7nm miner thread

[philipma1957]

so two technical guys say it is a long wait.

and every economical  analysis I do shows no need for them  to make them  this year or next year or two years from now.


Plus no one  meaning :

samsung
apple
intel
nvidia
amd

 has shown  working quality 7nm gear  for the present  moment.

this is a long wait.

please remember the three screen shots above and that the L3 is 3.3%  for power and 96.7%  for profit on hashnest

and bit main has shown us they don't give a darn about which coin makes the money.  we will not be seeing 7nm gear from bitmain  for years.

[1714914100]

1714914100

[1714914100]

1714914100

[1714914100]

1714914100

[1714914100]

1714914100

[1714914100]

1714914100

[NotFuzzyWarm]

so two technical guys say it is a long wait.
and every economical  analysis I do shows no need for them  to make them  this year or next year or two years from now.

Plus no one  meaning :

samsung
apple
intel
nvidia
amd

has shown  working quality 7nm gear  for the present  moment.
this is a long wait.
<snip>

Much less the fact that they have yet to get current 16/14nm chips as stable/reliable as the 22nm node is/was. I believe you have made mention of vid card failure rates going up as node size went down... I know I keep seeing pre-order ads for the Latest Samsung Galaxy (10nm CPU and base-band chips) but are they shipping yet?

[philipma1957]

so two technical guys say it is a long wait.
and every economical  analysis I do shows no need for them  to make them  this year or next year or two years from now.

Plus no one  meaning :

samsung
apple
intel
nvidia
amd

has shown  working quality 7nm gear  for the present  moment.
this is a long wait.
<snip>

Much less the fact that they have yet to get current 16/14nm chips as stable/reliable as the 22nm node is/was. I believe you have made mention of vid card failure rates going up as node size went down... I know I keep seeing pre-order ads for the Latest Samsung Galaxy (10nm CPU and base-band chips) but are they shipping yet?

I know there  was  a higher failure rate for the i7 6700x cpu's.

 I had one fall myself  and it is the only intel cpu I ever killed since 2005 other then 1 i7 in a 2011 mac mini.  this would be out of over 350 intel cpus I have worked with on modded gear.

Even forgetting  all the engineering feats need to stop the bleeding or leaking of electrons in the 16/14's .

 I have read the 7's  will not be silicon at all.
 My chip design/build knowledge does not touch yours, but the economic reality is  that the 7nm is not needed for mining  bitmain is making a fortune with  16/14 mode.

[QuintLeo]

Going by commentary out of IBM, *THEIR* 7nm process will be using a "hybrid" Silicon/Germanium wafer.
No solid word out of anyone else except an Intel comment that "10nm will be the end of the line for pure silicon".


 Economic reality is that no miner maker will be making new gear on the new node(s) for a while, but that they WILL have to make the move for competative reasons probably 2 years more-or-less from now when the new node(s) start getting affordable yield figures and the relative efficiency has become known between them.

If the new nodes take longer to deploy than the current projections (like the 14/16nm generation did), it could be 3-5 years before miner makers will be ABLE to move to them.

[philipma1957]

Going by commentary out of IBM, *THEIR* 7nm process will be using a "hybrid" Silicon/Germanium wafer.
No solid word out of anyone else except an Intel comment that "10nm will be the end of the line for pure silicon".


 Economic reality is that no miner maker will be making new gear on the new node(s) for a while, but that they WILL have to make the move for competative reasons probably 2 years more-or-less from now when the new node(s) start getting affordable yield figures and the relative efficiency has become known between them.

If the new nodes take longer to deploy than the current projections (like the 14/16nm generation did), it could be 3-5 years before miner makers will be ABLE to move to them.

yeah  2-5  covers it  and my gut tells me  about 4 on the button.

also  if we are all around in 2021  my guess is that 7nm chip  will be around longer then the 16/14 chip say 10 years or more

much like  the gasoline engine does not get much improvement anymore chips will reach their limits very soon.

[wmabern]

Should be amusing to read what the masses of folks here folks with no ties to the Semiconductor Industry come up with for their Speculative dreaming thoughts.  

300ths at 500W, shipping Nov 2017!  

I will be getting three of these. Sounds great!

Make that four. I've always to have over 1ph of my own gear! 

~~~~~~~~~~~~~~~~

I was just trying to be a smart-A$$ in the above post - hopefully twas taken as such!  

Seriously, I have no clue regarding time-frame, specs, design, and true hash rate specs. But...

I'm also wondering is the current design of just about all the models that have come around going to remain the same or come out with some very innovative new design none of us have thought about.

I certainly hope I'm around, and will be able to get one from whomever puts it out. Most probably Bitmain, but who knows.? Price? Who knows. But if they have made the extraordinary leap to 20TH or even 25TH. That would be amazing and the cost might be that amazing also. Maybe twice the price? Is that out of line with what others may think?

If it's something like 20-25 TH/s, difficulty is going to Jupiter!!

Perhaps it would also be wise to wait for the 2nd or maybe 3rd iteration of the miner just to get any of the production trouble ironed out. Be amazing if the first batch was rock-solid. doubtful...

Anyhow, I'm just rambling with a nice Friday early-evening beer buzz.  

OK, I can dream. But it would be very nice to have four new 7nm miners. Then I will have that psychological point of  I shall simply have 100TH/s. But who knows what the whole BTC and mining scene / landscape will be like then. The "having" 100TH/s is a nice psychological place to be.

Of course I would still have my approximate 60TH/s mini farm.( add another 13TH when I can get my S9 controller card replaced and get the one bad hash board repaired. Not looking forward to dealing with it at present. It is in pieces and I have used some parts swapping around.
Anyway, I seriously doubt I have the ability to have four more miners in my house for many reason, power, heat nose,yada.. yada... yada  Might be time to start thinking of hosting somewhere.

OKAY, I know I'm starting to ramble... I'm making plans on power sources in my house for a new appliance (miners) that is not even a solid, material object that can be touched. Imagination-ware!Bye!

Let us hit some BLOCKSSSSSSSS!!!

That is all. You may now return to your regularly scheduled programs already in progresssssssssssss  

[Sr.Urbanist]

It sounds like there is some skepticism on the 7 nm.

 Do you think the Japanese firm claiming they'll be out with a 7 nm next year is full of hot air or do you take them serious?

[Tomintx]

It sounds like there is some skepticism on the 7 nm.

Do you think the Japanese firm claiming they'll be out with a 7 nm next year is full of hot air or do you take them serious?

Unlike most people on this list, I am in the chip biz.  I cannot say much.  I can say this:  A US$300M investment in a 7nm-based chip design will get you pretty much that.  Then you'll need more money to get your chips fabbed and you systems built and installed and everything else.  A few $millions will go a long way towards physical plant for installation, but it takes a lot of money to play in the 7nm game. 

There were some comments in this thread about the move from 22/20nm to 16nm.  The 20nm node was largely skipped by the industry because 16nm node added so much value with the introduction of finFETs, that it wasn't worth it to continue investing in 20nm.  The 16nm node still has a ways to go, and has had a number of rounds of refinements.  I don't know which actual 16nm process is being used by BMT, but they could probably get some mileage just by re-spinning into a newer 16nm process.  If they cared about improving the product.  However they're in a situation where they can sell all they can make, and price is hardly an object right now, so why waste the engineering resources? 

7nm is a huge wall to climb for anyone, and I don't think anyone's going to be building a 7nm miner for at least another year, maybe 2. 

If someone does come along and proves me wrong, I'll buy it.

[QuintLeo]

It sounds like there is some skepticism on the 7 nm.

Do you think the Japanese firm claiming they'll be out with a 7 nm next year is full of hot air or do you take them serious?

Unlike most people on this list, I am in the chip biz.  I cannot say much.  I can say this:  A US$300M investment in a 7nm-based chip design will get you pretty much that.  Then you'll need more money to get your chips fabbed and you systems built and installed and everything else.  A few $millions will go a long way towards physical plant for installation, but it takes a lot of money to play in the 7nm game. 

There were some comments in this thread about the move from 22/20nm to 16nm.  The 20nm node was largely skipped by the industry because 16nm node added so much value with the introduction of finFETs, that it wasn't worth it to continue investing in 20nm.  The 16nm node still has a ways to go, and has had a number of rounds of refinements.  I don't know which actual 16nm process is being used by BMT, but they could probably get some mileage just by re-spinning into a newer 16nm process.  If they cared about improving the product.  However they're in a situation where they can sell all they can make, and price is hardly an object right now, so why waste the engineering resources? 

7nm is a huge wall to climb for anyone, and I don't think anyone's going to be building a 7nm miner for at least another year, maybe 2. 

If someone does come along and proves me wrong, I'll buy it.

 22nm is big in the memory industry and was for Intel, but most others skipped it as it wasn't ENOUGH of an improvement over well-optimised 28nm to be worth the hugh development costs.

 I'm pretty sure that Japanese company is way optimistic on their announced timeframe - not so much blowing smoke but haven't looked at ALL the factors.

[NotFuzzyWarm]

I'm pretty sure that Japanese company is way optimistic on their announced timeframe - not so much blowing smoke but haven't looked at ALL the factors.

Ja. Rather like that infamous line from a BFL exec over their 28nm node Monarch delays that went something along the line of, "In the semiconductor industry delays are very common and are to be expected" as he shrugged off the growing clamor of very pissed off customers who had pre-ordered and Paid In Full months earlier..

In other words, they know damn well they are basing their PR projections on Fairy tale assumptions that all ongoing process problems will be miraculously solved and no other glitches popping up. They also damn well know that is NOT how the Real World works.

[mustangy]

I doubt we'll see 10 nm (TSMC) or 7nm (GF) available to miner makers before 2019 or VERY VERY LATE 2018.

 The first experimental chips might show up in early 2018 or very late 2017, but there won't be capacity to spare from the BIG long-term contract customers for months after that.


 With that said - I'd predict the S11 will be around 35 TH at around 1300 watts (at the wall with the AP3++) and available in Febuary 2019 at the soonest.

thanks, i was about to sell all s9 and l3 d3
because of good market price on eBay.
the problem is no colo now and I m stuck to run machines desperately look for good amp industrial zone small rentals, keep searching

[Tomintx]

I'm pretty sure that Japanese company is way optimistic on their announced timeframe - not so much blowing smoke but haven't looked at ALL the factors.

I'm wondering if our friend Sonny Vleisides has gone to Japan...  This sounds a bit like his type of operation.

[QuintLeo]

I doubt we'll see 10 nm (TSMC) or 7nm (GF) available to miner makers before 2019 or VERY VERY LATE 2018.

 The first experimental chips might show up in early 2018 or very late 2017, but there won't be capacity to spare from the BIG long-term contract customers for months after that.


 With that said - I'd predict the S11 will be around 35 TH at around 1300 watts (at the wall with the AP3++) and available in Febuary 2019 at the soonest.

thanks, i was about to sell all s9 and l3 d3
because of good market price on eBay.
the problem is no colo now and I m stuck to run machines desperately look for good amp industrial zone small rentals, keep searching

 Since the D3 is at best the #3 miner CURRENTLY available for X11, it might be worth selling it off early batch ones soon before the BIG difficulty jumps hit.
 S9, no point - it's the best AND the most cost-effective on the market WHEN it works (though BitFury based gear SHOULD be capable of getting close).
 L3+ is pretty much a tossup for the best on the market WHEN it works AND is the most cost-effective.

[lottery248]

Should be amusing to read what the masses of folks here folks with no ties to the Semiconductor Industry come up with for their Speculative dreaming thoughts. 

300ths at 500W, shipping Nov 2017! 

whelp that would be epic, but why not making those cores more universal? like GPU they can mine all the altcoins including bitcoin, that would be more profitable lol.

[philipma1957]

Should be amusing to read what the masses of folks here folks with no ties to the Semiconductor Industry come up with for their Speculative dreaming thoughts. 

300ths at 500W, shipping Nov 2017! 

whelp that would be epic, but why not making those cores more universal? like GPU they can mine all the altcoins including bitcoin, that would be more profitable lol.

a universal asic chip hmmm

 i think  that is like saying why not make the color black  look like every color.

As I understand an asic chip is more or less hard wired to one al-gore-rhythm

and a gpu chip can be programed  to do any al-gore-rhythm

[QuintLeo]

Should be amusing to read what the masses of folks here folks with no ties to the Semiconductor Industry come up with for their Speculative dreaming thoughts. 

300ths at 500W, shipping Nov 2017! 

whelp that would be epic, but why not making those cores more universal? like GPU they can mine all the altcoins including bitcoin, that would be more profitable lol.

 They would be less profitable as there would be a ton of wasted space on the chip and a bunch of circuitry that even in a "powered down" state is STILL going to pull a little power.

 The whole REASON that an ASIC is efficient is that it is dedicated to doing ONE thing VERY VERY WELL, as opposed to something like a GPU that has to be set up to do many different things and is not as good at each one.

[QuintLeo]

Bit of food for thought for those that think 7nm will show up next year.

https://arstechnica.com/gadgets/2017/08/intels-next-generation-chip-plans-ice-lake-and-a-slow-10nm-transition/

Given the issues with 14/16nm, and the reported issues getting 10nm up to speed, I'm starting to think that *2019* might be excessively optimistic for actual 7nm production to arrive.

[NotFuzzyWarm]

Should be amusing to read what the masses of folks here folks with no ties to the Semiconductor Industry come up with for their Speculative dreaming thoughts. 

300ths at 500W, shipping Nov 2017! 

whelp that would be epic, but why not making those cores more universal? like GPU they can mine all the altcoins including bitcoin, that would be more profitable lol.

Been tried, failed miserably, re: SFARADS attempt in 2015 for a BTC and x11(?) combo chip. You can find it with a Github search.

At least the chip lives on as fodder from just about all of the scam vaporware rigs these days....
Still biggest problem is that being hard-wired each different algo requires different cores so in essence you get something more like a SOC - 1 big chip comprised of different IP blocks. Just not efficient use of real-estate.

[QuintLeo]

whelp that would be epic, but why not making those cores more universal? like GPU they can mine all the altcoins including bitcoin, that would be more profitable lol.

Been tried, failed miserably, re: SFARADS attempt in 2015 for a BTC and x11(?) combo chip. You can find it with a Github search.

At least the chip lives on as fodder from just about all of the scam vaporware rigs these days....
Still biggest problem is that being hard-wired each different algo requires different cores so in essence you get something more like a SOC - 1 big chip comprised of different IP blocks. Just not efficient use of real-estate.

 3 options that have ever been built as ASIC.

 Gridseed and the GC 3355 - which was very quickly not competative on SHA256 efficiency if it EVER was, but was the only option for some months for Scrypt.
 SFARDS SF100 - second-generation of the GC3355, again by the time it shipped it was no longer competative on the SHA256 side but was fairly close on the Scrypt side.

 SFARDS was the merger of Gridseed and WiiBox - they appear to have built and sold ONE batch of the SF100 then quietly died, or got out of cryptocoin entirely.

 The Baikal technically is multi-algo, but X15 INCLUDES all of the other algos supported by the Baikal miner so it was just a matter of "breaking out" input and output to the applicable sub-algos to allow it to mine on more than one algo (but not at the SAME time, the Gridseed and SFARDS chips were designed to do both Scrypt AND SHA256 at the same time).

 I'm inclined to say that the Baiklal is the only one that was SUCCESSFUL at multi-algo, the GC3355 lived for a long time as Scrypt but not SHA256 and the SFARDS was never a success at all.

[promojo]

Going by commentary out of IBM, *THEIR* 7nm process will be using a "hybrid" Silicon/Germanium wafer.
No solid word out of anyone else except an Intel comment that "10nm will be the end of the line for pure silicon".


 Economic reality is that no miner maker will be making new gear on the new node(s) for a while, but that they WILL have to make the move for competative reasons probably 2 years more-or-less from now when the new node(s) start getting affordable yield figures and the relative efficiency has become known between them.

If the new nodes take longer to deploy than the current projections (like the 14/16nm generation did), it could be 3-5 years before miner makers will be ABLE to move to them.

Interesting... Thanks for pointing that out...  Off to learn what Germanium wafer is I go.... lol