Can KNCMiner really deliver 28 nanometers?

[Anenome5]

Can KNCMiner really deliver 28 nanometers?

most of us would even just love to know if they can deliver anything at all, even if it's 500 nm, that would make them at least legit...  

They produced a working Mars FPGA based on their ASIC design at least. That proves at least chip-designing chops and the ability to code it to a working state.

[1715221523]

1715221523

[1715221523]

1715221523

[dree12]

People like to use OSRoC involvement as a reason why Kncminer is legit and going to do what they say they are. I agree they are likely making an ASIC and not a complete scam, however their promises of shipping time and power consumption seem like they'll be hard to keep.

Anyone that thinks that BFL didn't use an ASIC design company as reputable (or more reputable) than OSRoC is fooling themselves. Also, kncminer is using newer tech than BFL, which is more difficult to design and develop. Both companies are using newer tech than ASICMINER / Avalon and both companies are going to run into their share of complications. Except you just haven't seen the complications that Kncminer will run into yet because they're just getting started... they will come IMO.

Except for all of those companies, these were the first ASICs they'd ever made/ordered. Only BFL had some experience with FPGAs, and that doesn't necessarily copy directly over to ASICs.

Are you sure BFL used a design company? Explain the power problems then. Any seasoned ASIC company has lots of experience avoiding the common pitfalls in producing designs the BFL apparently fell into.

Orsoc by contrast, it's their business to produce these things. I agree they'll look like geniuses if they pull this off. But is it outside the realm of possibility for them to deliver? Nope. They'd just have to be extremely professional to pull this off. Are they? Seemingly they are. Now it's just a matter of waiting upon execution and seeing if they deliver.

Avalon also has experience with FPGA technology.

[coinedBit]

They produced a working Mars FPGA based on their ASIC design at least. That proves at least chip-designing chops and the ability to code it to a working state.

Actually, you would normally develop an ASIC based on a FPGA design and not vice versa: http://en.wikipedia.org/wiki/FPGA_prototype
RTLs like VHDL/Verilog need to be specifically translated into ASIC designs, FPGAs are programmable after all - when using an FPGA there's simply no such thing as dealing with things at the wafer-level.
FPGAs are comparatively easy to use, even by hobbyists with some background experience (CS/EE), with very little resources  being required. These days you are exposed to FPGA and VHDL in college.
Real ASICs designs are a hugely different thing however.

Saying that an ASIC company "has FPGA experience" is as informative as saying that Albert Einstein knew  arithmetics. There's simply NO way for an ASIC company not to know about FPGAs.
You would expect an airline pilot to be able to land a single-engine airplane, but you would not necessarily expect a single-engine pilot to land an A380, roger ?

[Anenome5]

They produced a working Mars FPGA based on their ASIC design at least. That proves at least chip-designing chops and the ability to code it to a working state.

Actually, you would normally develop an ASIC based on a FPGA design and not vice versa: http://en.wikipedia.org/wiki/FPGA_prototype

Yes, but there are certain pitfalls when you go to actually tape-out that takes seasoning and sophisticated simulation software to know how to avoid. BFL seems to have hit these snags considerably. If anyone can avoid them, it's an org such as Orsoc.

Saying that an ASIC company "has FPGA experience" is as informative as saying that Albert Einstein knew  arithmetics. There's simply NO way for an ASIC company not to know about FPGAs.

That's basically my point. Just having shipped FPGAs doesn't mean BFL could make an easy transition into ASICs.

You would expect an airline pilot to be able to land a single-engine airplane, but you would not necessarily expect a single-engine pilot to land an A380, roger ?

Indeed. So we'd be calling KNC true airline pilots, and BFL experienced in little more than Cessnas, whereas Avalon is somewhere in the middle to where at least they pulled off shipping.

[coinedBit]

You would expect an airline pilot to be able to land a single-engine airplane, but you would not necessarily expect a single-engine pilot to land an A380, roger ?

Indeed. So we'd be calling KNC true airline pilots, and BFL experienced in little more than Cessnas, whereas Avalon is somewhere in the middle to where at least they pulled off shipping.

Well, at least for the time being, I beg to differ, I will only call someone a "true airline pilot" once I have seen him actually fly, takeoff and land an airliner - but without such a track record, it's all hearsay, because I haven't actually seen too many ASIC vendors actually deliver upon their promises, despite some of them appearing more legit than others admittedly.

[aeronautical]

If anyone can deliver it's the guys from opencores.org and theres a very strong link between them and KnCminer. I would in fact say that their normal side of there business at ORSoc
was a little slow because of the EU problems so they simply turned there attention to Bitcoin. There starting position would have been nothing like BFL and Avalon, their starting experience
could only be bettered by someone like Cadence Design Systems who have been buying up industrial IP core for some years and bought lots electronics CAD companies for the past decade.
So the next step up in hardware design would be a contracted CDS by a larger company that wanted to really make an impact or damage Bitcoin.
Lots of the major E-CAD companies used opencores as a starting based for design work. I know i worked for CDS.

[KS]

If anyone can deliver it's the guys from opencores.org and theres a very strong link between them and KnCminer. I would in fact say that their normal side of there business at ORSoc
was a little slow because of the EU problems so they simply turned there attention to Bitcoin. There starting position would have been nothing like BFL and Avalon, their starting experience
could only be bettered by someone like Cadence Design Systems who have been buying up industrial IP core for some years and bought lots electronics CAD companies for the past decade.
So the next step up in hardware design would be a contracted CDS by a larger company that wanted to really make an impact or damage Bitcoin.
Lots of the major E-CAD companies used opencores as a starting based for design work. I know i worked for CDS.

But they are entering a whole new ballgame: shipping complete retail products.

Designing the core is one thing, turning it into a product and managing the whole pipeline is quite another. These are two completely different businesses.

[Anenome5]

If anyone can deliver it's the guys from opencores.org and theres a very strong link between them and KnCminer. I would in fact say that their normal side of there business at ORSoc
was a little slow because of the EU problems so they simply turned there attention to Bitcoin. There starting position would have been nothing like BFL and Avalon, their starting experience
could only be bettered by someone like Cadence Design Systems who have been buying up industrial IP core for some years and bought lots electronics CAD companies for the past decade.
So the next step up in hardware design would be a contracted CDS by a larger company that wanted to really make an impact or damage Bitcoin.
Lots of the major E-CAD companies used opencores as a starting based for design work. I know i worked for CDS.

But they are entering a whole new ballgame: shipping complete retail products.

Designing the core is one thing, turning it into a product and managing the whole pipeline is quite another. These are two completely different businesses.

They have a company making them, they have a company assembling them, what makes you think they don't have a company shipping them too?

[aeronautical]

I think the others ASIC companies got them selves in a tangle with issues like setting up SMT lines, not sure why you would bother.
I think they will get it done and in a reasonable manner so long as they use a contract manufacture and good shipping logistics.
ORSoc are smart, way above the starting point of BLF, reading the blogs it nearly sound like there's some bitchy comments and right now and everyone is an expert SHA 256 ASIC designer.

But like i said for a team at CDS to crank out a 28nm chip would be no problem that wiped the Bitcoin market so hard, they are doing all day long go check out the IP list at
http://ChipEstimate.com

[J35st3r]

But they are entering a whole new ballgame: shipping complete retail products.

Designing the core is one thing, turning it into a product and managing the whole pipeline is quite another. These are two completely different businesses.

They have a company making them, they have a company assembling them, what makes you think they don't have a company shipping them too?

What's more, they will be shipping prototypes. They don't even have a chip test strategy in place yet! It will be several months after first silicon that they will be in a position to ship production quality goods. But in this respect they are in exactly the same boat as Avalon and BFL. Welcome to the bleeding edge of electronics design

[KS]

If anyone can deliver it's the guys from opencores.org and theres a very strong link between them and KnCminer. I would in fact say that their normal side of there business at ORSoc
was a little slow because of the EU problems so they simply turned there attention to Bitcoin. There starting position would have been nothing like BFL and Avalon, their starting experience
could only be bettered by someone like Cadence Design Systems who have been buying up industrial IP core for some years and bought lots electronics CAD companies for the past decade.
So the next step up in hardware design would be a contracted CDS by a larger company that wanted to really make an impact or damage Bitcoin.
Lots of the major E-CAD companies used opencores as a starting based for design work. I know i worked for CDS.

But they are entering a whole new ballgame: shipping complete retail products.

Designing the core is one thing, turning it into a product and managing the whole pipeline is quite another. These are two completely different businesses.

They have a company making them, they have a company assembling them, what makes you think they don't have a company shipping them too?

Not what I said.

I meant they're going from and OEM/ODM business to a full retail business.  A completely different beast. Already they don't have enough personnel just for support. So, like the others, they're bound to run into snags, which means delays. I'm sure they can iron things out eventually (basically like everyone before them), if they are serious about staying in business. As I said many times already: it's wait and see time.

[Bitcoinorama]

Yo, update!:

Re: BitcoinOrama Report on the KnCminer/OrSoC Open-day Mon 10/06/13 (Stockholm)

https://bitcointalk.org/index.php?topic=232852.msg2503674#msg2503674

Haz videoz!!

[KS]

So they ARE using Altera, Stratix4 (not 5 - so Quartus II?)...only full ASIC instead of HardCopy, but on the fab. So it's either Altera or eASIC (FPGA agnostic) or another 3rd party? meh...

tic toc

[tacotime]

It's kind of funny hearing everyone talk about ORSoC like they're some kind of master chip designing company...  They're listed on LinkedIn as having between 11 and 50 employees.  Butterfly Labs, comparably, has about 40 employees.

[dree12]

It's kind of funny hearing everyone talk about ORSoC like they're some kind of master chip designing company...  They're listed on LinkedIn as having between 11 and 50 employees.  Butterfly Labs, comparably, has about 40 employees.

# of employees doesn't really matter. Avalon has far fewer than BFL, but finished earlier. ORSoC's employees are probably more efficient than BFL's, especially when you consider Josh Zerlan is included in that 40 BFL employees.

[Anenome5]

It's kind of funny hearing everyone talk about ORSoC like they're some kind of master chip designing company...  They're listed on LinkedIn as having between 11 and 50 employees.  Butterfly Labs, comparably, has about 40 employees.

Lol, and yet Orsoc still has infinity-times more chip designers, since BFL employs -zero-. Orsoc employs at least 4 people who are outright ASIC designers, professionally and full-time.

[MrHempstock]

Thanks, guys. This thread has given me more information about and insight into KNCminer than some threads that were much, MUCH longer.
 

[AstroBoy]

I'm glad this thread exists.  I was curious about KNCminer as it is one of two 28nm ASIC producers at the moment ActiveMining (a.k.a VMC)  is going with a Standard 28nm design.  It'll be interesting who hits the market first and who hits their power and hashing targets.  Both are using eASIC too which really evens the odds.

It'll be fun watching in September!

[CoinHoarder]

I will be really surprised if KNC keeps their timeline.

They are the next BFL. You heard it here first. 

If it makes you feel any better, Avalon is the first new BFL anyways (delays, customer service, etc.).

[sickpig]

I'm glad this thread exists.  I was curious about KNCminer as it is one of two 28nm ASIC producers at the moment ActiveMining (a.k.a VMC)  is going with a Standard 28nm design.  It'll be interesting who hits the market first and who hits their power and hashing targets.  Both are using eASIC too which really evens the odds.

It'll be fun watching in September!

Firstly dunno anything about VMC but KnC's not using easic, secondly hashfast is claiming to use 28nm in their product.