Let’s talk about ASIC mining


#5644

So is http://www.innosilicon.com/ the one site now selling ASIC miners?

That seems to be very bad for Zcash


#5645

Jeps, Bitmain sold out a while ago I believe.


#5646

Yes, I thought it was bad enough that it was limited to Bitmain and them…but now. Only 1 source for ASICs ouch…

Im sorry ASICs helped ruin Zcash, I was as vocal and against ASICs as I could be when they were annouced. I dont know how or why Zcash flipped so quickly and easily and accepted them into the network. Now with talks of punishing GPU miners(time-locked only them) IF they bring them back? I mean wtf…WHY SUCH A CHANGE?.

It seems like someone was flipped on the inside, but I have no proof of this.


#5647

it takes just single zkon0 conference to flip.

omg! asics are inevitable!

we have to protect miners who invested in asics, because they can’t switch algos, that means they are dedicated!

if we decide to change algo we will warn at least 12 months in advance, because we care for our community!

well… you created coin which strong point was GPU pow, asic-resistance… Miners chose to mine zcash and nothing else (like me). Very short time later you value asic miners rights much more comparing to gpu miners rights. Gpu miners like never made this whole thing.

zooko did not warn any member of dedicated zcash gpu miners. bitmain just announced asics (to the public), meanwhile there simply were no comment from dev team… then zcon0… and then total asic protection!


#5648

ETC 51% attack confirmed. Over 75 blocks:


#5649

ETC is no longer under devolopment last I heard. Sounds like its a dead coin now.


#5650

I feel for you. I am still very sad the company didn’t take stronger and swifter action against asics, and feel some of the “vibe”/positive atmosphere of the company/project was lost as a result, along with other actions & inactions of a similar nature.


#5651

For me personally (and I was staunchly anti asic in the beginning) my perception of the situation changed when I questioned my preconceived notions of everything we were yelling back and forth at each other and whether that aligns with the majority of the Community Values which obviously are not fully represented here
I could only assume that would equate to doing the right thing and could I never correlate that (forking for the sole purpose of blocking miners of a particular type) with the core values of the project


#5652

The problem with ASIC resistance is the unknown. Could an ASIC be developed in secret and then overwhelm the GPU network, leading to a possible 51% attack? Could an ASIC developer turn ‘fair mining’ into unfair mining while the community thinks ASICs are impossible? In my opinion this is a function of the price. At some price it makes sense for well funded groups to attempt these unfair practices. As an aside, governments looking to take down private transactions might see ASICs as a tool in a GPU-mined situation, given how long it takes ZCash to make PoW changes.

If ASIC resistance is actually possible and proven to be true, then we should make it happen. If there is any doubt about that, then it seems like an extremely large risk to use non-hardware-optimized algorithms shared amongst many coins. Overall, I’m sympathetic to the sentiment that @arielgabizon expressed and part of me feels the same way. It’s all about how we weigh the probability of security threats.


#5653

This post was flagged by the community and is temporarily hidden.


#5654

I came to the very same conclusion in princip. The more coins on a given algorithm, the more insecure it gets, just logical. Most exposed the smaller projects on a given algorithm but not limited to them over time.

But the whole security story is more difficult than just this in my opinion. Exactly this topic here made me think about the whole impact of Asics, GPU’s, FPGAs and i have allready several time expressed my conclusion that the whole POW approach is doomed to fail long term. While most here still argue about asics & gpu’s i think it’s a whole POW problem in generally. With Asics we just see things happen faster than with other mining hardware as they are more efficient, but the network security problems are absolutly the same in my opinion.

After i made my worldwide electricity price list and saw that some regions chinese and other regions have access to electricity at just </= 2 cents per kw/h i’am even more convinced that securing a network with POW nowadays isn’t anymore a reliable option, even less when prices fall and new more efficient hardware batches are released.

The current attack on Ethereum Classic / ETC, a top 20 by market cap coin, just backs up my concerns. With an around 8 TH/s network, mixed more or less leveled asic/gpu mining hardware, it obviously was easy to compromise and attack their network. I have said it bevor that for now only the top coins on a given algo are more or less safe from attacks on the network, but this will change as well soon or late, expect BTC for now.

Back to Zcash, i remember that at some time the calculated cost of attack per hour was around $50,000 (not sure if it was 54k or 46k hence i use ~50k), today with a 2+G/s network the attack cost is allready down to $13,000 per hour. There seems to be an easy formula (which isn’t limited to asics) in my opinion: The more coins per algo + more efficient hardware (future) and/or more hardware (future) in generally + centralized mining operations in low electricity cost regions + lower/falling coin prices = higher risk of (future) attacks.

While we are still on the secure side we shouldn’t forget that mining ZEC even at these low prices is still profitable at many regions worldwide. Hence, more mining hardware is still to come, more effiecient ones to be released. With 22 equihash asic mineable coins for sure more hardware is coming, no doubt here, short or long this will further lower the attack cost per hour. I expect it going below $10,000 the next 6 months.

This said, i consider by now the whole POW mining approach no more a problem of asic resistance or not but in generally questionable, wrong by now, insecure and absolutly flawed and leading to mining centralization in the mid and long term.


#5655

I generally agree about your long term mining perspective. Ignoring the economies of scale and electrical pricing unfairness, energy usage is going to be highly scrutinized in the next decade. New issuance via proof of work has already been an easy target and it’s one of the larger complaints against cryptocurrency adoption.
In my opinion we should learn from other proof of stake coins and follow their lead if those coins make it work in a decentralized and secure way. Still, I have some serious reservations about proof of stake.


#5656

#5657

For those reading this article, keep in mind that David Vorick is the creator of Siacoin and owns/runs the associated Obelisk company, which manufactures ASICs. The Sia development team hard forked on October 31st, 2018 to brick every other ASIC - at the time 99% of other miners - except for the one they sell. This article presents him vaguely as an Ethereum developer.


#5658

In conversation with CoinDesk, Spraul also said that pending its implementation into ethereum, the company will research the feasibility of building specialized ASIC hardware for ProgPoW.

“I can publicly confirm today that we intend to study the feasibility, and then build, ProgPoW ASICs,” Spraul said.

Because ProgPoW changes ethereum’s underlying mining algorithm, Ethash, to be more memory-heavy, the code switch is said to make GPU hardware competitive with ASICs.

Proponents of ProgPoW say that if hardware designers try to build ProgPoW ASICs – which is to say a specialized chip with the sole function of computing ProgPoW – it would just end up resembling GPU hardware.

Still, Spraul denied this, stating that “Hardware innovation is non-linear,” and “We can accelerate ProgPoW by a factor of 3x to 8x.”

For comparison, Equihash’s current ASICs are ~11x more efficient (at lower cost).


#5659

I think it’s a threat so they don’t add ProgPOW. I’ll believe it when I see it. The hardware would be more expensive and the difference would be minimal IMHO.


#5660

Maybe? It’s hard to know. It would be a bold bluff to call given how much capital these miner manufacturers have.


#5661

This is not entirely true. After using ASIC mining began an exponential growth of network hashrate , and as a result increase of mined coins. For example with the ASIC’s arrival from June to December 2018, the hashrate increased 5 times in 6 months. When using GPU mining in 6 months hashrate increased 2 times. Also, ASIC contributes to the fact that few people can hold a lot of coins and speculate in price.


Zcash is the worse cryptocurrency from top 100 capitalization
#5662

No, the difficulty adjusts to keep an approximate constant blocktime (currently 150 seconds).


#5663

The network will generate one block each 150 seconds regardless if there’s a single Nintendo 64 mining or millions of ASICs