ASICMINER: Entering the Future of ASIC Mining by Inventing It

[Whtwabbit]

Can anyone tell me more about this immersion cooling? Would they be individual cooling units that anyone could use to cool down any ASIC? Are there any risks of damaging your machine? How much money could these units save someone who is mining?

The precise details are still to come, but the basic idea behind immersion cooling is that air is a terrible heat conductor, which means that you need fans spinning rapidly to pump as much air past the heatsinks/chips in order to keep it cool. Those fans use power (and generate noise, which is less important in a datacenter) of their own and in a datacenter there's enough heat being generated that the ambient air in the room needs to be cooled as well, requiring AC units, which add to the power usage.

With immersion cooling, you can immerse all machines in one giant container of oil, which has been selected to have low viscosity (= easy to pump around) and high thermal conductivity (= takes up a lot of heat). Unlike air, you can actually finetune the properties of the oil. The oil is then pumped around and can, in theory, simply be cooled in the outside air, provided you have a big enough heatsink.

If set up properly, immersion cooling is more efficient than simple air cooling. The hardware is less accessible, so for a datacenter where hardware is regularly replaced or added, it's not so ideal. But a mining farm is pretty much set-and-forget.

So unless you actually had a mining farm it probably wouldn't be worth the effort to buy a cooling unit? The thing that worries me is that the potential market might not be that big if it doesn't include hobbyists who buy one or two machines. What percentage of people involved in mining are running farms? Then how many of those will buy cooling instead of just opening the window in the winter?

I'm not sure..... but I think mining farms could be growth industry in the future

[1715399008]

1715399008

[hlynur]

Can anyone tell me more about this immersion cooling? Would they be individual cooling units that anyone could use to cool down any ASIC? Are there any risks of damaging your machine? How much money could these units save someone who is mining?

The precise details are still to come, but the basic idea behind immersion cooling is that air is a terrible heat conductor, which means that you need fans spinning rapidly to pump as much air past the heatsinks/chips in order to keep it cool. Those fans use power (and generate noise, which is less important in a datacenter) of their own and in a datacenter there's enough heat being generated that the ambient air in the room needs to be cooled as well, requiring AC units, which add to the power usage.

With immersion cooling, you can immerse all machines in one giant container of oil, which has been selected to have low viscosity (= easy to pump around) and high thermal conductivity (= takes up a lot of heat). Unlike air, you can actually finetune the properties of the oil. The oil is then pumped around and can, in theory, simply be cooled in the outside air, provided you have a big enough heatsink.

If set up properly, immersion cooling is more efficient than simple air cooling. The hardware is less accessible, so for a datacenter where hardware is regularly replaced or added, it's not so ideal. But a mining farm is pretty much set-and-forget.

So unless you actually had a mining farm it probably wouldn't be worth the effort to buy a cooling unit? The thing that worries me is that the potential market might not be that big if it doesn't include hobbyists who buy one or two machines. What percentage of people involved in mining are running farms? Then how many of those will buy cooling instead of just opening the window in the winter?

just my two satoshis here,
imo immersion cooling isn't a technology solely for use in mining coins.
i like the approach AM is doing there.
sooner or later AM has to concentrate much more on several sectors in the whole mining market and try to push technological innovations.
(though from my amateurish perspective i can't answer if this cooling concept could even be useful in other markets e.g. renderfarms).
it looks very difficult to me to run succesfully a mining farm and develop chips/ asic designs for endconsumers as well as mining companies at the same time
as a small company AM has to diversify somehow and get a foot in the door for coming demands of the future mining market. (if that should include hobbyists is another question)
If they have innovative ideas on the side during their evolution as a company it's not bad to give it a try and perhaps even specialize.

[kmtan]

it will be great to see the financial report soon on Oct 20.

I suppose that depends on the contents of the report. I do have high expectations though.

Hope that AM will give good contents and excited news to be announce towards.

[Vycid]

imo immersion cooling isn't a technology solely for use in mining coins.

No kidding. It isn't AM's idea either.

http://www.datacenterknowledge.com/archives/2012/10/08/3m-demos-new-immersion-cooling-technique/
http://techreport.com/news/20766/immersion-cooling-comes-to-data-centers
http://en.wikipedia.org/wiki/Computer_cooling#Liquid_submersion_cooling

Of course, if it was practical, it'd already be deployed on a large scale. Bitcoin mining is unusually power-intense and so exotic cooling solutions make more sense.

[stripykitteh]

imo immersion cooling isn't a technology solely for use in mining coins.

No kidding. It isn't AM's idea either.

http://www.datacenterknowledge.com/archives/2012/10/08/3m-demos-new-immersion-cooling-technique/
http://techreport.com/news/20766/immersion-cooling-comes-to-data-centers
http://en.wikipedia.org/wiki/Computer_cooling#Liquid_submersion_cooling

Of course, if it was practical, it'd already be deployed on a large scale. Bitcoin mining is unusually power-intense and so exotic cooling solutions make more sense.

I imagine there would be some interesting design possibilities. If a design assumption is the hardware will only ever run immersion-cooled, much higher densities and clock-rates could be possible.  Plus you can fry your chips.

[VeeMiner]

sorry for the crappy image, I only have ms paint here

[bbxx]

imo immersion cooling isn't a technology solely for use in mining coins.

No kidding. It isn't AM's idea either.

http://www.datacenterknowledge.com/archives/2012/10/08/3m-demos-new-immersion-cooling-technique/
http://techreport.com/news/20766/immersion-cooling-comes-to-data-centers
http://en.wikipedia.org/wiki/Computer_cooling#Liquid_submersion_cooling

Of course, if it was practical, it'd already be deployed on a large scale. Bitcoin mining is unusually power-intense and so exotic cooling solutions make more sense.

I imagine there would be some interesting design possibilities. If a design assumption is the hardware will only ever run immersion-cooled, much higher densities and clock-rates could be possible.  Plus you can fry your chips.

With immersion cooling you can ged rid of heatsink, it takes so much space and is heavy,
You can double your density that way.
Heat in liquid form is much more usefull too, you can use it for heating hot tube, or whatever.

In few months from now mining will depends on margins, here you can sell/reuse heat easily.

[VeeMiner]

imo immersion cooling isn't a technology solely for use in mining coins.

No kidding. It isn't AM's idea either.

http://www.datacenterknowledge.com/archives/2012/10/08/3m-demos-new-immersion-cooling-technique/
http://techreport.com/news/20766/immersion-cooling-comes-to-data-centers
http://en.wikipedia.org/wiki/Computer_cooling#Liquid_submersion_cooling

Of course, if it was practical, it'd already be deployed on a large scale. Bitcoin mining is unusually power-intense and so exotic cooling solutions make more sense.

I imagine there would be some interesting design possibilities. If a design assumption is the hardware will only ever run immersion-cooled, much higher densities and clock-rates could be possible.  Plus you can fry your chips.

With immersion cooling you can ged rid of heatsink, it takes so much space and is heavy,
You can double your density that way.
Heat in liquid form is much more usefull too, you can use it for heating hot tube, or whatever.

In few months from now mining will depends on margins, here you can sell/reuse heat easily.

yeah, immersion cooling is a great idea and it will definitely be a game changer for the bitcoin mining. I only hope that the gen 2 and 3 chips have priority over this...

[Lohoris]

I only hope that the gen 2 and 3 chips have priority over this...

I'm not sure.
Apparently we have already lost with gen2, so pushing it further might be just wasting money.
Starting something new like specialised cooling systems instead, might give us the lead in that field, in case we are successfull.

[VeeMiner]

I only hope that the gen 2 and 3 chips have priority over this...

I'm not sure.
Apparently we have already lost with gen2, so pushing it further might be just wasting money.
Starting something new like specialised cooling systems instead, might give us the lead in that field, in case we are successfull.

it's true that if the gen 2 were supposed to be 60nm then it would be outdated by the time that they came out. I just hope that the gen 3 would be around 28nm full custom so that we stay competitive with the likes of Cointerra, Hashfast and KnCminer

[Lohoris]

it's true that if the gen 2 were supposed to be 60nm then it would be outdated by the time that they came out. I just hope that the gen 3 would be around 28nm full custom so that we stay competitive with the likes of Cointerra, Hashfast and KnCminer

If they managed to fail gen2 despite such an headstart, how can you hope they will succeed in a much more difficult task on the same field, like gen 3 is?
On the other hand, completely switching field might make sense, at least there is some chance they'll again use their creativity to build something the market wants.

[ex-trader]

On the other hand, completely switching field might make sense, at least there is some chance they'll again use their creativity to build something the market wants.

Seriously? They had a great lead in ASICs and dominated the market for a few months, paid great divs which everyone thought would last forever (big mistake). Now they are a tiny player, have lost the lead and have yet to show clear guidance about their next steps, other than a big batch of outdated chips about to arrive.

The idea that they'd be worth anything based on them finding another business to make money from is naive.

[robix]

On the other hand, completely switching field might make sense, at least there is some chance they'll again use their creativity to build something the market wants.

Seriously? They had a great lead in ASICs and dominated the market for a few months, paid great divs which everyone thought would last forever (big mistake). Now they are a tiny player, have lost the lead and have yet to show clear guidance about their next steps, other than a big batch of outdated chips about to arrive.

The idea that they'd be worth anything based on them finding another business to make money from is naive.

Sadly true.

[VeeMiner]

On the other hand, completely switching field might make sense, at least there is some chance they'll again use their creativity to build something the market wants.

Seriously? They had a great lead in ASICs and dominated the market for a few months, paid great divs which everyone thought would last forever (big mistake). Now they are a tiny player, have lost the lead and have yet to show clear guidance about their next steps, other than a big batch of outdated chips about to arrive.

The idea that they'd be worth anything based on them finding another business to make money from is naive.

Sadly true.

it's only partially true, the writer omited some important facts:

- they have enough money for new gen chip development
- they are based in china - cheap workforce, electricity, lots of assembly lines
- they have proven themselves time and time again in the past

so yeah, I don't think we will have to solve problems like "Will AM hashrate be over 51 percent" again, but I still think that AM is a strong and important player.

[Lohoris]

The idea that they'd be worth anything based on them finding another business to make money from is naive.

And why is that?

They failed when they needed to do the next step and build something more complex.
But they showed they are able to create something from scratch, which basically is what they would do if they started studying fields like immersion cooling.
As I said, it's like a new IPO, only they already have the funds and we already have the shares, and we at least know it's not a scam (much better than many others).

[pascal257]

Did burnside do the last dividend payment (directly to BTC address) or was that already done by friedcat?

When do I know my shares are transformed into direct shares?

[dexX7]

After some digging I have to say: immersion cooling is awesome!

Great explaination what immersion cooling is:
http://www.allied-control.com/immersion-cooling

Video demonstration with computer hardware:
http://www.youtube.com/watch?v=EN3m1bJvS4Q

"Two Phase Open Bath Immersion Cooling Thread" by DeathAndTaxes:
https://bitcointalk.org/index.php?topic=313087

[Kosmatos]

Looks like dividends for Oct 23 will be higher than for Oct 16.

Make of that what you will.

[lophie]

Looks like dividends for Oct 23 will be higher than for Oct 16.

Make of that what you will.

hopefully but because? You have to actually deploy to mine.

[notme]

http://anonymouse.org/cgi-bin/anon-www.cgi/http://anonymouse.org/cgi-bin/anon-www.cgi/http://img854.imageshack.us/img854/2175/hm9v.jpg

sorry for the crappy image, I only have ms paint here

http://gimp-win.sourceforge.net/

You're welcome.