Bitmain E3 Ethash Miner ASIC (Batch 1 - Shipping:16-31 July. $800 USD)

[siampumpkin]

It seems like these new E3 can be upgraded once ETH will fork and still mine it. I heard this from "insiders"..

Can you provide more information? Is it using and FPGA versus ASIC? Is it really GPU cores inside?

How exactly could they upgrade it after the fact?

If you look at their main site they have AI research chip, could be that the chips in E3 are modified version of these chips. Also they promise better speed when the units start to ship, firmware upgrade???

This is just typical Bitmain smoke and mirrors. They cannot be trusted for anything. By the time they ship I hope all the ETH coins have forked to a totally new algo and make this trash worthless!!!

[1713587009]

1713587009

[1713587009]

1713587009

[1713587009]

1713587009

[MBW_2]

At this price that's the dead of GPU miners. Now cost only 1/3 and have same hashrate.
Lucky I just sold all of my GPU rigs after 3 months used still profitable. lol

You have time util August when bitmain will deliver first batch...

[Searing]

It seems like these new E3 can be upgraded once ETH will fork and still mine it. I heard this from "insiders"..

Can you provide more information? Is it using and FPGA versus ASIC? Is it really GPU cores inside?

How exactly could they upgrade it after the fact?

If you look at their main site they have AI research chip, could be that the chips in E3 are modified version of these chips. Also they promise better speed when the units start to ship, firmware upgrade???

This is just typical Bitmain smoke and mirrors. They cannot be trusted for anything. By the time they ship I hope all the ETH coins have forked to a totally new algo and make this trash worthless!!!

won't matter to Bitmain these are PRE-ORDERS for what end of July 2018 was it? So you see, they have your coin...it forks and ETH goes POS coin and Bitmain still wins!

(Bitmain: professional evil....professionally done...we do evil well!) (tm: Bitmain)

[QuintLeo]

I think the E3 uses 16 x BM1680 Neuralprcessing Chips and one Asic or FPGA for the controlling.

If the E3 has 6 cards, I must correct myself:
6 cards with 2 x BM1680 + 1 FPGA or ASIC each
That´s 12 x BM1680 + 6 x FPGA or ASIC.

Or 18 x BM1680 and one Asic or FPGA for the controlling, if Bitmain connect the BM1680 with each other.

Even Bitmain sell a card with 1 x BM1680 and FPGA since January. They announced a card with 2 x BM1680 and FPGA.  

Why are people even talking about the Bitmain AI chip and this thing?

That chip is optimized for FLOATING POINT operations, like any AI chip - it is NOT going to process INTEGER operations very well like mining requires, NOR is it going to be efficient at doing so.

The BM 1680 is NOT A MINING CHIP and is not going to offer decent performance per $ or watt since the bulk of the chip would go to waste AND WASTE SOME POWER if someone tried to mine on it.

[Metroid]

(Bitmain: professional evil....professionally done...we do evil well!) (tm: Bitmain)

Bitmain looks like an American company where only money really matters, hard to say it's Chinese. I wonder where is the socialism and equality. That jihan guy maybe spent most of his life with evil american manufactures. Learnt from them.

[QuintLeo]

This must be bitmain's smokescreen, for Vitalik Buterin, to fool him into thinking they don't have the F3 1.5 g/h miners,  since bitmain announced they were making these e3's last year.

Bitmain didn't announce the E3 last year, they announced 2 GPU-based rigs (G1 and G2 as I recall) last year.

The alleged F3 is still a "one video as the only source of information" rumor.
That video is VERY uninformative about any hard facts, VERY short, and it's entirely likely it wasn't about an ethash miner anyway even if the miner pictured DOES actually exist.

Videos are EASY to manipulate - we've seen "fake" ones too often in this industry before after all.

[jstefanop]

With current difficulty, this Bitmain Garbage can only mine for a round $4 a day. Imagine all the batches are come a live and then difficulty sky rocketed way higher.
Both this ASIC miner and GPU have to say good bye to ETH.
Btw, are you sure that this is the best Bitmain can do? I somehow have intuition that they have more powerful hashing board to mine ETH under their wrap still. Way more powerful that this garbage.

This garbage has officially killed ETH gpu mining for now. Until ETh dev comes up with something or AMD or Nvidia develops a new gpu specifically designed for mining.
This garbage is still very low cost and very scalable. Good for mining farms. So it will increase difficulty on all the coins as big gpu farms will switch coins.

This wont kill eth mining anytime soon. These are not A3's/D3's etc. Anyone worried about these has no sense of the magnitude of Ethereum's current hashrate.

Its currently at 270 TERAHASHES. Thats NINE MILLION 570's/580's

In order to double the hashrate, bitmain would have to make and sell 1.5 MILLION E3's. At most these batches have 10-20 thousand units, which wont even make a dent in the hashrate. To put things in perspective, Litecoin's entire hashrate is made up of a couple hundred thousand L3+'s and some of that hashrate is shared by innosilicon and BW.

The only reason A3's and D3's killed GPU mining is because they were orders of magnitude faster and more efficient than the GPUs they replaced. E3's are exactly the same speed and efficiency as GPUs, they only advantage they have is that they cost less than current GPUs, and Bitmain does not have to be at the mercy of getting GPU packages from AMD/NVIDIA( and thats about the only advantage an ethereum ASIC can have, maybe slightly higher effenciey if they go full custom cores which they wont since eth will go POS before then).

Lets all calm down

[e97]

Appears they are not selling to well.

Were initially limit 1, now its limit 5.

Bitmain model is:

1) develop ASIC,
2) mine until hitting profit target
3) sell old miners to public

Repeat.

As one of the few ASIC mfgs still around it seems to be quite a successful model. They also made more rev/profit than NVIDIA this year and are buying up capacity from TSMC for manufacturing. As a company they are hitting all the right notes.

As for us miners, we basically get scraps after they've eaten the main course.

With the crypto craze seemingly over, all the casual day traders are leaving and the institutional investors are starting to enter. Volatility is decreasing and once we hit the bottom should be a gradual build up.

[Smilee]

Something caused the Ethereum hashrate to Double in less than two months.....I dont think it was this E3.  I dont think it was new GPU's.
It was the hidden Bitmain Eth Asic.  Lets hardfork and watch the hashpower lose at least 35%!!!  Its the only way to know for sure.

[RobbMines]

Something caused the Ethereum hashrate to Double in less than two months.....I dont think it was this E3.  I dont think it was new GPU's.
It was the hidden Bitmain Eth Asic.  Lets hardfork and watch the hashpower lose at least 35%!!!  Its the only way to know for sure.

Steady rise since December = batch 1 units hit network. Bitmain is rumored to have mined for 3 - 6 months with equipment before refreshing hardware and selling the old stuff. Very likely what's going on here.

[nsummy]

What is up with everyone in this thread acting like ETH is the only mineable coin?  There are a shitload of other algos out there.  Unless you are using old AMD cards, mining ETH hasn't been the most profitable coin to mine in almost a year.  Blows my mind how people act like the sky is falling and its the end of mining because Bitmain is releasing an "ASIC" that is underpowered and overpriced and not even available for 3 months.

What is up with you and your head in the clouds not realizing that ETH is the main source of income of GPUs both directly and indirectly. Yes, its not the most profitable coin to mine but its what FARMS mine for a steady source of income. If ETH tanks all those mining farms will eat up your little GPU rigs and nothing will be profitable. You are a little fish who is lost in a big sea of miners.

for real 😳 who are these newbies that have no idea how bad the gpu mining buisness was before ether ,ether saved gpu mining

it absolutely blows my mind HOW MANY PEOPLE do not understand the severity of this situation from the POV of a GPU Miner let alone bitmain ruling crypto / centralization and so forth

Also I heard you never receive merits, so here's a merit my friend lol

As the dude in the Big Lebowski says, “That’s just like, your opinion, man”   ETH did not save GPU mining.  It helped popularize it, there is no proof all of these other coins wouldn't have came around if Ethereum never existed.  As for these mining farms, what kind of cards do they have?  I'm guessing the people who parrot about these "farms" are clueless to the size, scope, and equipment.  It doesn't matter how large your operation is, the fact is that if you have 5 or 5000 GTX 10XX cards, ETH has not been the most profitable coin to mine for a long time.  And if these huge mining farms have AMD cards, what the hell else are they going to switch their hashrate to?   There are a million different coins to mine.

god you are an imbecile, amd cards cost a third of the price of nvidia cards when most of these farms were build, for each 1070 or 1080ti there are 3 or more rx470 or 570 cards purchased for the same price.  Two 470s easily do the work of 1 1070at any algo, 3 the work of a 1080ti.

these cards are undervolted to the max and on top of that power is not a concern for large farms since they pay .02 cents per watt or less.

This so called Nvidia advantage is false when it comes to scaled up opperation. 

the fact thst you think a 5 vs a 5000 gpu farm faces the same issues shows how clueless you are

carry on tho 😂

You are missing my point.  I am saying if ETH went away tomorrow, where would this hashrate switch to exactly?  Like you said, these farms are all using AMD cards, and the only advantage AMD has is with ETH and Cryptonight.  Pretty much every other algo gives the advantage to Nvidia.  I'm guessing a lot of these AMD cards at the farms would simply get sold.  And if there are giant nvidia farms, they are already mining something besides ETH.

[senseless]

(Bitmain: professional evil....professionally done...we do evil well!) (tm: Bitmain)

Bitmain looks like an American company where only money really matters, hard to say it's Chinese. I wonder where is the socialism and equality. That jihan guy maybe spent most of his life with evil american manufactures. Learnt from them.

You've obviously never been to china.

[m.vina]

With current difficulty, this Bitmain Garbage can only mine for a round $4 a day. Imagine all the batches are come a live and then difficulty sky rocketed way higher.
Both this ASIC miner and GPU have to say good bye to ETH.
Btw, are you sure that this is the best Bitmain can do? I somehow have intuition that they have more powerful hashing board to mine ETH under their wrap still. Way more powerful that this garbage.

This garbage has officially killed ETH gpu mining for now. Until ETh dev comes up with something or AMD or Nvidia develops a new gpu specifically designed for mining.
This garbage is still very low cost and very scalable. Good for mining farms. So it will increase difficulty on all the coins as big gpu farms will switch coins.

This wont kill eth mining anytime soon. These are not A3's/D3's etc. Anyone worried about these has no sense of the magnitude of Ethereum's current hashrate.

Its currently at 270 TERAHASHES. Thats NINE MILLION 570's/580's

In order to double the hashrate, bitmain would have to make and sell 1.5 MILLION E3's. At most these batches have 10-20 thousand units, which wont even make a dent in the hashrate. To put things in perspective, Litecoin's entire hashrate is made up of a couple hundred thousand L3+'s and some of that hashrate is shared by innosilicon and BW.

The only reason A3's and D3's killed GPU mining is because they were orders of magnitude faster and more efficient than the GPUs they replaced. E3's are exactly the same speed and efficiency as GPUs, they only advantage they have is that they cost less than current GPUs, and Bitmain does not have to be at the mercy of getting GPU packages from AMD/NVIDIA( and thats about the only advantage an ethereum ASIC can have, maybe slightly higher effenciey if they go full custom cores which they wont since eth will go POS before then).

Lets all calm down

I've read your analysis and now i'm curious. Given your arguments, i wouldn't say you are PRO bitmain E3 ASICs but you definitely aren't against them. Are you, jstefanop or QuintLeo, personally buying some of these E3 ASICs? Do you think it would be a wise move to sell a GPU-rig and instead buy one of these now?

[stash2coin]

Why are people even talking about the Bitmain AI chip and this thing?

That chip is optimized for FLOATING POINT operations, like any AI chip - it is NOT going to process INTEGER operations very well like mining requires, NOR is it going to be efficient at doing so.

The BM 1680 is NOT A MINING CHIP and is not going to offer decent performance per $ or watt since the bulk of the chip would go to waste AND WASTE SOME POWER if someone tried to mine on it.

Could be not the same chip but something very similar, if you look at the block diagram it resembles very closely GPU without texturing and rasterization units(you dont need these for mining) and pcie bus is moved to an fpga chip. They just have to substitute these "NP" units with something more appropriate for mining. 

[QuintLeo]

I've read your analysis and now i'm curious. Given your arguments, i wouldn't say you are PRO bitmain E3 ASICs but you definitely aren't against them. Are you, jstefanop or QuintLeo, personally buying some of these E3 ASICs? Do you think it would be a wise move to sell a GPU-rig and instead buy one of these now?

I am not planning to buy any because of the uncertainty about "when" on the move to Proof of Stake and that I'm currently very short on spare cash, coupled with the relatively LOW reliability of most Bitmain options.
I am not planning to sell a rig to finance buying one (or more likely 2).

I'm on the fence on "would it be a good idea" - if you had sold the rig a month ago when GPU pricing was still crazy-inflated, it would have made sense even WITH the uncertainty of "will an E3 manage to pay for itself" if the E3 was going to ship in the next week or two, and probably even with the "not ship 'till July" issue.
I'm not going to say it's a BAD idea - but it's a bit higher risk than I'm willing to take right now.

If I did buy an ASIC miner in this month, there are other options that I suspect will be more profitable longer than the E3 would be.


There's also the lingering uncertainty about the rumored "F3", if it exists, and what algorithm it's FOR if it does actually exist.
If it turns out the numbers in that one video that caused all of the rumors are correct, and *if* it was mining on Ethereum, that thing WOULD have a noticeable impact sooner than the E3 would - unless the E3 is a 28nm effort and the F3 is a 14/16 nm effort that has to compete for very limited foundry capacity that would reduce it's available numbers to be built A LOT.

[QuintLeo]

In order to double the hashrate, bitmain would have to make and sell 1.5 MILLION E3's. At most these batches have 10-20 thousand units, which wont even make a dent in the hashrate. To put things in perspective, Litecoin's entire hashrate is made up of a couple hundred thousand L3+'s and some of that hashrate is shared by innosilicon and BW.

Close - there are probably stlll some "last gen" units like the Innosilicon A2, KNC Titan, and Alcheminer in use since they're still PROFITABLE if you have 4 cent/kwh or less electric cost - but the L3+/A4/L21/A6 have long since become the bulk of Scrypt mining hashrate and likely account for close to 95% of that hashrate.

[badfad]

If this shipped sooner I'd buy it, still you can get 2 eth for what this is worth now

[bigjee]

okay so looks like we have a few updates from the Eth dev team after they discussed the relevant eip.

-Casper implementation is scheduled / budgeted around July mid (when these ASICS are due for launch). Progress is good so far and making in roads.
-There is currently a 2.5x cost reduction with e3 but no real gain in hash rate; threat for 51% attack is minimal.
-No emergency measures will need to be taken at the moment; however plan for alternative algo like sha-3 will be integrated into next hard fork.
-If bitmain threatens with higher hashing power and possible 51% attack eth will counter with a cryptonight like hard fork OR they may expedite casper POS launch without bug testing etc. (option 1 seems to be the more widely accepted scenario).
-eth dev will be able to detect any spikes in hashrate etc that signals a high hash rate asic (like F3 eg.) and may choose hard fork earlier.

https://www.youtube.com/watch?v=SoPfoNpqG0k

[rdluffy]

Let's buy everything guys, let's give power to Jihan

[QuintLeo]

okay so looks like we have a few updates from the Eth dev team after they discussed the relevant eip.

-Casper implementation is scheduled / budgeted around July mid (when these ASICS are due for launch). Progress is good so far and making in roads.
-There is currently a 2.5x cost reduction with e3 but no real gain in hash rate; threat for 51% attack is minimal.
-No emergency measures will need to be taken at the moment; however plan for alternative algo like sha-3 will be integrated into next hard fork.
-If bitmain threatens with higher hashing power and possible 51% attack eth will counter with a cryptonight like hard fork OR they may expedite casper POS launch without bug testing etc. (option 1 seems to be the more widely accepted scenario).
-eth dev will be able to detect any spikes in hashrate etc that signals a high hash rate asic (like F3 eg.) and may choose hard fork earlier.

Even if the F3 exists and meets the stats in that video, it's STILL not a short-term threat - only double the hashrate of the E3, with no information about how much power it draws.
Bitmain would have to sell almost a MILLION of those to match the existing ETH hashrate, and over a million when you toss in the existing hashrate from the other ethash coins like ETC.

NOT HAPPENING in a few months timeframe - though to be fair initial Casper implementation is intented to be the 1%/99% hybrid and still mostly POW.