Cross posted from newbies with permission of OP (not quoted for readability, but I'll edit if this is confusing)
https://bitcointalksearch.org/topic/about-kncminer-227625OP is SwedErik
https://bitcointalksearch.org/user/swederik-127256Yes, please crosspost... I still have to be logged in for two more hours...
There is quite a few things I read there that I could/would like to comment on, but it's not possible now..
Found the threads about KNCminer on this forum and since I know a little about this company. Well actually the formal company "kncminer" is soo new so the one I know about is orsoc, which is
one of it's two parent companies. The other parent company that is owned by two guys do I not know much about, but they does not seem to be have with anything similar in the past, but has worked in the financial sector...
I created an account and will post what I know. I saw a good summary posted by someone who attended the "openday":
There's just not much info to tell.
Are ORSoC part of KnCMiner? Yes they are.
Is ORSoC a legit company? Yes it is.
Is ORSoC's area of expertise desgning FPGAs, ASICs and embedded systems? Yes it is.
Is there a working FPGA prototype hashing at 6.2GH/s? Yes there is.
Is there an ASIC prototype? No there isn't.
What's the plan? To make the ASIC from the FPGA, order the chips with preorder money, and send them directly to assembly.
Are there any gerbers for the ASIC PCB's, etc. that can be showed to us? No, they say they ASIC PCBs will be just based on the FPGA PCBs.
Is this feasible? Mmmm, yeah, why not. That was BFL's plan too - right?
Is this extremely expensive, knowing they are planning to do 28nm? Yes it is. VERY expensive.
Is this project at least x10 bigger in terms of revenue streams, etc. compared to anything else ORSoC has done before? Yes it is. It's a huge project for a tiny company.
I can just confirm everything above and start with the "posstive" things. I do not know the relation between ORSoC or KnCMiner, but I can confirm that ORSoC has been around for quite some time and has made profit in the last 3-5 years. In range of $80k-$250k. They do various things but profile themselfe as fpga consultants. I honestly do not belive that they are trying to make an obvious scam (like; scam people and flee the country), but I would NEVER pay a single dollar in advance. See below for reasons;
They have a history with the "openrisc" cpu that they have been promoting for many years. People who are paying them for a-yet-to-be-produced asic should know that this is not the first time they are collecting money in advance for an asic design. Previous try was with a promise to produce an openrisc cpu (at a MUCH less advance process than .28), they gathered some money from the community (the "campaign" may still be active, I do not know, google for it!). However, when the project did not take off, I think most of the money just went into their company...
As far as I know there was nu public records for the accounting, allthought it was "pitched" as a "community project".
Among other local companies within the same sector, they are quite well known for being "close to impossible to work with" and quite a few companies has bad relations with them. Ask around!!!
My personal belief is some of the people have little touch with reality and tends to be better on producing visions rather than code/hw! That's why I said above that I do not thing they are doing an "obvious scam", but I fear that the result may be the same.
Among the five people listed I think would say that four of them are 100% "non-technical". I guess others are also working in producing working code. To my knowledge no one has ever been involved in asic production before (I could be wrong on this) and they are now about to go directly to a 28nm tape out... . Yeah right!
But since there are quite a few "non-technical" people on the project (at least relativly) I'm not suprised that they do a good job in marketing and come out with new ideas about lotto and other stuff instead of actually showing working stuff.
Or is it 100% confirmed that the mars prototype is actually is working a specified speed??! If soo, why is it not up for sale?
Some other thoughts and things I would think is an absolute must before even going public with something like this:
* I read about canceling mars.... well, if you can not even meet deadlines or results on the fpga, why even bother moving along to an asic? It's a bit contradictive when they claim they basically will use the same pcb for asic as fpga, but they need to focus on asic. To my knowledge it more or less standard to verify that the design is working in fpga (at much lower clock/density) before moving to asic. Why not release this then if it's so magical... Even if the fpga is working PERFECTLY it's a LONG way before doing a tape out (escpecially at 28nm)...
* At least a few people with experince of asic production.
* SOME budget that at least shows that there is a possibility that they can handle a production run at 28nm.... Are you guys aware of the HUGE costs involved??! My guess would be several millions of USD.... however, I have not either done anything similar so it's a guess, but should at least give SOME indication.
* + Some financial backing for doing a tape out at 28nm.... not just relying on massive preorders.
* + some proof that the company and owners really risk their own money, not money collected from users preorders
* It's not that many FAB:s that do 28nm, is it public who they will use? If not, if this is legit, why should it be keept secret? All chip manufactures are quite open about who produce their devices.
Well, these are just my thought, based on what I know and heard from others in the industry. Sure it would be cool if they could produce such a device, but I highly doubt it based on what I know and have read about kncminer. If it sounds to good to be true, it is probably to good to be true.
