Swedish ASIC miner company kncminer.com

[s1gs3gv]

Their engineering is disappointing, I did not expect so many months of designing for that cheap construction.

I know what you mean but considering the product lifetime of these things is so short, does it matter ?

[1714921341]

1714921341

[s1gs3gv]

Here is a better math. No preorders, buy 3Th worth of S2 for $6.3K or $5.2K if you got coupons TODAY.
Conclusion: most preorders are not worth it.

I'll wait to see what the S3 looks like before I buy any more equipment.

[btc_uzr]

Their engineering is disappointing, I did not expect so many months of designing for that cheap construction.

I know what you mean but considering the product lifetime of these things is so short, does it matter ?

What do you mean ? Their design goal was -mentioned by Bitcoinorama when they where in January still in an early design phase- to build a long lasting unit, therefore it has to be very power-efficient which requires a total redesign of the chip instead of a simple die shrink..

[jimmothy]

Here is a better math. No preorders, buy 3Th worth of S2 for $6.3K or $5.2K if you got coupons TODAY.
Conclusion: most preorders are not worth it.

I'll wait to see what the S3 looks like before I buy any more equipment.

Are you really considering buying more hardware after the datatank announcement?

They are offering $0.93/gh and $0.025/kwh..

[RoadStress]

Here is a better math. No preorders, buy 3Th worth of S2 for $6.3K or $5.2K if you got coupons TODAY.
Conclusion: most preorders are not worth it.

I'll wait to see what the S3 looks like before I buy any more equipment.

Are you really considering buying more hardware after the datatank announcement?

They are offering $0.93/gh and $0.025/kwh..

So you are selling cheap inefficient use of power in an expensive immersion cooled setup. Much marketing, less hardware.

[dropt]

Are you really considering buying more hardware after the datatank announcement?

They are offering $0.93/gh and $0.025/kwh..

Here's a link to what Jimmothy is talking about, since most of you are probably not lurking through the Securities SF.

https://bitcointalk.org/index.php?topic=655464

[jimmothy]

So you are selling cheap inefficient use of power in an expensive immersion cooled setup. Much marketing, less hardware.

Roadstress sp30 offering:$0.88/gh 0.5w/gh $0.15/kwh (starts hashing in a few months)

Datatank offering: $0.93/gh 0.6w/gh $0.025/kwh (starts hashing in a few weeks)

0.6w/gh is not inefficient and <$0.5/w for cooling is not expensive.

[Phoenix1969]

 Biodom...   always respected you and your posts...    didn't mean to sound rash, but indeed a "Razz"
All my selling days are over tho I'm thinking....
I'm with Searing's latest theory right now

[Guy Corem]

So you are selling cheap inefficient use of power in an expensive immersion cooled setup. Much marketing, less hardware.

Roadstress sp30 offering:$0.88/gh 0.5w/gh $0.15/kwh (starts hashing in a few months)

Datatank offering: $0.93/gh 0.6w/gh $0.025/kwh (starts hashing in a few weeks)

0.6w/gh is not inefficient and <$0.5/w for cooling is not expensive.

My misguided favorite AM SH. It's late and I'm going to sleep, but I need to explain something.
Assuming you want to deploy, you need to look at $/GHs cost at the deployment level. Whether it will be DC, AC IC or KNC Chicken Farm (Boden).

We have customers that knows how to deploy (without miners cost but with building, electricity, air cooling etc) at sub $200K/MW
Which is extremely cheap. Even if you do 1.4 PUE, but with $0.025/KWh (depends on your locale only), you get much better numbers than what you wrote. Especially with SP30 ...

Do the math.

[s1gs3gv]

Their engineering is disappointing, I did not expect so many months of designing for that cheap construction.

I know what you mean but considering the product lifetime of these things is so short, does it matter ?

What do you mean ? Their design goal was -mentioned by Bitcoinorama when they where in January still in an early design phase- to build a long lasting unit, therefore it has to be very power-efficient which requires a total redesign of the chip instead of a simple die shrink..

What I mean is the useful life of a Neptune or a SP30 or an S3 is probably going to be 12 months or less.
For example at a diff increase of only 15%, electricity costs at $.10/kwh and BTC at today's prices, a 6TH machine consuming 2800 watts is operating in the red as early as April 2015.

Why waste money on building a short lifecycle product like a tank ?

[tolip_wen]

10 k for 3 TH and 5 X-boxes?


FAIL.

Huge FAIL indeed.... they could have delivered that in simple 28 mm ....

Under promise and over deliver   .... Bullshit !

I luv how everyone judges the cuisine before it's served. Adds a bit of spice....

It's their OWN press-release:  5 boxes only 3 TH and about 2100 Watt with 0.7 W/g, delivery starts next week...

Why the hell release this news to the world, if you have in mind to Serve something much better next week?   It will not be a 5 TH system on less than 0.5 W/g.......  So FAIL !

I was in error.

I had not seen the 'press release'
Oddly, I rarely go there. (I have noted previously I don't get out much)

For reference
http://forum.kncminer.com/forum/news-and-anouncments/46242-the-first-20-nanometer-chip-devoted-to-bitcoin

5 boxes it is.

I wonder what happenes to the 600x400x300 dimensions that are on the actual order form.

YMMV

[jimmothy]

My misguided favorite AM SH. It's late and I'm going to sleep, but I need to explain something.
Assuming you want to deploy, you need to look at $/GHs cost at the deployment level. Whether it will be DC, AC IC or KNC Chicken Farm (Boden).

We have customers that knows to deploy (without miners cost but with building, electricity, air cooling etc) at sub $200K/MW
Which is extremely cheap. Even if you do 1.4 PUE, but with $0.025/KWh (depends on your locale only), you get much better numbers than what you wrote. Especially with SP30 ...

Do the math.

What is the monthly rent on a 1mw warehouse? I don't doubt $200k/MW is possible but that is only $300k less than an immersion container but rent savings alone will probably be more than $300k. That is in addition to the electricity savings of ~$90k/year (400KW @ $0.025/kwh)

Also how can you get better numbers with an sp30? Datatank hardware production costs are $0.53/gh vs your material costs of $0.5/gh. Will you be selling SP30 in a few weeks for $3200 each? (including shipping)

[raskul]

So you are selling cheap inefficient use of power in an expensive immersion cooled setup. Much marketing, less hardware.

Roadstress sp30 offering:$0.88/gh 0.5w/gh $0.15/kwh (starts hashing in a few months)

Datatank offering: $0.93/gh 0.6w/gh $0.025/kwh (starts hashing in a few weeks)

0.6w/gh is not inefficient and <$0.5/w for cooling is not expensive.

My misguided favorite AM SH. It's late and I'm going to sleep, but I need to explain something.
Assuming you want to deploy, you need to look at $/GHs cost at the deployment level. Whether it will be DC, AC IC or KNC Chicken Farm (Boden).

We have customers that knows how to deploy (without miners cost but with building, electricity, air cooling etc) at sub $200K/MW
Which is extremely cheap. Even if you do 1.4 PUE, but with $0.025/KWh (depends on your locale only), you get much better numbers than what you wrote. Especially with SP30 ...

Do the math.

he can't do math, he's jimmothy and he's angst-ridden about not getting his latest round of AM dividends, this is why he takes it out on the rest of the forum.
that's why he is one of the very few on my ignore.

[Biodom]

Biodom...   always respected you and your posts...    didn't mean to sound rash, but indeed a "Razz"
All my selling days are over tho I'm thinking....
I'm with Searing's latest theory right now

I have a thick skin, like an elephant   it's all cool, bro...
I am simply in search of some current mining equipment since the mining bug got me; was too late for Jups, S1(s) are getting too small to matter much, etc., etc.
Had an eye on mini, but it has to be $2K to matter. Datamining "in the tank" is swell, but i did not go through the math yet. Typically, such large scale employment does not work for investors all that well.

[jimmothy]

he can't do math, he's jimmothy and he's angst-ridden about not getting his latest round of AM dividends, this is why he takes it out on the rest of the forum.
that's why he is one of the very few on my ignore.

Really?

You invested in CBCM..

CBCM powered by spongebobtech: $2.8/gh 1w/gh $0.25/kwh
Datatank: $0.93/gh 0.6w/gh $0.025/kwh

CBCM is around 3 times more expensive, half as efficient, and running costs are 10 times as much.

Or is my math wrong?

[Guy Corem]

My misguided favorite AM SH. It's late and I'm going to sleep, but I need to explain something.
Assuming you want to deploy, you need to look at $/GHs cost at the deployment level. Whether it will be DC, AC IC or KNC Chicken Farm (Boden).

We have customers that knows to deploy (without miners cost but with building, electricity, air cooling etc) at sub $200K/MW
Which is extremely cheap. Even if you do 1.4 PUE, but with $0.025/KWh (depends on your locale only), you get much better numbers than what you wrote. Especially with SP30 ...

Do the math.

What is the monthly rent on a 1mw warehouse? I don't doubt $200k/MW is possible but that is only $300k less than an immersion container but rent savings alone will probably be more than $300k. That is in addition to the electricity savings of ~$90k/year (400KW @ $0.025/kwh)

Also how can you get better numbers with an sp30? Datatank hardware production costs are $0.53/gh vs your material costs of $0.5/gh. Will you be selling SP30 in a few weeks for $3200 each? (including shipping)

By sub $200K/MW, I meant TOTAL cost. And it's actually very big sub.

(How did you get to the conclusion that SP30 material cost is $0.5/GHs ? Don't you understand the meaning of "assume" ?)

[s1gs3gv]

Will you be selling SP30 in a few weeks for $3200 each? (including shipping)

That could well happen, depending on how diff increases go ...

[jimmothy]

My misguided favorite AM SH. It's late and I'm going to sleep, but I need to explain something.
Assuming you want to deploy, you need to look at $/GHs cost at the deployment level. Whether it will be DC, AC IC or KNC Chicken Farm (Boden).

We have customers that knows to deploy (without miners cost but with building, electricity, air cooling etc) at sub $200K/MW
Which is extremely cheap. Even if you do 1.4 PUE, but with $0.025/KWh (depends on your locale only), you get much better numbers than what you wrote. Especially with SP30 ...

Do the math.

What is the monthly rent on a 1mw warehouse? I don't doubt $200k/MW is possible but that is only $300k less than an immersion container but rent savings alone will probably be more than $300k. That is in addition to the electricity savings of ~$90k/year (400KW @ $0.025/kwh)

Also how can you get better numbers with an sp30? Datatank hardware production costs are $0.53/gh vs your material costs of $0.5/gh. Will you be selling SP30 in a few weeks for $3200 each? (including shipping)

By sub $200K/MW, I meant TOTAL cost. And it's actually very big sub.

(How did you get to the conclusion that SP30 material cost is $0.5/GHs ? Don't you understand the meaning of "assume" ?)

- SP30 BOM (without NRE) is $3,000

I'd love to know how you can get much below $200k for a 1mw DC. Rent alone should be ~$100-150K/year.

[xstr8guy]

No: one box is ONE chip.... just look at their earlier released photo of the PCB: that fits in one box (see the 6 pin connector on both photo's).

No: KnC puts more than 1 PCB in each box.

The exception was the Mercury.

Saturn, 2 ASIC 1 box
Hoopiter, 4 ASIC 1 box
Frank, 6 ASIC and lots of loose parts 1 box

YMMV

FrankenJup was 6 boxes, each with 3 ASICs.

[AGcrypto]

We will see. 

Neptune In Assembly.

We’re pleased to announce touch-down of the first batch of our new 20 nanometer Neptune processors. They are now going through the testing/tweaking/assembly stage in Sweden and shipping will commence this week.

A World First 20-Nanometer Design

The Neptune is a formidable technological achievement. It is the first commercially available 20 nanometer processor shipped to end users. KnCMiner can take pride in beating all competitors to the 20nm stage, including the likes of Intel, Apple and Samsung.

The Neptune is a 1440-core 55x55 mm package tailor-made for the Bitcoin market. Each Neptune consists of 5 chips and comes in new sturdier enclosure and packaging. Using the speed offered by a true 20nm die process each Neptune churns out more than 3 TH/s while consuming only 2.1 KW at the wall. That works out to 0,7 watts per GH/s or less.

However, there are margins built into our design and the software running our chips can only improve from today on. All future software updates to raise the hashing speed will be free of charge.

Software Updates

We've reached our initial targets on speed and power consumption. But as this is the first day we’ve had the Neptune in our lab our technicians have already implemented a few tricks to increase hashing speed and we’ll release updates to our firmware shortly.

The first two batches of the Neptune miner is all sold-out. Batch three is available for pre-order now in our web shop. Watch this space for more info on upcoming products based on our new state-of-the-art 20nm design process.

We expect to ship to Batch 1 pre-order customers this week and to Batch 2 pre-order customers before the end of the month.

Best regards,

The KnC Team, June 17, Stockholm, Sweden.