Swedish ASIC miner company kncminer.com

[Equilux]

I deeply apologize, I didn't know.

crap.

Delete your post please!

[Bitcoinorama]

Damn! I clicked. And now?  

It was an innocent mistake by Noah, but just to be safe;

http://www.malwarebytes.org/

(It's freeware, as always update the malware definitions before first run), or if it was an iOS device, you'll be fine...

EDIT: Noah, re-edit your post and at the very least use the line through tool, it looks like; s, and remove the hyperlink, then mention the new domain that's been suggested. For those helpful enough to correct Noah, it may be best you also re-edit and remove the hyperlinkage as well.

[dhenson]

For those helpful enough to correct Noah, it may be best you also re-edit and remove the hyperlinkage as well.

I edited mine before I originally quoted. (cxxm vs. com)

[gateway]

its funny how things get hijacked and questions that people want from KNC get lost in the noise :/

[erk]

its funny how things get hijacked and questions that people want from KNC get lost in the noise :/

That's because there is no news coming out of KNCminer for a week, so first the FUD starts to hit the thread, then it just degrades into random chatter like most unmoderated forums do.

[Ytterbium]

I guess the question I have is... what is the chance of getting a hole-in-one with their first stab at the chip.  It seems that bitfury missed his target by about 50% in performance, but it is still a useful chip and they are going with it.

They said they would stuff as many chips in the box as they needed.  From the cad files they posted the box is pretty large and could probably fit more chips.

The question is... can KnC/OrSoC get it all right the first time around.

The fact that a company with no experience in semiconductor design had problems doesn't mean a company with years of experience building semiconductors for paying customers, presumably on time, is going to have the same problems. 

Also all these parties have to communicate and work together:

 KnC <-> ORSoC <-> unknown FPGA-ASIC conversion house <-> unknown ASIC vendor

There are so many things which can go wrong when you make an ASIC. Getting it right the first time is hard.

ORSoC is an ASIC design place, and their upstream provider is a company called eASIC, which I think works with TSMC.

I don’t think it is possible to write a calculator that would predict at which difficulty people will start shutting down their ASICS and the difficulty will stop growing. IMO the calculator by The Genesis Block is a “pessimistic” variant which was really necessary because all calculators that accounted for difficulty increase lacked the delayed mining start parameter. Still, I think in a year it gives quite accurate prediction because a long time, much longer than a year has to pass before the hashrate race is over.

The problem with all the calculators is that they assume an exponential growth in mining forever.  That would require ether an exponential growth month by month in ASIC providers forever, or an exponential growth in the amount of product they ship out the door. Is BFL going to shipping out 200x chips each month in a year, and 40,000x a year from now? Is Avalon going to be shipping millions of chips, then billions? 

Come on.

Other then moore's law (which is much, much slower then bitcoin network growth so far) you're much more likely to see a linear growth in network hashrate over the long term once things re-stabilize after the current tsunami of ASICs goes out the door.

[erk]

The problem with all the calculators is that they assume an exponential growth in mining forever.  That would require ether an exponential growth month by month in ASIC providers forever, or an exponential growth in the amount of product they ship out the door. Is BFL going to shipping out 200x chips each month in a year, and 40,000x a year from now? Is Avalon going to be shipping millions of chips, then billions? 

The stuff that goes out the door is reasonably linear, just bumpy,  the exponential diff increases came about mostly by new vendors coming online. When that stops the diff increases will be more linear too.

[kingcoin]

Also all these parties have to communicate and work together:

 KnC <-> ORSoC <-> unknown FPGA-ASIC conversion house <-> unknown ASIC vendor

There are so many things which can go wrong when you make an ASIC. Getting it right the first time is hard.

ORSoC is an ASIC design place, and their upstream provider is a company called eASIC, which I think works with TSMC.

Depends how you define "ASIC design place".  ORSoC is more an embedded software, FPGA, and front-end RTL ASIC design house. They are not back-end layout engineers so they will have to get this service elsewhere.

I also thought that eASIC was their vendor, but somebody made a comment and said that they are using a standard cell ASIC, in which they would have to outsource the back-end, layout, place & route work. Which is why I made the figure containing the chain of vendors above.

As I mentioned earlier eASIC devices has FPGA structures (LUT's and routing) and are more similar to FPGA's than ASIC's (cell based), even though they are OTP (one time programmable). However, the turn-around time for eASIC is shorter than cell based ASIC's since you don't have to create all the masks as you do with cell based ASIC's

[cypherdoc]

Has KNC reached tapeout yet?

[gateway]

Has KNC reached tapeout yet?

I have asked the same question several times over the last days as well..

The natives are getting restless KNC

[erk]

Has KNC reached tapeout yet?

I have asked the same question several times over the last days as well..

The natives are getting restless KNC

No, only the people that don't understand concepts like "shipping the first boxes in September" are getting restless. We are in July.


It's like children in the back seat moaning "Are we there yet?"

Wake me up when it's September 29th.

[gateway]

Has KNC reached tapeout yet?

I have asked the same question several times over the last days as well..

The natives are getting restless KNC

No, only the people that don't understand concepts like "September delivery" are getting restless. We are in July.

Well almost August, with September not to far away 

[Anenome5]

Has KNC reached tapeout yet?

I have asked the same question several times over the last days as well..

The natives are getting restless KNC

No, only the people that don't understand concepts like "shipping the first boxes in September" are getting restless. We are in July.

It is getting awfully close to September. Maybe late September is a more realistic shipping goal >_>

[HeRetiK]

Don't panic yet. There's still more than a month to go. When we reach the end of September and they need "two more weeks", that's the point where we can start panicking.

[kingcoin]

Don't panic yet. There's still more than a month to go. When we reach the end of September and they need "two more weeks", that's the point where we can start panicking.

Don't panic yet. There's still more than a month to go. When we reach the end of September and they need "two more weeks", that's the point where we can start panicking.

If they are making a standard cell ASIC and have not yet taped out it's not likely that they will be shipping in September. But who knows, they might have taped out long time ago...

[cypherdoc]

Don't panic yet. There's still more than a month to go. When we reach the end of September and they need "two more weeks", that's the point where we can start panicking.

Don't panic yet. There's still more than a month to go. When we reach the end of September and they need "two more weeks", that's the point where we can start panicking.

If they are making a standard cell ASIC and have not yet taped out it's not likely that they will be shipping in September. But who knows, they might have taped out long time ago...

You'd think that they would've made an announcement about something like that.

Anyways I put in for a refund. Haven't heard anything back yet.

[eve]

if no news by end of August most of us will get restless and when September comes with no concrete information some will opt for refund option. Mr.Sam is not involved in engineering or designing why is he not replying to emails and gives us an update? and where is Andreas?

[HyperMega]

Also all these parties have to communicate and work together:

 KnC <-> ORSoC <-> unknown FPGA-ASIC conversion house <-> unknown ASIC vendor

There are so many things which can go wrong when you make an ASIC. Getting it right the first time is hard.

ORSoC is an ASIC design place, and their upstream provider is a company called eASIC, which I think works with TSMC.

Depends how you define "ASIC design place".  ORSoC is more an embedded software, FPGA, and front-end RTL ASIC design house. They are not back-end layout engineers so they will have to get this service elsewhere.

I also thought that eASIC was their vendor, but somebody made a comment and said that they are using a standard cell ASIC, in which they would have to outsource the back-end, layout, place & route work. Which is why I made the figure containing the chain of vendors above.

As I mentioned earlier eASIC devices has FPGA structures (LUT's and routing) and are more similar to FPGA's than ASIC's (cell based), even though they are OTP (one time programmable). However, the turn-around time for eASIC is shorter than cell based ASIC's since you don't have to create all the masks as you do with cell based ASIC's

I agree, ORSoC is only an experienced ASIC front-end design house. There must be a design enablement partner for the 28nm physical implementation.

Besides a structured ASIC service, eASIC has a service called "easycopy". The related design flow looks like a standard cell ASIC flow. So in principle eASIC could be a choice.

But I don't think that they are doing the layout, because it was said at the KnC open day, that the ASIC will be manufactured in Asia.
 
eASIC has 2 foundry partners: GLOBALFOUNDRIES and Fujitsu.

Fujitsu has no own 28nm fab (45nm is the smallest feature size they produce in their own fabs). So the only foundry left would be GF.
BUT GF has only one 28nm fab, which is located in Europe (fab1 in Dresden, Germany). So if they did not change the foundry after the open day, it's most likely not GF and not eASIC.

Anyway, I think it's better that it is not eASIC, because they just about to start to provide 28nm, there are no proven first time right tape-outs. And the other funny point is, that VMC is claiming to go through eASIC too. Could be a small conflict of interests, if the same team implements two miner ASICs at the same time.

[Phoenix1969]

if no news by end of August most of us will get restless and when September comes with no concrete information some will opt for refund option. Mr.Sam is not involved in engineering or designing why is he not replying to emails and gives us an update? and where is Andreas?

I can answer that in 3 words...
I'll give you a hint... What happens during the month of July in Sweden?

[Gyrsur]

if no news by end of August most of us will get restless and when September comes with no concrete information some will opt for refund option. Mr.Sam is not involved in engineering or designing why is he not replying to emails and gives us an update? and where is Andreas?

I can answer that in 3 words...
I'll give you a hint... What happens during the month of July in Sweden?

summer, sex and swimming?