Pages:
Author

Topic: 21 co introducing bitcoin [mining+] computer for $399.99 (unofficial thread) - page 3. (Read 5192 times)

donator
Activity: 1414
Merit: 1051
Spondoolies, Beam & DAGlabs
.......

As usual, you are very well informed, Guy!  Wink

To be fair, with 0.16 J/GH this would be the most efficient chip in the wild so far.
The 0.6V BM1384 value is just quadratic extrapolations of the 0.66V value. Try it, it fits perfectly. IMHO too perfect for a measured value.

Anyway, 21 really screwed it up with the GH/mm2 value. In best case they have here about 2 GH/mm2 (for sure not with 0.16 J/GH). This is a good GH/mm2 value for 28nm, but for 22nm it should be 3 ... 4 GH/mm2. Otherwise this chip will be simply too expensive in production to be competitive for standard miners.



You have got to stop thinking about conventional economics here. The IoT bandwagon is impervious to conventional economic considerations. After all, folks are going to just be so happy to pay an extra $25 for their light bulb, it doesn't matter if the ASIC price is a little high. The fact that it's going to be part of the IoT relieves it from those pesky metrics.

If it's connected to the Internet, it MUST be great!!!  Smiley Smiley Smiley

Isn't Raspberry PI 2 + Sidehack's 1384 dongle is essentially the same for quarter of the price ?
alh
legendary
Activity: 1846
Merit: 1052
.......

As usual, you are very well informed, Guy!  Wink

To be fair, with 0.16 J/GH this would be the most efficient chip in the wild so far.
The 0.6V BM1384 value is just quadratic extrapolations of the 0.66V value. Try it, it fits perfectly. IMHO too perfect for a measured value.

Anyway, 21 really screwed it up with the GH/mm2 value. In best case they have here about 2 GH/mm2 (for sure not with 0.16 J/GH). This is a good GH/mm2 value for 28nm, but for 22nm it should be 3 ... 4 GH/mm2. Otherwise this chip will be simply too expensive in production to be competitive for standard miners.



You have got to stop thinking about conventional economics here. The IoT bandwagon is impervious to conventional economic considerations. After all, folks are going to just be so happy to pay an extra $25 for their light bulb, it doesn't matter if the ASIC price is a little high. The fact that it's going to be part of the IoT relieves it from those pesky metrics.

If it's connected to the Internet, it MUST be great!!!  Smiley Smiley Smiley
member
Activity: 118
Merit: 10
It's a bluff, seriously.. 400$ for such device  Shocked
Chinese will do it for 50$ - why the heck?
full member
Activity: 129
Merit: 100
Assuming 0.16 is bottom-clock efficiency, that puts it at about 10% better than BM1385. Not really revolutionary by comparison.
+1

21 1st gen was TSMC 40nm
21 2nd gen was Intel 22nm, power numbers of about 0.57 J/GH at nominal working point.
21 3rd gen was another Intel 22nm, they aimed for 0.22 J/GH according to their investors materials (taped out 24/8 2014, silicon was expected in November 2014)

I believe this is the 3rd gen. It's very bad number for a Intel 22nn, which is FinFET process.

A quote I liked from their presentation: "Approaching Moore's law: best efficiency is likely ~0.15 W/GH/s @ 14nm"

So funny.

I believe that they're working on TSMC 16nm, or Intel 14nm 4th gen.

They're not doing custom design, taking a standard cell brute force approach.

Look at the "ASIC Design Engineer" job description here:
https://web.archive.org/web/20150821050459/https://21.co/#jobs

Pure waste of funds, IMHO.

As usual, you are very well informed, Guy!  Wink

To be fair, with 0.16 J/GH this would be the most efficient chip in the wild so far.
The 0.6V BM1385 value is just quadratic extrapolations of the 0.66V value. Try it, it fits perfectly. IMHO too perfect for a measured value.

Anyway, 21 really screwed it up with the GH/mm2 value. In best case they have here about 2 GH/mm2 (for sure not with 0.16 J/GH). This is a good GH/mm2 value for 28nm, but for 22nm it should be 3 ... 4 GH/mm2. Otherwise this chip will be simply too expensive in production to be competitive for standard miners.

legendary
Activity: 3374
Merit: 1859
Curmudgeonly hardware guy
And then there's the consideration that anything operating on a wireless network, especially something as "plug and play" simple as most knuckledraggers are going to need, will probably be incredibly hackable.
alh
legendary
Activity: 1846
Merit: 1052
I also found the "WiFi connected Garbage Cans" to be truly hilarious. In my house, the garbage goes out when it stinks too much, or it's pickup day. I would love to see what a "WiFi can will cost" and how it changes the need for the trash hauler to drive down the street on Monday morning. I am absolutely not interested in doing Tech Support at my house for any new Internet connected Gadgets. If you aren't married, this may work great for you. My wife will have nothing to do with it.
alh
legendary
Activity: 1846
Merit: 1052
I don't really understand why a "machine" would need to send money to another machine. Why or when a machine would need or want money is beyond me. I thought money was an invention for people, not for machines.

How is your self-driven car going to pay for it self, when going through the car wash? Paying parking?
How are people going to pay your self driven uber car(taxi)?

The possibilities are endless!

http://www.nerdgraph.com/rise-internet-things/
<3 IoT



So do the JetCars arrive at the same time as the switchover to connected cars? The graphic referenced above is way cool and really impressive. Including the item where the IoT universe is worth $7.1 Trillion, which is supposed to be larger than the 2020 Russian GDP. I wonder how the Russian people feel about that? Looks to me like we need to get on the stick and start replacing cars at a furious pace if we want 90% of the cars on the road to be "Connected" in 2020. Maybe folks will just be so dazzled by the possibilities they'll run right out and replace their car. I am sure the.

The costs are staggering and completely out of line with reality. Maybe this can become part of the 2016 Presidential election discussion, after of course we dispense with Planned Parenthood and reducing taxes on the wealthy.   Smiley
hero member
Activity: 588
Merit: 500
It also smacks of a "Oh dear this chip is not that good, how are we going to present it?" At which point the marketing department got out the smoke & mirrors and came up a collection of stuff to wrap round it.  Smiley

Rich
legendary
Activity: 3374
Merit: 1859
Curmudgeonly hardware guy
I kinda figured the same, that it was a <28nm process and not a full-custom design.

A lot of what these guys want to do seems like a bad idea to me. I'm all for wider adoption and bitcoin promotion and such, but the idea of adding mining chips into non-miner devices is foolish. Just like with trying to deploy a farm full of stick miners, the support-component overhead per hash makes it not cost-effective.
We're already having fights over how to expand the block size to accomodate existing actual financial transactions in the chain (you know, the financial transaction recordbook), and we're approaching a halving wherein block rewards decrease, putting more emphasis on transaction fees as a source of mining revenue. So why do we want to promote a mass deployment of services based around spamming the blockchain with (likely zero-fee) dust transactions, either from the "constant stream of bitcoins" coming from your baby miner or microtransactions as a means of messaging? My hatred for IoT in general aside, I'm sure I'm not the only one considering how little sense it makes to use a public financial ledger (with finite resources), the cost of maintenance of which is supposed to be compensated by its users, as a free SMS service.

Seriously, 21. You want to spam a blockchain with tiny free transactions, go make your own and stop messing with ours.

Having a console from which you can send and receive coins in payment for goods and services, that's a great idea. Except it's been done already, for free.
donator
Activity: 1414
Merit: 1051
Spondoolies, Beam & DAGlabs
Assuming 0.16 is bottom-clock efficiency, that puts it at about 10% better than BM1385. Not really revolutionary by comparison.
+1

21 1st gen was TSMC 40nm
21 2nd gen was Intel 22nm, power numbers of about 0.57 J/GH at nominal working point.
21 3rd gen was another Intel 22nm, they aimed for 0.22 J/GH according to their investors materials (taped out 24/8 2014, silicon was expected in November 2014)

I believe this is the 3rd gen. It's very bad number for a Intel 22nn, which is FinFET process.

A quote I liked from their presentation: "Approaching Moore's law: best efficiency is likely ~0.15 W/GH/s @ 14nm"

So funny.

I believe that they're working on TSMC 16nm, or Intel 14nm 4th gen.

They're not doing custom design, taking a standard cell brute force approach.

Look at the "ASIC Design Engineer" job description here:
https://web.archive.org/web/20150821050459/https://21.co/#jobs

Pure waste of funds, IMHO.
legendary
Activity: 3374
Merit: 1859
Curmudgeonly hardware guy
Assuming 0.16 is bottom-clock efficiency, that puts it at about 10% better than BM1385. Not really revolutionary by comparison.
full member
Activity: 140
Merit: 100
I don't really understand why a "machine" would need to send money to another machine. Why or when a machine would need or want money is beyond me. I thought money was an invention for people, not for machines.

How is your self-driven car going to pay for it self, when going through the car wash? Paying parking?
How are people going to pay your self driven uber car(taxi)?

The possibilities are endless!

http://www.nerdgraph.com/rise-internet-things/
<3 IoT

full member
Activity: 129
Merit: 100
Analyzing the pictures on the 21 inc website, the die size should be about 58 mm².

With 50...125 GH/s it is probably an 28nm ASIC. With 0.16 J/GH most likely a full custom design.
legendary
Activity: 1456
Merit: 1000
I was looking at could not find it.  What NM chip is this?   

Was curious with it showing more efficient then S7.  Would really like to know more about the chip itself.
hero member
Activity: 742
Merit: 500
Just bring out an 5th miner! simple Grin
legendary
Activity: 1036
Merit: 1001
/dev/null
$400 for an API, a couple apps, and 50-125GH of hashing @ 0.16gh/j ?

yes exactly, this is what it is. but..reason why they will sold something is this statement: "Buy digital goods with the constant stream of bitcoin mined by a 21 Bitcoin Chip"

it sounds so cool, imagine..you will order some small box and it will comes with wallet, node (you are helping network, wow) and MINER. and all this with geeky feeling of command line for 400USD. this device is not for bitcoiners but for wannabe geeks and kids. I really can't consider it as serious product.
legendary
Activity: 1498
Merit: 1030
I don't see this "first product" being of intrest to many folks. WAY too low on performance, WAY too espen$ive, even for a "development tool" it's bloody overpriced.

 I've got to wonder what the REAL "intended use" for this thing is, the comments 21 has made on it so far make exactly ZERO sense to me.
 Technology demonstrator perhaps?
alh
legendary
Activity: 1846
Merit: 1052
So it looks like the whole point of this gadget would be to allow folks to use/abuse the Bitcoin network to do "stuff" that isn't actually related to Bitcoin (perhaps). That would make sense within the context of the "Blockchain is preloaded on the SD card". Kinda like using a BTC wallet address instead an IP address to communicate with somebody maybe? The blockchain seems poorly engineered with rising difficulty and such, but maybe this is the idea. I don't really understand why a "machine" would need to send money to another machine. Why or when a machine would need or want money is beyond me. I thought money was an invention for people, not for machines.

Oh well, fun to watch, since I have zero of that $116M at risk. Maybe I should blow $400 to just find out what's really special, and what's just cobbled together from Micro-Center and the Raspberry Pi foundation.

I also noticed the "datacenter backend" mentioned. Is that something that 21Inc is running or what? I always wonder if when a company supplies a "service" that's part of your product, what happens if/when they go under. I remember a class of DVD player that had specially encoded DVD's that you would purchase cheap, and then get a two day play period. You could pay more and then "own" the disk. That only worked until the company went under, and then your "owned" disks were unplayable, because the server disappeared when the company died.
alh
legendary
Activity: 1846
Merit: 1052
I would say that probably the most valuable aspect of this announcement and mid-November delivery will be to demonstrate their actual chips. I expect it will help satisfy their investors (i.e. the $116M of VC money), that they actually produced something like they said they would. It also will allow other folks to actually tinker with it and perhaps do something with it. That's probably secondary in their mind.

As a $400 100GH miner it will fall flat. It may well be more successful than the $200 SFARDS development board though. After all, you still have some useful hardware in 6 months (e.g. Pi, SD card, etc). 

I will also breathlessly await the next IoT thing from 21Inc, so I can laugh at how silly it is to try and make everything that has electricity do something with BTC.  Smiley
newbie
Activity: 1
Merit: 0
Just wanted to say that this things main purpose is not mining.

See https://21.co/faq/#making-a-profit and https://21.co/faq/#making-a-profit-part2

"If you do want to pursue the commodity business of mining-to-sell-for-dollars, you can certainly try to use the 21 Bitcoin Computer for that purpose as it does have a highly energy efficient chip— but that is not the intended use."

This more like a payment system for APIs and IoT devices, where amounts for service/work are tiny and mucking around with credit cards, payments gateways and handling subscriptions is too tedious. In other words, machine-to-machine payments.

They definitely need to improve their copywriting, but on the other hand I can see that people are locked into bitcoin + hardware == mining mindset and thats hard to change.
Pages:
Jump to: