I thought Segwit eliminated AsicBoost

[Wind_FURY]

https://twitter.com/slush_pool/status/977499667985518592

Our block #514882 is the first-ever block mined using the active version-rolling aka overt AsicBoost. According to the available evidence, it was mined by @HalongMining hardware. Congrats to lucky miner from Guangdong, China 🇨🇳

Plus isn't AsicBoost patented technology which causes an unfair environment on Asic producers and miners in an already not so fair mining environment?

Maybe Cobra was not that crazy when he proposed a change in mining algorithm. Another Bitcoiner proposing it before him is Luke Dashjr. Although they were not taken seriously, but I believe soon they will be.

[ranochigo]

Segwit didn't eliminate ASICBoost completely. There are two kinds of ASICBoost; covert(can't be detected) or overt(can be detected easily).

SHA256 splits the 80byte block header into 2 chunks (of 512bits each) before hashing them. The merkle root overlaps from the 1st chunk to the 2nd. With ASICBoost, the last 4 bytes of the merkle root in the 2nd chunk is to be kept the same while the 1st chunk is varied. This allows for an improved efficiency since they don't have to calculate the second chunk over and over again. The merkle root can be changed by shuffling the order of transactions.

However, with Segwit, there is a witness commitment inside the coinbase transaction. With this, the covert attack with transaction reordering can't work since the witness commitment has to match the ordering of the transactions in the block.

Overt ASICBoost will still work since the versionBit is in the 1st chunk and the 2nd chunk wouldn't need to be recalculated. Everyone can see the change in the VersionBit so it's easy to tell.

[Xynerise]

Overt ASICBoost will still work since the versionBit is in the 1st chunk and the 2nd chunk wouldn't need to be recalculated. Everyone can see the change in the VersionBit so it's easy to tell.

The change in the nVersionBits actually triggers warnings about soft fork signalling for full node operators.
And it also reduces the amount of possible soft fork deployments available to bitcoin from 29 to 13

Stranger still, btcdrak, owner of HalongMining whose company produces miners that use overt ASIC boost created a pull request to make it standardised in Bitcoin Core.

[TheQuin]

Plus isn't AsicBoost patented technology which causes an unfair environment on Asic producers and miners in an already not so fair mining environment?

There's a lot more information over in the DragonMint 16TH/S halongmining.com thread. It's pretty long but there's a lot of good info buried in there mainly from -ck who has been writing the drivers for them.

Bitmain had it hidden from the public and had the patent for it meaning no one else could use the technology giving them an unfair advantage. The halong group paid dearly for the patent to buy it and then opened it up with the blockchain defensive patent license - this license means anyone can use the asicboost technology provided they don't patent anything else in their hardware. Think about it - they paid for the license, only to open it up for everyone. No one has an unfair technology advantage now.

[nullius]

As others have already explained, Segwit prevents the covert form of ASICBOOST (and I have always been careful to make that distinction when discussing ASICBOOST on this forum).

ASICBOOST is not evil per se.  If everybody has it and uses it openly, then it is simply an optimization.  But if one party has a patent on it and stealthily uses the covert form (prevented by Segwit), then that party has an unfair advantage—leading to worse miner centralization.  That constitutes an attack on the Bitcoin network.

I like the sound of an open, defensively-licensed overt ASICBOOST being sold by a major new competitor to Bitmain.  This could solve some serious problems with centralization not only of mining, but of sources of mining hardware.  I hope it will ultimately lead to greater hardware availability, which is necessary to decentralize the global hashrate.  Altogether, it is also an excellent reason to stick with SHA-256 for Bitcoin’s PoW algorithm!

Times are about to get more interesting...


Thanks @BtcDrak for doing this.  If you want to understand who is behind this new Halong Mining, peruse the bitcoin-dev archives and the Core commit history a bit.

[andrew1carlssin]

As others have already explained, Segwit prevents the covert form of ASICBOOST (and I have always been careful to make that distinction when discussing ASICBOOST on this forum).

ASICBOOST is not evil per se.  If everybody has it and uses it openly, then it is simply an optimization.  But if one party has a patent on it and stealthily uses the covert form (prevented by Segwit), then that party has an unfair advantage—leading to worse miner centralization.  That constitutes an attack on the Bitcoin network.

I like the sound of an open, defensively-licensed overt ASICBOOST being sold by a major new competitor to Bitmain.  This could solve some serious problems with centralization not only of mining, but of sources of mining hardware.  I hope it will ultimately lead to greater hardware availability, which is necessary to decentralize the global hashrate.  Altogether, it is also an excellent reason to stick with SHA-256 for Bitcoin’s PoW algorithm!

Times are about to get more interesting...


Thanks @BtcDrak for doing this.  If you want to understand who is behind this new Halong Mining, peruse the bitcoin-dev archives and the Core commit history a bit.

When I decided to buy an ASIC miner (yes I was one of the pissed Butterfly's costumer back in 2012 ) instead just buy bitcoin .. there was an conversation about how Application-specific integrated circuit could help prevent malicious bot-nets etc ... Today one thing that annoys me a lot (mainly in academic field) is the criticism about waste of electricity. I used to say that this is the collateral of a more security system.  

I am not sure that Intel's Proof of Elapsed Time (PoET) will solve the problem since Intel SGX is patented and we going to end up in ASICBOOST polemic of unfair competition. But I just wondering ... how complicate is to build ASIC to mine PoW such as primecoin ?



  

[Wind_FURY]

As others have already explained, Segwit prevents the covert form of ASICBOOST (and I have always been careful to make that distinction when discussing ASICBOOST on this forum).

ASICBOOST is not evil per se.  If everybody has it and uses it openly, then it is simply an optimization.  But if one party has a patent on it and stealthily uses the covert form (prevented by Segwit), then that party has an unfair advantage—leading to worse miner centralization.  That constitutes an attack on the Bitcoin network.

I like the sound of an open, defensively-licensed overt ASICBOOST being sold by a major new competitor to Bitmain.  This could solve some serious problems with centralization not only of mining, but of sources of mining hardware.  I hope it will ultimately lead to greater hardware availability, which is necessary to decentralize the global hashrate.  Altogether, it is also an excellent reason to stick with SHA-256 for Bitcoin’s PoW algorithm!

Times are about to get more interesting...


Thanks @BtcDrak for doing this.  If you want to understand who is behind this new Halong Mining, peruse the bitcoin-dev archives and the Core commit history a bit.

According to this it is under the "Blockchain Defensive Patent License".

https://www.asicboost.com/patent/

I could see the new ads for them already. "Introducing the new ASIC BOOST! Your Bitcoin mine will never be the same again!"