, with multi cpu this would be the case i guess.

ajareselde

legendary

Activity: 1722

Merit: 1000

Satoshi is rolling in his grave. #bitcoin

Quote from: Etlase2 on October 25, 2012, 12:31:23 PM

Quote from: MysteryMiner on October 25, 2012, 12:07:04 PM

I have read that Bitcoin can relatively easy replace SHA256 with another hash algorithm if SHA256 security is at question. This will cause all clients to update but it will not change how Bitcoin operates.

Relatively is a relative word. It is a hard fork and not an easy thing at all to accomplish. It will also void all ASIC hardware.

Voiding asic mining was the general idea

it is not so super duper awesome to everyone.
cheers

Etlase2

hero member

Activity: 798

Merit: 1000

Quote from: MysteryMiner on October 25, 2012, 12:07:04 PM

I have read that Bitcoin can relatively easy replace SHA256 with another hash algorithm if SHA256 security is at question. This will cause all clients to update but it will not change how Bitcoin operates.

Relatively is a relative word. It is a hard fork and not an easy thing at all to accomplish. It will also void all ASIC hardware.

MysteryMiner

legendary

Activity: 1512

Merit: 1049

Death to enemies!

I have read that Bitcoin can relatively easy replace SHA256 with another hash algorithm if SHA256 security is at question. This will cause all clients to update but it will not change how Bitcoin operates.

Etlase2

hero member

Activity: 798

Merit: 1000

Quote from: Come-from-Beyond on October 25, 2012, 10:31:13 AM

If an attacker reveals a way how to make an attack. He will, likely, keep it in secret. Also, if the attack lets to find a hash with ALL zeros then "difficulty" will make no sense anymore.

This is not how attacks on hashing algorithms work. At least, no one has ever studied the possibility of it because it would normally be retarded. In bitcoin's case, it would make a bit more sense, but I don't know if such a thing is even possible. Hashing algorithm attacks fall under the category of collision attacks, where m1 != m2 but h(m1) == h(m2) or pre-image attacks such that given a hash h find a message m that hashes to it. There is no such attack as "given m + nonce, find h < difficulty faster than brute force". Even MD4 and 5 are not broken in this sense at all and could potentially be used for bitcoin without repercussions other than being only 128-bit.

Come-from-Beyond

legendary

Activity: 2142

Merit: 1010

Newbie

Quote from: MysteryMiner on October 25, 2012, 11:29:04 AM

But in my opinion the greatest threat for Bitcoin existence is global internet blackout. Bitcoin might be completely decentralized, the internet infrastructure is very centralized.

I agree. Peer-to-peer (WiFi) network of mobile devices could help us though. But it's for discussion in other thread.

MysteryMiner

legendary

Activity: 1512

Merit: 1049

Death to enemies!

Quote

If an attacker reveals a way how to make an attack. He will, likely, keep it in secret. Also, if the attack lets to find a hash with ALL zeros then "difficulty" will make no sense anymore.

The probability of SHA256 being completely cracked is very low. All zeroes will not make sense of difficulty anymore, that is completely true.

But in my opinion the greatest threat for Bitcoin existence is global internet blackout. Bitcoin might be completely decentralized, the internet infrastructure is very centralized. ISP providers must comply with ever increasingly totalitarian laws to operate. Most countries in European Union have silently passed laws that enable total internet kill switch in "cases of emergencies, natural disasters or civil unrest". This might not be relevant to original question about SHA256 being cracked, but this total blackout is more realistic threat to be worried about. The alien invasion also can cause Bitcoin to fail because aliens might start >50% attack with spaceship's onboard computer, but we don't think that this is a way how Bitcoins will fail, right?

Come-from-Beyond

legendary

Activity: 2142

Merit: 1010

Newbie

Quote from: MysteryMiner on October 25, 2012, 10:10:42 AM

Quote from: Come-from-Beyond on October 25, 2012, 09:22:52 AM

Quote from: Etlase2 on October 25, 2012, 09:12:07 AM

Quote from: Come-from-Beyond on October 25, 2012, 09:05:38 AM

Sorry, but I disagree. If coming up with a hash with lots of zeros is faster than simply finding a nonce,

But this is extraordinarily, unbelievably, impossibly unlikely. The block has to be 1) a valid bitcoin block, which heavily limits what data can be used to find a collision and essentially goes back to using a nonce, and 2) limits you to ONE SPECIFIC hash whereas searching for ANY hash with the correct leading number of zeros is many, many magnitudes easier. Bitcoin mining is essentially already a partial-collision attack.

I mean other type of attack. Not attempt to find a collision for an existing block, but attempt to find nonces for new ones with insane rate.

So the miner software needs to be updated that will find new blocks via attack method. This is still serving as a valid proof-of-work and the difficulty will adjust for new block rate.

If an attacker reveals a way how to make an attack. He will, likely, keep it in secret. Also, if the attack lets to find a hash with ALL zeros then "difficulty" will make no sense anymore.

Topic: Bitcoin algorithm change - page 2. (Read 7819 times)