Topic: Will there be more efficient mining hardwares? (Read 386 times)

I am looking at the difficulty chart and it seems to disagree with your statement, consumers are stacking mining gears like there is no tomorrow, and mining manufacturers don't seem to be worried about their miners not being as profitable as anyone would imagine them to be, in fact, nobody gives a damn if someone who has a 20c/kWh rate is trying to compete in a field which depends 90% on the power rate.


So no, I don't think mining manufacturers need to do anything different, you know who needs to do something? it's the people who are under the impression that mining is a life hack to make you rich in a few months, those people need to wake up and smell the coffee, people can't just order a dozen mining gears and think they will end up making a profit, for the same exact reason people don't go around starting their own coffee shops, clothes shop or casinos. Mining is a complex business at its core, it's nothing different than starting a traditional business, if you open up a bar in Saudi Arabia -- it's nobody's fault that nobody entered your shop, and the beer company has to do nothing about because that beer sells just fine in the Czech Republic.

Of course, everything is true in order to be profitable, you need to reduce energy costs by using a more subtle technological process for chip production, but the latest sanctions against microelectronics manufacturers in China are unlikely to contribute to this, let's see how this process will develop in the near future.

Er?
This BITCOIN area is as alive as ever. Being dedicated only to BTC topics, anything regarding alt shitcoins are OT here so ja, nothing to be found here for them...

As for the altcoin areas -- folks in this section could care less what happens there.

Crypto mining companies need to start building more profitable miners for consumers to jump right into mining again, it looks like alt mining is dead right now, look how silent this section have come out to be after the merge, I'm sure things aren't looking so good right now but do not give in.

Um - what problem does liquid cooling solve?
Efficiency is about the same, output heat is about the same.

Liquid cooling is really only about how you remove the heat from the miner.

Moore's Law is pretty much over with.

Will we see a 10 kwatt unit with 10 watts a th basically a  1000th miner.

I do not think they will do that.
I do think a twin rack miner
1 rack for cooling--liquid based
1 rack for the gear.
the gear rack would have 6 c19 hook ups with six units doing 3000 watts basically 18k to 20k

whatsminer has a clip on what I like

https://twitter.com/whatsminer_mbt/status/1511880682129133572?lang=en

I think it is the future of mining.

the units are close to a pair of server racks in size.
more on the m53

https://bixbit.io/en/blog/post/full-review-of-microbt-whatsminer-m50-m50s-and-m53-asics-features-profit-payback

and to power it simply use these cables
https://www.ebay.com/itm/324998473461?

a full sized unit has 6 seperate slide in racks each needs 1 20 amp c19 cable

the plus's are gear stays clean
you can do 1 to 6 units per rack

If you are smart you do 5 a rack not six and you have six setups

so

A) 5/6
B) 5/6
C) 5/6
D) 5/6
E) 5/6
F) 5/6

so 25 miners and 6 mining racks and 6 cooling towers

means you have a spare if a mining rack breaks
means you have a spare if a cooling rack breaks
a true cold room as cooling racks could be in a separate room or even outdoors
and gear stays much cleaner due to sealed liquid cooling

my variation above allows for l6-20p to c19 cables or l6-30p to c19 cables.

directly set to wall l6-20r or l6-30r

these would be around 3kwatt x 5 if partially  filled 15 kwatts
or 3kwatts x 6 if filled 18 kwatts.

right now bitmain does 21watts a th so a slide in miner would do about 142th

so 1 rack is 710th or 852th depends on 5 or 6 in the rack.

if you go to 10 watts a th  you could do 1500th or 1800th per mining rack.

I do not see anything better than this,for the small commercial miner under 500 kwatts

But even a guy with a mega watt and space would prefer this over a container
 
800kwatts/20 = 40 setups one room would be as much as you want to do.
as it means 200-240 miners each with a cable to a wall socket.

Once you go to 1 megawatt containers seem the way to go.

If we follow Moore’s law, we will achieve better and more efficient miners. Moore’s law indicates that the density of the transistors or any integrated circuit inside a component would be significantly smaller and produce more and more power and efficiency.

I think everyone wants to improve their stuff and make it even more efficient. It’s not just with integrated circuits, but also in life in general, because we all know that we don’t want to be stuck on the slower than we are because we would be able to produce more and be effective.

There will always be a cost to it and having more components in a device because compacting and making things more smaller would be hard to do, so it definitely would cost more. Understanding that would be beneficial for you because you would know why it is like that.

I was thinking more about cost and simplicity. Most data centers probably can't fully power and cool everything if they were all miners.
BUT, once you start changing floor layout and the high power wiring you are changing the dynamics of the facility.
Once you start going down that path you hurt the potential resale value of the facility and can limit options later.

There are a lot of data centers out there that started as top of the line places and still are *for when they were built* but new better ways of doing things came along. Now they are not so great but still usable. Mining is one option. If you re-do it o be better for mining the next use will cost you a lot more money to re-configure it back. As would re-configuring if you are not just using regular racks.

So are you better having more spaced out miners in a place that you can re-sell later if you bought, or will not have to pay a fortune fix if you leased it. Or making it better for mining? Would need some engineers and accountants to be sure.

-Dave

Migrating a data center to a mining facility isn't the best idea in the world, the power consumption of miners is way too high compared to those regular servers, the bottom line is that you can't easily find PDUs that can supply 5kw per socket, most currently used PDUs will be capped at 4kw at best, so a lot of re-constructing will need to be done to be able to run 5kw gears which is why I don't think they will ever make anything like that.

Also, most data centers probably use 42U racks instead of 48u, but then, either way, going with the assumption that the average server consumes no more than 500w, we are talking about 25kw per rack enclosure (which I am sure is a very rare as most of them will not be above 10-15kw), that would hardly fit 4*5KW machines, in size of 2258cm x 750cm, a lot of space wasted.

So in terms of space and power, if a data center has access to excessive power, it would be best to just get rid of those racks and install normal mining shelves + new electric wiring, but do data centers really have any power to spare? I mean looking at the data it seems like Facebook which has one of the largest combinations of data centers only uses 50MW of power, where does that put the average data center at?

Yes and no, I was thinking of a miner in a footprint like one of these that are used in data centers everywhere:

https://www.apc.com/us/en/product/AR3357/apc-netshelter-sx-server-rack-enclosure-48u-black-2258h-x-750w-x-1200d-mm/?parent-subcategory-id=88954

You feed the power in from the top or bottom depending on the cooling setup and if it's a raised floor or not and that's it.
For the most part you can get 2 or 3 phase or whatever you want. As I have said a few times before, there are a lot of empty data centers out there with power and cooling and so on. If you can just roll in the miner cabinet, roll out the empty server cabinet, wire it up and go it would seem like an easy sell.

I could just be missing something but it's a standard size / form that has been around for decades. Would think it would be a simple way to do it.

-Dave

This would be a problem for many large miners, most mining PDUs out there use c19 sockets, the largest PDU i saw was rated at 4.1kw max per port at 220v.

They use a split cables c19 to 2*c13 to run Bitmain gears and c19 to c20 to run Whatsminers

The most common set up is 24 ports, so one PDU, 24 gears, max is 4.1kw they use 3.5kw at most so pretty safe.

Pushing past this will cause two issues.

For Whatsminer since they use a single PSU, you need to change the PDU (about 600$ for managed metered and 300$ for basic).

For Bitmain, you can use the same PDU with no split cables but then you lose half the number.

 Besides, all the cables running from the distributer box will likely need to be re-sized, thats a ton of work.

Besides running 5kw will probably require 3 phase set up, so all in all, i can only see them do such a thing with things like the s19 hydro where you buy the whole package pre-wired and just feed 3 phase wires to the container, but using these existing set up will be very costly for most people.

I was actually wondering if they are going to go bigger and bigger for more industrial locations. 3 to 5 Kw is nice but if you could I would think larger is better.
1 mega-miner so to speak. Let it take up a full rack space or 2 and you drop in 100A 220V power at the top and be done kind of thing. It is in theory a single point of failure for a lot of hash-rate which would be bad. But on the other side, it is just 1 device instead of 10 to deal with.

-Dave

If you can make more efficient and smaller chips just pack more in there. I like the idea of the liquid cooled bitmain, solves allot of problems.

More efficient and faster - yes there is still a slight amount improvement that is possible by using the 3nm node provided that Biitmain, Canaan, et al think it is worth the multi-millions of $$$ it would cost them to develop the chips for it.

Use less power (which = less heat) -- no. Miner manufacturers have zero incentive to make lower power miners as their biggest customer base
 - farms - are perfectly happy with ~3-5kw per miner. So, the makers will just pack more chips in the miners to hit that power usage point just they have always done.

Looking at the history of bitcoin mining and how the technology behind it has progressed continuously over the years, Can we expect more efficient miners than the ASIC miners soon? Miners that will be faster, less energy consuming and will produce lesser heat.

https://www.nicehash.com/blog/post/the-history-of-cryptocurrency-mining