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

[Bitcoinorama]

good read.

and your idea about AYBABTO is totally la
me.

Wha?!...and after I complimented you on your choice of signature the other day...!

 

That was totally 'cause *I* have good taste

In this instance we have an agreement:

http://www.planettribes.com/allyourbase/index.shtml

[J35st3r]

Granted... at least voltage tests and i/o tests seem simple enough...  but on the other hand, if they're capable of excluding defective engines through an automated process onboard - then there's really little reason to pre-test anything.

Well, there is the matter of an expensive test socket/interface to be designed and procured. And did you note they have not even selected their foundry yet. In another post I noted the over-optimism of ASIC customers (ie KNCMiner in this context, as the customer of the foundry), that was 20 years ago. Seems nothing has changed. There will be tears.

PS I edited for clarity, hence the difference from firefop's quote below.

[firefop]

Granted... at least voltage tests and i/o tests seem simple enough...  but on the other hand, if they're capable of excluding defective engines through an automated process onboard - then there's really little reason to pre-test anything.

Well, there is the mater of an expensive test socket/interface to be designed and procured. And did you note they have not even selected their foundry yet. In another post I noted the over-optimism of ASIC customers (ie KNCMiner in this context vs foundries), that was 20 years ago. Seems nothing has changed. There will be tears.

I'm inclined to agree... but you know every once in awhile someone wins 64btc playing satoshidice. So I guess we'll just wait and see.

[Bitcoinorama]

Granted... at least voltage tests and i/o tests seem simple enough...  but on the other hand, if they're capable of excluding defective engines through an automated process onboard - then there's really little reason to pre-test anything.

Well, there is the mater of an expensive test socket/interface to be designed and procured. And did you note they have not even selected their foundry yet. In another post I noted the over-optimism of ASIC customers (ie KNCMiner in this context vs foundries), that was 20 years ago. Seems nothing has changed. There will be tears.

I believe they have. I think this is mentioned later in the audio.

There are two competing foundries, we were told who they were, both have the final design, both are bidding for the gig. Sam mentioned something of huge interest here in front of us. So i'm not the only one who heard this, but I'm not sure I can repeat it; so I won't, but it involves meeting the September deadline.

The decision was to made on Tuesday/Wednesday to place the order with the chosen foundry I believe was mentioned. I attended Monday.

I'm also only through the first page of questions.

[KS]

so, did you pay for a jupiter/saturn in the end? you don't seem to want to answer the question publicly.

I'll keep on riding you for that, not that I particularly enjoy it, if you keep on painting KNC in an especially good way yet don't invest yourself.

[Bitcoinorama]

so, did you pay for a jupiter/saturn in the end? you don't seem to want to answer the question publicly.

I'll keep on riding you for that, not that I particularly enjoy it, if you keep on painting KNC in an especially good way yet don't invest yourself.

Yes, I answered that in the first post on this thread!

[suryc]

Is anyone clear on what the physical size of these going to be?
There site says a Jupiter is "160 x 160 x 200", but doesn't specify units. If that is cm, which I would assume it is, that is really massive.

[J35st3r]

There are two competing foundries, we were told who they were, both have the final design, both are bidding for the gig. Sam mentioned something of huge interest here in front of us. So i'm not the only one who heard this, but I'm not sure I can repeat it; so I won't, but it involves meeting the September deadline.

Yes, but the point I am making is that this s very late in the game to be selecting your foundry. OK, there has been a lot of progress since I left the industry 20 years back, but this is a very tight schecdule. Perhaps I should quote again the article linked by KS in the other thread  http://www.edn.com/design/systems-design/4319891/The-economics-of-structured-and-standard-cell-ASIC-designs (2006 vintage but still relevant).

I wish some of the current ASIC design gurus would comment on this (several very informed guys post on these threads).

[Bitcoinorama]

There are two competing foundries, we were told who they were, both have the final design, both are bidding for the gig. Sam mentioned something of huge interest here in front of us. So i'm not the only one who heard this, but I'm not sure I can repeat it; so I won't, but it involves meeting the September deadline.

Yes, but the point I am making is that this s very late in the game to be selecting your foundry. OK, there has been a lot of progress since I left the industry 20 years back, but this is a very tight schecdule. Perhaps I should quote again the article linked by KS in the other thread  http://www.edn.com/design/systems-design/4319891/The-economics-of-structured-and-standard-cell-ASIC-designs (2006 vintage but still relevant).

I wish some of the current ASIC design gurus would comment on this (several very informed guys post on these threads).

But ORSoC have a strong relationship with both foundries, it's not like they are selecting people they haven't worked with. They are negotiating delivery for Sept.

They were to resell the FPGAs in Mars at an amazing price because of a direct relationship with ALTERA. The deal was though the chips could not be harvested and resold under the agreement. Customers could in theory if they could release the chips from the board, but that's why KnC couldn't offer a trade-in deal with the FPGAs for their scrypt miners to be made at a later date.

[J35st3r]

But ORSoC have a strong relationship with both foundries, it's not like they are selecting people they haven't worked with. They are negotiating delivery for Sept.

Yeah, good point. My experience was with new entrants with wide-eyed expectations. I'm still rather worried about the lack of a test strategy. But in my day yields of 30-60% were common, if they can manage 95% today, them maybe the build it and pray approach will work (it certainly seems to be what BFL used). Read the artice I linked. It shows the sort of timescales the professionals work to.

[KS]

so, did you pay for a jupiter/saturn in the end? you don't seem to want to answer the question publicly.

I'll keep on riding you for that, not that I particularly enjoy it, if you keep on painting KNC in an especially good way yet don't invest yourself.

Yes, I answered that in the first post on this thread!

Yeah, but "paid for" is the keyword I'm looking for. You already had a Jupiter (or 2) and a Saturn on order before.

[KS]

but that's why KnC couldn't offer a trade-in deal with the FPGAs for their scrypt miners to be made at a later date.

So you're saying they scrapped the Mars because of a trade-in deal they couldn't do for a non-existing miner??

I'm pulling your leg, but there is still a lot of BS coming from your source. Either you're drunk in the coolaid or you're genuinely deaf to it.

[Syke]

Marcus: The die size will be...very large.

That sounds like trouble waiting to happen.

[DPoS]

how about:

I drink your bitshake

http://www.youtube.com/watch?v=RKQ3LXHKB34


or combine them

All your bitshake are belong to I  (or  us)

[rizzman]

Would it be; "All your Bitcoin are belong to us", or "All your Bitcoins are belong to us"?

It would be the singular, right? As 'base' was never plural...

Haha! I would gladly pay an extra $25 bucks for that... would be awesome!

[erschiessen]

how about:

I drink your bitshake

http://www.youtube.com/watch?v=RKQ3LXHKB34


or combine them

All your bitshake are belong to I  (or  us)

There's no cow on the ice

[erschiessen]

Good work thus far, bitcoinorama

OTHER ATTENDEES:

Please feel free to chime in, more voices are needed in the choir!

[RoadStress]

Marcus: The die size will be...very large.

That sounds like trouble waiting to happen.

Why is that?

[Syke]

Marcus: The die size will be...very large.

That sounds like trouble waiting to happen.

Why is that?

So many things can go wrong. Timing signals/voltage drops across the whole die, low wafer yields, too much heat, etc.

[Bitcoinorama]

Is anyone clear on what the physical size of these going to be?
There site says a Jupiter is "160 x 160 x 200", but doesn't specify units. If that is cm, which I would assume it is, that is really massive.

To me Mars was about 16cm by 16cm, by 20cm, That pcb board they gave me fits a space the exact same dimensions as an iPad screen area.

I know this as I'm positioning it over one now.