Did some more reading on the crypto forums and it seems that each time that an ASIC miner is developed for mining a specific currency, the rabble get the coin developers to do a hard fork in the algorithm totally negating the ASIC miner. It seems that the general populace that is mining doesn't want the currency to become concentrated in only the few entities that can buy and deploy these very costly ASIC miners. There is already talk about hard forking the ETHash coins that this new Bitmain E3 ASIC miner is supposed to mine. This happened for a different coin and ASIC miner where the miner is now worthless. The company is sitting on millions of dollars of useless hardware. They have buy 1 - get five sale going on basically because they can't even give them away.

It also appears that this new Bitmain E3 ASIC miner that still hasn't opened for sale to the public had actually been put to use already by Bitmain itself to mine Ethereum. They posted a $3.5B profit last year. They are as profitable as Nvidia. If the crypto currencies actually do the hard forks to negate ASIC mining, it will put pressure back on the gpu market when the miners start grabbing every gpu available again. I hope this does not happen.

[Edit] Looks like Bitmain just announced ANOTHER ASIC miner, the Antminer Z9 mini which is used to mine Equihash coins like zcash. I also discovered that a very large proportion of production capacity at TSMC has been for producing chips for Bitmain. This could also be another reason why gpu cards have been hard to buy for the last year if 30% of the wafers have been used for ASICS and whatever else is left finally gets used to produce the Nvidia and AMD gpu silicon.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

What I am wondering about after reading your post Keith, is why would any mining hardware company announce to the world an ASIC miner that can efficiently process coins, because I'd think if you had developed something that was head and shoulders above the current GPU mining rigs capabilities, and by advertising it you PO those miners enough to have them cause a fork, you'd be shooting yourself in the foot.

If I was in this, I'd do my best to keep it as much on the downlow as possible, fly under the radar for as long as you can, and have my company crank out tons of coins for as long as possible, and when it doesn't work any longer, walk away. I know there's the miner vs. shopkeeper analogy from the 1800s gold rush (many/most ppl selling the stuff to the miners did quite well, as opposed to how most miners did), but if it's obvious that your business model will crash as soon as your product is released due to the inevitable backlash against it, I'd think the risk/reward ratio is pretty slanted. As evidenced by your example above. But then, what do I know, I don't mine, or make mining hardware... ;-)

---

I dunno, the whole setup sounds like racketeering to me. No wonder that governments want to get involved to monitor it.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

I agree, it seems to be a little on the sketchier side, that's why I never got into it, even though there had been decent money to be made, especially for the early adopters.

---

It's just another form of gambling.

---

Grant
Darwin NT

Somewhere between a racket, a ponzi scheme and gambling.

Which ever - late adopters run away or get burned

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

What I am wondering about after reading your post Keith, is why would any mining hardware company announce to the world an ASIC miner that can efficiently process coins, because I'd think if you had developed something that was head and shoulders above the current GPU mining rigs capabilities, and by advertising it you PO those miners enough to have them cause a fork, you'd be shooting yourself in the foot.

If I was in this, I'd do my best to keep it as much on the downlow as possible, fly under the radar for as long as you can, and have my company crank out tons of coins for as long as possible, and when it doesn't work any longer, walk away. I know there's the miner vs. shopkeeper analogy from the 1800s gold rush (many/most ppl selling the stuff to the miners did quite well, as opposed to how most miners did), but if it's obvious that your business model will crash as soon as your product is released due to the inevitable backlash against it, I'd think the risk/reward ratio is pretty slanted. As evidenced by your example above. But then, what do I know, I don't mine, or make mining hardware... ;-)

Well it seems that Bitmain actually did keep the new hardware as low key as possible . . . . for a while. I don't know their business model, but it IS NOT CHEAP to develop an IC or ASIC design, produce the lithography masks and contract with TSMC to lock up a certain percentage of production capacity for the contract period or quantity.

I wonder if the exclusive run with the new miner by Bitmain itself was to payoff the design investment and TSMC up front costs and only when they have balanced the books, sell the hardware to the public which would expose the effect on the algorithm much faster than what Bitmain did itself. Only then might it incur the wrath of the coin developers to consider a hard fork against the new mining hardware and possible negation of the new miner.

I see some between the lines comments that there is a faction within each coin that considers these new hyper efficient miners "a good thing" And I do believe that sentiment is held by the new investment communities and hedge funds. Will be interesting to see how much influence those interests have on the algorithm steering committees of each coin. I have seen comments by some coin inventors that they would allow ASIC mining "over their dead body" and that would be possible because the steering committee numbers 1, themselves.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Coinbase plans revamp to lure institutional and high-speed traders

NEW YORK (Reuters) - Cryptocurrency exchange Coinbase is planning a revamp of its trading technology and creating a new suite of services to attract more institutional investors, including hedge funds and high-frequency trading firms, it said on Tuesday.

The San Francisco-based startup said it is building a new matching engine that will allow firms to match buy and sell orders for cryptocurrencies like bitcoin in fractions of seconds and is building up a team that will cater exclusively to institutional clients.

New services, such as the ability for trading firms to colocate their servers as close to Coinbase’s operations as possible for faster trading, will be rolled out within the year to help attract more trading participants, Coinbase said. Critics of colocation, a feature of many electronic markets, say it is unfair because it allows firms that can afford it to pay for faster access.

But Coinbase said the move will lead to better results for all traders on its exchange, including retail investors, who currently trade the bulk of cryptocurrency volumes.

---

According to economist Alex de Vries electricity consumption by miners is of the order of 66 TWh/year, similar to energy needs of countries like Austria and Finland.
Tullio

---

So, it's been a while since the last time we've discussed this, and I was, for the first time in a couple months, over at EVGA's site this morning just for giggles, checking out what is in stock currently. Man, what a difference 6 months or so makes, huh? Pretty much everything is available, but my memory is a little foggy, it doesn't appear that the pricing has really yet taken much of a hit?

I saw mentioned in another thread that apparently Nvidia cranked up their production of the current gen to cover the mining boom just about the time that it turned into a bust (figures), and that is going to be pushing back the release of their newest GPU till possibly sometime next year? Even if this is true, I'm surprised that it hasn't seemed to have much effect on the pricing of the current gen cards. Might it be because there is still an issue with memory pricing, even though I think I read that the major mfg's plead pretty much no contest to the price fixing business?

Anyone have a good grasp on what the current situation in the GPU market is mid-2018?

---

From what I have read, the current card prices are back to MSRP. But no good fire sale prices because the OEM's don't feel the need to unload inventory yet. And that is because of no new products in the pipeline forcing the usual condition of needing to get rid of old generation cards from inventory just as you are releasing new generation cards where all new sales will come from.

That and the possibility of GDDR5/X memory still being in short supply and I think that is because fab lines are being turned over to produce the new GDDR6 memory.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

There have been forecasts of price drops of up to 20% by the end of this month on some GPUs from the vendors, so depending on stock levels it could take 1-6 weeks for those drops to flow through to retail.

However, there has been some major flooding in the provinces in China that are the main home to crypto miners.
What isn't know is just which miners are affected (Bitcoin which is mostly ASIC, or Ether & other GPU mined currencies), and whether there has been damage to their hardware, or just infrastructure. Also many smaller miners are getting out due to the decreasing return, and larger miners may or may not spend the money to rebuild their systems in light of the diminishing returns.

So what it boils down to is that GPU prices may, or may not, drop considerably over the next month or 2.

---

Grant
Darwin NT

AMD and nVidia are no dummies. They are not going to leave any more blood on the table than they have to.
Even though they have allowed themselves to be 'manipulated' to a point by the crypto miners, they grabbed everything they could while the boom lasted, and now they will probably be content to work off their fat production pipeline rather than rush their new product to market.
Either way, the average Joe GPU user suffered either higher prices or delayed new product.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

Yes, you are correct. The graphics chip makers are no dummies. But I saw a report that one OEM was returning 300,000 Pascal chips to Nvidia because they didn't have a market for finished product anymore. That is going to take quite a bit of time to work off that much inventory or finally decide to take a loss for scrapping what they finally end up with.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Jezuz, 300k chips? Any guesses as to how much production capacity/time that quantity of chips is? I have no idea how fast they would produce that quantity of chips, days/weeks maybe? Or is this a blip on their production schedule? From what you are saying, that sounds like it might be somewhat significant.

---

Jezuz, 300k chips? Any guesses as to how much production capacity/time that quantity of chips is? I have no idea how fast they would produce that quantity of chips, days/weeks maybe? Or is this a blip on their production schedule? From what you are saying, that sounds like it might be somewhat significant.

That is just a blip. TSMC produces 12M 300mm wafers a year. That is 32K per day. A 300mm wafer will produce 150 400mm square dies which is a good approximate for the latest gpu dies with infeasible 100% yield.

So 300K dies need about 2000 wafers or about 1 1/2 hours worth of work.

One Bitcoin miner is buying 20,000 16nm wafers from TSMC per month

TSMC is moving to 450mm diameter wafers for 7nm production. Apple's A12 cpu is currently in production using 7nm lithography.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Yes, you are correct. The graphics chip makers are no dummies. But I saw a report that one OEM was returning 300,000 Pascal chips to Nvidia because they didn't have a market for finished product anymore.

Rather strange considering all the GTX 1030, GTX 1050, GTX 1060, GTX 1070 & GTX 1080s & GTX 1080Tis still being sold, even with the present still high prices.
Maybe once the next generation of GPUs is finally released, and widely available I can see the market for Pascal chips drying up, eventually. Even so, the GTX 750Ti Maxwell GPUs are still selling.
If that OEM was making mining only cards with no video output they might now no longer have a market for them, although I don't see why NVidia would take them back unless they had yet to be paid for.

---

Grant
Darwin NT

I wondered about that too. Why would Nvidia agree to take them back? I was thinking some sort of commission arrangement or advance shipment conducive on a contract arrangement to buy many more wafers in the future. Don't know how the business arrangements work between foundries and OEM's

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

The prices at NewEgg aren't that much higher than they were before Christmas...if you keep an eye out.
A couple of weeks ago the GIGABYTE GV-N1060WF2OC-3GD was down to $229 with a $20 rebate, making it only $10 more than pre-Chrismas. So, I bought another one.
Then, the EVGA 03G-P4-6167-KR keeps dropping down to $239 for a day or so. I bought one of those at $239.
Right now they are back to slightly higher prices, but, they are Much better than they were a few months ago.

Today is a good day to pick up some bundles on GPUs and PSUs at EVGA

https://www.evga.com/products/featured-bundles.aspx

---