Technical information on ASIC mining

[Consentia]

Hi everyone,

even if the search for online info has been long, I have not found the answers to the questions I have about ASIC mining, so I have to ask you! 

1) Is it correct to say that the cryptographic base used by ASICs is the digital signature?
2) Does this mean that encryption is asymmetric?
3) Is the underlying algorithm SHA256 or ECDSA? I also read double SHA256... I think I've gone a little confused 

Thanks in advance!

C.

[1715478558]

1715478558

[1715478558]

1715478558

[1715478558]

1715478558

[ranochigo]

1) Is it correct to say that the cryptographic base used by ASICs is the digital signature?

Nope. A digital signature generally means that a signature is generated that can be authenticated to a known key. ASICs do not do anything like that; they perform SHA256 hashes twice on the block header and they result in a hash.

2) Does this mean that encryption is asymmetric?

No. The hashing methods in SHA is symmetric.

Sorry, I was wrong about this piece of info. SHA256 is a hashing algorithm and it is neither symmetrical nor asymmetrical.

3) Is the underlying algorithm SHA256 or ECDSA? I also read double SHA256... I think I've gone a little confused  

Thanks in advance!

C.

SHA256D. Its used to hash block header and the hash is considered in the validity of a block (ie. Determine whether it meets the target).

ECDSA is used in the transaction signatures itself. Both is used in the protocol but the mining doesn't consist of ECDSA.

[Consentia]

Thank you for your reply.
So what is the technology and what are the cryptographic methods used by ASICs?

Thanks again!

[HeRetiK]

Thank you for your reply.
So what is the technology and what are the cryptographic methods used by ASICs?

Thanks again!

ASIC stands for application specific integrated circuit -- they are used for all sorts of things, that require simple calculations, fast.

In the case of Bitcoin, they are basically the SHA-256 hashing algorithm implemented in hardware, rather than software. This makes ASICs faster than software being run on top of CPUs or GPUs. The downside is, that ASICs can do only one thing, and one thing only.

Note that cryptographic hashes are different from regular encryption -- they are neither symmetric or asymmetric, they are simply one-way cryptographic functions. That is, unlike symmetric or asymmetric encryption schemes there is no key that can be used to retrieve the original data.

[rickC137]

This is helpful, thank you.

X11 was once thought to be ASIC resistant, until the D3 Antminer.  If someone was going to Code up a new blockchain to be ASIC resistant, to re democratize mining, would one select X11 as the most likely hashing algorithm?

[ranochigo]

If someone was going to Code up a new blockchain to be ASIC resistant, to re democratize mining, would one select X11 as the most likely hashing algorithm?

Same thing happened with scrypt. The main motivation for it is obviously the profit that can be made from developing an ASIC. It isn't cheap to design and build the ASICs. While most coins has strived to be ASIC resistant, many have failed once they take off. Coins like Scrypt utilised the fact that it's memory intensive and it isn't cheap. They won't be ASIC resistant for a long time in that case.

Scrypt-N has a dynamic parameter and that makes them the least suitable for ASIC development.

[Consentia]

In the case of Bitcoin, they are basically the SHA-256 hashing algorithm implemented in hardware, rather than software. This makes ASICs faster than software being run on top of CPUs or GPUs. The downside is, that ASICs can do only one thing, and one thing only.
Note that cryptographic hashes are different from regular encryption -- they are neither symmetric or asymmetric, they are simply one-way cryptographic functions. That is, unlike symmetric or asymmetric encryption schemes there is no key that can be used to retrieve the original data.

Ok, if I have understood correctly the ASICs do not implement the cryptographic algorithms but they directly apply a cryptographic function, is it correct?

And if in any way the ASICs "contain" the SHA256 (regardless if algorithm or function), is this content symmetrical or asymmetrical?

Thanks again,
C.

[ranochigo]

Ok, if I have understood correctly the ASICs do not implement the cryptographic algorithms but they directly apply a cryptographic function, is it correct?

Yes. Bitcoin ASICs are only designed to hash SHA256 twice and nothing else. The data is fed from a computer.

And if in any way the ASICs "contain" the SHA256 (regardless if algorithm or function), is this content symmetrical or asymmetrical?

Sorry, I've made a mistake. It is neither.

Asymmetrical and symmetrical are terms that are used with encryption. In Bitcoin, there is neither the encryption or decryption of any information at the protocol level. SHA256 is used for hashing and not for encrypting; SHA256 cannot be reversed to get the original text.

[Consentia]

Ok, if I have understood correctly the ASICs do not implement the cryptographic algorithms but they directly apply a cryptographic function, is it correct?

Yes. Bitcoin ASICs are only designed to hash SHA256 twice and nothing else. The data is fed from a computer.

And if in any way the ASICs "contain" the SHA256 (regardless if algorithm or function), is this content symmetrical or asymmetrical?

Sorry, I've made a mistake. It is neither.

Asymmetrical and symmetrical are terms that are used with encryption. In Bitcoin, there is neither the encryption or decryption of any information at the protocol level. SHA256 is used for hashing and not for encrypting; SHA256 cannot be reversed to get the original text.

Ok! Now I have better understood and I thank you and the other contributors.

Another doubt: why is SHA256 double?

Thank you again!
C.

[ranochigo]

Another doubt: why is SHA256 double?

In Bitcoin, when SHA256 is used, its usually hashed twice and thats what we meant. So its basically SHA256(SHA256()) or SHA256D.

It's used to primarily prevent the length extension attack[1] for which SHA256 is vulnerable to but not SHA256D.

[1] https://en.wikipedia.org/wiki/Length_extension_attack

[bitChipper]

I have heard over some podcast that the improvement of asics is really slowing down, just like the improvement of cpu processors. I heard it in a debate between Johnny from blockstream and Roger Ver, but basically Johnny said that at our current tech we wont be able to improve very much on the actual design of asics....

Whats your guys thoughts on this? I have always wanted to get a perspective from miners who have good understanding of asics