|
Bitmain expected to release Ethereum ASICs in April:quote:Bitmain, which has already been supplying mining ASICs for Bitcoin in the past several years and is one of the top-5 clients of TSMC, is ready to release ASIC products for mining the cryptocurrency Ether in April at the earliest. The products are expected to shore up the performances of related upstream suppliers including TSMC, Advanced Semiconductor Engineering (ASE) and Global Unichip (GUC).
|
#
¿
Mar 8, 2018 19:41
|
|
|
|
| # ¿ May 21, 2024 13:29 |
|
|
22 Eargesplitten posted:So this is basically going to crash the GPU market hardcore, right? Or will people just switch to some other shitcoin? I actually just did the math on this and came up with the following guesstimates. Top 5 coins by market cap:
Obviously there is a tier of shitcoins below that but at a rough guesstimate, Ethereum makes up about 3/4 of the entire GPU-mining economy. So yeah, if ASICs push GPUs out of the Ethereum pool then there's enough GPUs there to quadruple the difficulty on every single other coin, ish. In reality it's a little more complicated than that, because AMD GPUs are comically bad at mining everything except Ethereum (Polaris) and Cryptonight (Vega), so 10 million AMD GPUs is only equal to 5 million NVIDIA GPUs on most of the other coins, but that's still doubling the difficulty on every other coin. Also, as difficulty rose, a certain amount of people would cash out their rigs and so on, but it would not be a good time for profits.  My assertion is that ASICs should be no more than 5-10x more powerful than a general-purpose GPU on ASIC-resistant coins. On the other hand, if you assume that difficulty goes up by 5x then gross income goes down by 5x too, so a 570 (one of the most efficient cards for Eth) would go from making $1.53 per day before electricity to making $0.306 per day before electricity, and at $0.16 per kWH that would be $0.46 per day... so they'd be underwater. And while mining with Polaris or Vega on altcoins would be still profitable at current difficulty, they're so inefficient that they'd probably be underwater if difficulty doubled. Early rumors had the Bitmain ASICs at 2.5-4x the efficiency of Polaris at Ethash. However, it could also be significantly cheaper to produce them, since you can use many small dies instead of one big die. So it's hard to quantify what the impact will be on difficulty, apart from "up quite a bit". It all depends on how many they produce and how quickly they hit the market, and how quickly GPU miners go offline. In combination with NVIDIA potentially stockpiling for a new series of cards, and perhaps even a mining-optimized GPU series coming later this year... I think the smart money was on selling your cards a month ago and said as much at the time. It's probably not going to instantly crash when Bitmain launches, but it'll be downhill from there. (now watch as Bitcoin prices quadruple and miners snap up every card for the next 6 months again). Paul MaudDib fucked around with this message at 23:11 on Mar 8, 2018 |
|
#
¿
Mar 8, 2018 22:46
|
|
|
Risky Bisquick posted:https://bitcointalk.org/index.php?topic=2932282.0 Point is, it is a second source, a little more reputable, and hopefully they did a little bit of independent verification (one can hope). Yeah, that's the thread where I sourced the efficiency numbers. 650 MH/s at 750W is 3.7x the efficiency of a 570, or about 3.5x the efficiency of a 1070. Which is honestly just about where I'd expect a genuine Ethash ASIC to be. The memory subsystem (chips+controller) of a GPU makes up about 10-20% of total power consumption, if you zeroed the power of the rest of the card you'd expect a 5-10x efficiency improvement. Of course you can't actually have a processor consume zero power, but ASICs can easily be hundreds/thousands of times more efficient, so you can get close enough that it's a rounding error. The central idea of memory-hard ("ASIC-resistant") coins is that you can't really make the huge gains on memory performance like you can on processor efficiency. DDR3 is supposed to be a bit more efficient than GDDR5 on the small transfers that Ethash does, but it won't be 100x more efficient, I'd be surprised if it was more than 2x the efficiency, and that's a one-time gain. Maybe a little more if you went to DDR4 or DDR5, but not hundreds/thousands of times faster. So, potentially still some gains for a couple more generations of ASICs to squeeze out, but it sounds about right for a first-gen ASIC on an ASIC-resistant algorithm. The bigger question is how Vitalik would respond... tweak the algorithm a bit to break the ASICs? It'd be a pain for miners, but it'd kill this generation of ASICs, at least. Speed up epochs so they hit the 4 GB barrier? That would break the current generation of ASICs but it would also screw anybody with 290s, 570/580 4 GBs, 1060 3 GBs, etc, and there's nothing stopping Bitmain from releasing an 8 GB version. Some altcoins have proposed switching algos regularly, but it would have to be done on short notice to keep Bitmain from implementing them ahead of time, which would become a logistical nightmare (and effectively a form of centralization). And Bitmain can always just mine in-house and not tell anyone, so it would have to be done regularly and often, not ad-hoc. If true, it's put-up-or-shut-up time for Vitalik on Proof of Stake. Proof of Work is inherently vulnerable to ASICs once it reaches a certain size and it becomes profitable to make ASICs. He's years behind schedule and has no progress to show for it. Paul MaudDib fucked around with this message at 00:00 on Mar 9, 2018 |
|
#
¿
Mar 8, 2018 23:29
|
|
|
Baikal has a cryptonight ASIC out. Efficiency works out to 9.73x as efficient as Vega 56.
|
|
#
¿
Mar 12, 2018 06:14
|
|
|
Kazinsal posted:Isn't there also an Ethereum ASIC on the way too? Between Ethereum and Cryptonight that's like 90-ish percent of the GPUs in the global hash pool. Fake news, everyone knows that it's impossible to make a chip that processes Ethereum or Cryptonight, which is why all cryptomining is done with a pen and paper, idiot  (see also: "surely they will be able to invent enough random shitcoins to absorb 90% of the GPUs on the market, God is dead and Santa isn't real and in the grimdark future we will all be gaming on consoles")
|
|
#
¿
Mar 12, 2018 06:29
|
|
|
Cryptonight ASICs are real, 40x GPU efficiency
|
|
#
¿
Mar 15, 2018 15:20
|
|
|
22 Eargesplitten posted:
Monero has said they'll hardfork (edit: in 12 days) to get them off the network, hence the huge "NO REFUNDS" on Bitmain's order page. I think it's logistically impossible to do that regularly, though.
|
|
#
¿
Mar 15, 2018 16:26
|
|
|
If it's anything like previous ASICs it probably also spews all kinds of RF noise and the FCC may track you down and politely ask you to turn it off
|
|
#
¿
Mar 15, 2018 17:24
|
|
|
The serious players are using ASICs, the bag-holders are still HODL'ing. lol if you actually believed that whole "GPUs will always be around" bullshit that some of the self-interested miners were spouting, it's been obvious for at least a month that the jig's up, have fun flipping on ebay this month or I guess maybe the ASICs are all a lie, what even is a Bitmain, who can say Paul MaudDib fucked around with this message at 16:11 on Mar 18, 2018 |
|
#
¿
Mar 18, 2018 16:03
|
|
|
Stealthgerbil posted:Gridcoin is legit but no one mines it because its not profitable. The whitepaper is discussing a hypothetical possibility (hence "untested"), the current Ethash algorithm does not make contract execution part of the proof-of-work, it operates on pseudorandom data. There's really no reason for that paragraph to be in there at all, it just sounds cool. quote:
Paul MaudDib fucked around with this message at 17:39 on Mar 19, 2018 |
|
#
¿
Mar 19, 2018 17:28
|
|
|
Torvalds is kind of an idiot. NVIDIA actually does have a solid performance argument for why their approach is superior, even if it's not the hardware-abstract approach Torvalds wanted. He was way off on the latest AMD vulnerability too. Root password is absolutely not authentication to bypass signature authentication on PSP/IME firmware. That was a Poettering-level hot-take from Torvalds - yet another insufferable "why would that be a problem" from a person who's used to having complete+total privileges on a system. The tech community cuts him an insane amount of slack for his outbursts, there are lots of big projects that get done without a scathing rant from the founder every 3-6 months. Probably only going to get worse as he ages. Paul MaudDib fucked around with this message at 20:33 on Mar 24, 2018 |
|
#
¿
Mar 24, 2018 09:02
|
|
|
snip
Paul MaudDib fucked around with this message at 20:34 on Mar 24, 2018 |
|
#
¿
Mar 24, 2018 10:29
|
|
|
snip
Paul MaudDib fucked around with this message at 20:33 on Mar 24, 2018 |
|
#
¿
Mar 24, 2018 10:45
|
|
|
snip
Paul MaudDib fucked around with this message at 20:33 on Mar 24, 2018 |
|
#
¿
Mar 24, 2018 10:49
|
|
|
CNBC confirms Bitmain Eth ASICs, says there's at least three others in the worksquote:"During our travels through Asia last week, we confirmed that Bitmain has already developed an ASIC [application-specific integrated circuit] for mining Ethereum, and is readying the supply chain for shipments in 2Q18," analyst Christopher Rolland wrote in a note to clients Monday. "While Bitmain is likely to be the largest ASIC vendor (currently 70-80% of Bitcoin mining ASICs) and the first to market with this product, we have learned of at least three other companies working on Ethereum ASICs, all at various stages of development."
|
|
#
¿
Mar 26, 2018 19:36
|
|
|
Dr. Fishopolis posted:I mean, it still requires a legit cert and local admin to accomplish. I wouldn't call that an "egregious flaw", if you're at the point where you could use the exploit you've already owned it so thoroughly it hardly matters if there's some cpu-specific vulnerability or not. Stuxnet this ain't. One of the vulnerabilities was a bypass for the signature validation, so you didn't need a signed binary, just local admin, and then you could flash whatever you wanted down to the PSP.
|
|
#
¿
Mar 27, 2018 01:48
|
|
|
lmao all the Ethereum miners on Reddit are super in denial about ASICs "it's probably just a fancy GPU rig!"
|
|
#
¿
Mar 28, 2018 23:00
|
|
|
Stickman posted:I love how every goddam libertarian loves freedom until they suddenly stop winning. Then it just isn't fun anymore and they need their special protections. Bioshock is extremely on-point about this aspect of libertarians
|
|
#
¿
Mar 29, 2018 10:32
|
|
|
put your funds into my dip when I dip you dip we dip
|
|
#
¿
Apr 1, 2018 16:54
|
|
|
Running at tighter VRAM timings won't physically affect the card. The voltage is the same and everything. As long as they flash it back, it's no different to any other card. It could be a problem if they didn't flash it back, and games aren't stable at the tighter timings, but in principle Ethereum is a pretty heavy workout and if it's stable there it should be stable anywhere. I don't see much of a problem with miner cards beyond the relatively high number of power-on hours at unknown core voltage and fan settings/temps. Like, what miner isn't going to say that they ran it nicely for just two months before they decided to sell? And frankly that's assuming they tell you at all. I mean, unless they list 10 of them at once, how do you know they're a miner? Hell, how do you know it's not a third-hand card that was mined on before they got it? At this point, unless you have a card with a known provenance (a friend who you know doesn't mine/etc) then you pretty much have to assume every card newer than GCN 1.0 could have been mined on. The cards that are least likely to have been mined on are Fermi, Kepler, and Maxwell, but really you just don't know. Treat every used card like it was potentially mined on. (Truth is, unless you can buy from a friend you'll just have to take what you can get, this is already factored into the prices of used cards, and it's truthfully not that big a deal. GCN 1.0 and 1.1 cards pulled tons of power vs today's cards and had lovely blowers that ran crazy hot, and there's a lot of cards that mined back in 2014 that are still doing fine for gamers today.) Paul MaudDib fucked around with this message at 23:49 on Apr 2, 2018 |
|
#
¿
Apr 2, 2018 23:40
|
|
|
Bitmain announced their ASIC, the E3 miner, producing 180 MH/s at 800W The efficiency is actually terrible, effectively the same as a GPU in terms of hash/w (actually very slightly worse). They do note that these are "conservative" numbers but still, that's a little disappointing after all that build-up. Maybe some of the other ASIC companies have had better luck.
|
|
#
¿
Apr 3, 2018 16:48
|
|
|
There's no way it's a coincidence that Bitmain is advertising exactly the same efficiency as a GPU, they have to be throttling them. Bitmain is trying to get a foot in the door here, so Ethereum doesn't immediately hardfork like Monero did. Then next generation they will start opening up the throttle (knowing Bitmain they are probably already pre-mining on a newer generation). Gotta boil those frogs slowly. That or it's actually a bunch of GPUs on MXM cards or something, but I don't see any reason to doubt that an ASIC company is releasing an ASIC. The whole "oh well maybe 180 MH/s or maybe faster, gee I dunno guys" schtick only adds to my certainty here. Don't bullshit me, Bitmain premines like crazy, they know exactly how fast these things are. Paul MaudDib fucked around with this message at 18:14 on Apr 3, 2018 |
|
#
¿
Apr 3, 2018 17:54
|
|
|
Pretty sure there are private ASICs for most of the major coins at this point, hence why profits have been driven so low. People on Reddit (and in the crypto community in general) are in deep denial about it though.
|
|
#
¿
Apr 24, 2018 21:09
|
|
|
In "poo poo everyone who paid attention already knew" news, researchers say Tether was artificially inflating the price of Bitcoin last year.
|
|
#
¿
Jun 14, 2018 02:55
|
|
|
Fauxtool posted:I really dont think the big use GPU sell off people are wishing for is going to happen. Most current gen GPUs that are paid off are still profitable to run and will be for awhile Used 1080 Tis are already under $550, 1080s are pushing $350.
|
|
#
¿
Jul 27, 2018 04:34
|
|
|
Waste of electricity, at this point.
|
|
#
¿
Aug 24, 2018 22:18
|
|
|
|
| # ¿ May 21, 2024 13:29 |
|
|
BangersInMyKnickers posted:We're hitting a hard wall with per-core process improvements and in the absence of that you run custom silicon to accelerate the bottlenecks in your workload. This is an inevitable consequence of the world hitting the limits of silicon die-shrinks and GF is following the money. It's a little more complicated than that. GF could choose to chase 7nm/10nm like Intel, Samsung, and TSMC but the huge capital costs to develop the process need to be amortized over a large number of wafers, and GF didn't want to pay to expand their facilities to the extent that would be required for that. Custom silicon is well and good but there were process improvements that GF could have made that they chose not to. The article says that their investors are getting antsy with GF continuing to lose money and wanted to go for profitability today rather than another big capital outlay trying to stay competitive with the big dogs. Of course, there are many tiny fabs out there that did the same thing, you just don't know them because nobody cares about them anymore, and if GF goes down that road then eventually they'll fade out too. Charlie Demerjian was all-in on the idea that AMD was going to multi-source Zen2, GF made changes to their anticipated 7nm process that put their design rules more in-line with TSMC 7SOC, and then all of a sudden AMD announced they were single-sourcing from TSMC and GF announced they weren't doing 7nm. It's not clear which came first, because GF could definitely get cold feet if their lead customer pulled out, but AMD could also have pulled out because GF wasn't serious about building out enough capacity to get costs down sufficiently.
|
|
#
¿
Aug 28, 2018 13:53
|
|