Looking for system integrators for new asic

[RoadStress]

thank you helveticoin for providing EU ASICs. can't wait to be your customer

[repentance]

I've always hoped that there were a couple of organisations developing ASICs under the radar.  Let's hope this one is legitimate.  Sounds like they're doing what Avalon originally hoped to do.

[repentance]

The most important metric in current market is not power efficiency (Hash/J), but simply price per Hash/s. Sounds like you are far behind the competitors.
Another detail to note is that your professionally written call for partners is posted in a fucking internet forum. Any group of graduate students in Switzerland would be a better place to look for the solution you need, not to mention simple advertisement on a professional Website or in a journal, or even LinkedIn. I suspect this has got something to do with users of this forum having a proven competitive advantage compared to other groups I listed when it comes to sending advance payments to random story-tellers.

I think you've misunderstood what they're looking for.  They're just making chips and won't be supplying anything to end users.  They're looking for companies who want to get into the ASIC market without having to develop their own chips - how each company utilises those chips will determine which company's products offer the best value for end users.  There'll basically be 3 or 4 companies designing their product around the same chip - just like many companies use Intel's chips for their products.

The advantage to the companies is that they're starting off with a 28 nm chip which will be improved upon over time and not having to pay the chip development costs.  Smaller chips are one of the things which will inevitably happen with other ASIC developers, but not until they've recovered their R&D costs on their first generation and made a considerable profit to boot.  

It's unlikely this is going to be the only B2B chip producer to enter the ASIC market.  If these guys are for real, it signals the beginning of real competition in the ASIC market.

[2112]

Well, since the helveticoin user is tight lipped I had to search the STmicro press releases:

Semiconductor technology leaders ST, Soitec and CMP help universities, research labs and companies prototype next generation of Systems-on-Chip

The CMP multi-project wafer service allows organizations to obtain small quantities--typically from a few dozens to a few thousand units--of advanced ICs. The cost of the 28nm FD-SOI CMOS process has been fixed to 18,000 €/mm2, with a minimum of 1mm2.

http://www.st.com/web/en/press/en/t3343

Silicon-verified process technology delivers 30% higher speed and up to 50% improvement in power

Measurements on a multi-core subsystem in an ST-Ericsson NovaThor ModAp platform, with a maximum frequency exceeding 2.5Ghz and delivering 800 MHz at 0.6V, are confirming expectations and demonstrating the great flexibility of the technology and the extended voltage range exploitable through DVFS (Dynamic Voltage and Frequency Scaling).

http://www.st.com/web/en/press/en/t3370

Pretty soon anyone in EU-landia could probably order this process through Europractice, the same way as bitfury did.

[Gomeler]

Get in touch with Yohan, a rep from Enterpoint. It sounds like they are developing a second generation device in-house but perhaps, if what you have is competitive in price and performance, they could integrate your solution into a product for the market.

[Magnate]

I hope this gets off the ground-should be very interesting for the BTC ASIC market- numerous vendors of a powerful chip offering different features and price points to compete for market.

Reminds me of how the mother board market works more so than the video card market. You can have $200 mother boards using the same bridge chips as a $500 mother board, just the final features and implementation varies. But at each price point there is competition for market.

Will be watching for this with interest

[repentance]

I hope this gets off the ground-should be very interesting for the BTC ASIC market- numerous vendors of a powerful chip offering different features and price points to compete for market.

It would be a great thing from an end-user point of view, given that BFL is really the only company whose objective from the outset was to develop an ASIC unit for retail sale.  Avalon originally intended to manufacture chips and not complete units and ASICMiner intended to make their boards available for sale rather than complete units.  One or more companies wholesaling chips alone will likely lead to more options for those seeking to buy complete units at a retail level as well as for those who want to buy/develop complete boards only.

That someone would enter this market is something current ASIC vendors should definitely have foreseen, and it will be interesting to see how their business models accommodate an increasing amount of competition so soon after launch.

[Magnate]

This method could be very price competitive.

Take BFL. 8 chips per board, and those chips are a few $$s each to produce, but the development costs have to be spread over the units only BFL sells.

Avalong uses ALOT of slightly less expensive chips, but their cost to produce all those chips and boards etc would likely actually cost them more than BFL. Not to mention their volumes will be smaller- they just aren't set up for really high volumes of assembly it seems, and the power consumption will cost them sales.

For these new chips they will cost more than BFL's-say 5 times more for the integrator once you have paid costs and share of dev. But *if* you only need 2 of them for the same performance you have a small on cost to the integrator compared to BFL-but BFL still have to pay their development costs.

Look at the cost of a motherboard. Most good boards cost $200-$300AUD, and the fancier ones cost up to $500. Now consider all the bridge chips, controllers, LAN chips, audio chips, capacitors etc etc, there is a lot that goes on them. Since there are so many board models the only thing really made in huge volumes are the chips themselves which are used by many different vendors, the cost to develop each board is spread over smaller numbers.

Could this result in a $500 low power high hash ASIC? Possibly if enough integrators come along and compete with each other. It may not be possible for BFL to compete at that price, and I'd say Avalon will be done by the time those prices come around.
Could BFL become the 2nd 28Nm ASIC chip supplier so we get an Intel/AMD battle? Intel still makes their own boards as well. If BFL goes this path for their 2nd gen ASIC then mining really will stay affordable.

Still looking forward to Q3/Q4 to see these new ASIC integrators show their stuff. I can wait til then to part with my $$.

[repentance]

This method could be very price competitive.

Take BFL. 8 chips per board, and those chips are a few $$s each to produce, but the development costs have to be spread over the units only BFL sells.

Avalong uses ALOT of slightly less expensive chips, but their cost to produce all those chips and boards etc would likely actually cost them more than BFL. Not to mention their volumes will be smaller- they just aren't set up for really high volumes of assembly it seems, and the power consumption will cost them sales.

For these new chips they will cost more than BFL's-say 5 times more for the integrator once you have paid costs and share of dev. But *if* you only need 2 of them for the same performance you have a small on cost to the integrator compared to BFL-but BFL still have to pay their development costs.

Look at the cost of a motherboard. Most good boards cost $200-$300AUD, and the fancier ones cost up to $500. Now consider all the bridge chips, controllers, LAN chips, audio chips, capacitors etc etc, there is a lot that goes on them. Since there are so many board models the only thing really made in huge volumes are the chips themselves which are used by many different vendors, the cost to develop each board is spread over smaller numbers.

Could this result in a $500 low power high hash ASIC? Possibly if enough integrators come along and compete with each other. It may not be possible for BFL to compete at that price, and I'd say Avalon will be done by the time those prices come around.
Could BFL become the 2nd 28Nm ASIC chip supplier so we get an Intel/AMD battle? Intel still makes their own boards as well. If BFL goes this path for their 2nd gen ASIC then mining really will stay affordable.

Still looking forward to Q3/Q4 to see these new ASIC integrators show their stuff. I can wait til then to part with my $$.

I wonder whether BFL even considered this possibility in their risk management assessments.  This might not even be the only company which has developed chips which are near ready for release to the B2B market.  For all we know, there's another company out there we haven't yet heard of which has end user units almost ready for release, too.

This could force BFL to go to a smaller process sooner than they'd originally planned, or to drop the prices of their current units earlier than they'd hoped.  It's definitely going to be interesting as more and more players enter the ASIC market.

[kaerf]

3 2 1 ... que kano.

cgminer support for hardware.

[Puppet]

Still looking forward to Q3/Q4 to see these new ASIC integrators show their stuff. I can wait til then to part with my $$.

Q4? Delays are always possible, but dont be indoctrinated by BFL. Keep in mind these guys claim to already have chips, in that sense they are way ahead of BFL. I dont know how long it takes to design and manufacture a custom PCB for something like this, but certainly based on a ref design, those integrators should be ready when these guys get their chips. They will have had some time to have tested the boards and software using the preproduction chips. So if they do get the chips in june, why couldnt they ship in june?  BFL seems to think they can ship mere days after they get their first ever chips.

What I am curious about is when we will see those vendors announce their product and start accepting (pre)orders. BFL has shown the advantage to them of announcing early and since we might see multiple competing vendors with very comparable products soon, I dont think it will be very long.

[Micon]

Wow if this turns out to be real, and I'm calling this one early (with a lower degree of confidence 4 posts in) that OP seems extremely real from the onset.  This is how a good corporate citizen would act after building their first prototypes and testing the process. 

This seems like a no-bullshit business that has already successfully completed the R&D phase (unlike some guys out there...) and are taking the right steps to bring the chip to market. 

[repentance]

What I am curious about is when we will see those vendors announce their product and start accepting (pre)orders. BFL has shown the advantage to them of announcing early and since we might see multiple competing vendors with very comparable products soon, I dont think it will be very long.

Hopefully they've learned something from the other vendors and will unveil their product when they already have their supply chain in order and can offer a rapid turnaround time on orders.

Someone has to break the whole pre-order thing sooner or later and start offering ASICs with less than a week's turnaround time - it might as well be one of the new players because BFL is so hopelessly backlogged now that the only way they're going to be able to offer a reasonable turnaround time on orders this year is if they stop taking new orders for a while, and Avalon is likely to continue producing their ASICs in small batches for the foreseeable future.

[Magnate]

Q4? Delays are always possible, but dont be indoctrinated by BFL. Keep in mind these guys claim to already have chips, in that sense they are way ahead of BFL. I dont know how long it takes to design and manufacture a custom PCB for something like this, but certainly based on a ref design, those integrators should be ready when these guys get their chips. They will have had some time to have tested the boards and software using the preproduction chips. So if they do get the chips in june, why couldnt they ship in june?

Yeah I agree. *IF* there is a reference design and firmware/mining software in place then integrators with their own production equipment could have tested systems ready in Q3. If there is no reference, or different integrators want to do different implementations, might take longer.
Plus hey doesn't hurt to be pessimistic when thinking about when you are going to get some new mining hardware 

[mistfpga]

Great thread. 

email sent, looking forward to seeing how this progresses.

[MrTeal]

Q4? Delays are always possible, but dont be indoctrinated by BFL. Keep in mind these guys claim to already have chips, in that sense they are way ahead of BFL. I dont know how long it takes to design and manufacture a custom PCB for something like this, but certainly based on a ref design, those integrators should be ready when these guys get their chips. They will have had some time to have tested the boards and software using the preproduction chips. So if they do get the chips in june, why couldnt they ship in june?

Yeah I agree. *IF* there is a reference design and firmware/mining software in place then integrators with their own production equipment could have tested systems ready in Q3. If there is no reference, or different integrators want to do different implementations, might take longer.
Plus hey doesn't hurt to be pessimistic when thinking about when you are going to get some new mining hardware  

If the silicon is ready and well documented, I wouldn't be surprised to see boards available shortly after the first production runs get out of fab and packaging. Creating a nicely packaged product like BFL is going to be a relatively longer timeframe, but developing a board with several hashing chips, off the shelf POL modules for power and a small MCU or the included Cortex for USB comms in the style of the MMQ is something a competent PCB designer could do in a weekend. Even the firmware shouldn't take long, and miner support should be relatively quick. More optimized designs could come out later, but there's not reason that hashing units couldn't be in customer's hands no later than a month after chips are ready to be shipped in volume, and that's with a buffer for testing and needing to do a quick respin of the PCB.

[firefop]

This method could be very price competitive.

Take BFL. 8 chips per board, and those chips are a few $$s each to produce, but the development costs have to be spread over the units only BFL sells.

Avalong uses ALOT of slightly less expensive chips, but their cost to produce all those chips and boards etc would likely actually cost them more than BFL. Not to mention their volumes will be smaller- they just aren't set up for really high volumes of assembly it seems, and the power consumption will cost them sales.

For these new chips they will cost more than BFL's-say 5 times more for the integrator once you have paid costs and share of dev. But *if* you only need 2 of them for the same performance you have a small on cost to the integrator compared to BFL-but BFL still have to pay their development costs.

Look at the cost of a motherboard. Most good boards cost $200-$300AUD, and the fancier ones cost up to $500. Now consider all the bridge chips, controllers, LAN chips, audio chips, capacitors etc etc, there is a lot that goes on them. Since there are so many board models the only thing really made in huge volumes are the chips themselves which are used by many different vendors, the cost to develop each board is spread over smaller numbers.

Could this result in a $500 low power high hash ASIC? Possibly if enough integrators come along and compete with each other. It may not be possible for BFL to compete at that price, and I'd say Avalon will be done by the time those prices come around.
Could BFL become the 2nd 28Nm ASIC chip supplier so we get an Intel/AMD battle? Intel still makes their own boards as well. If BFL goes this path for their 2nd gen ASIC then mining really will stay affordable.

Still looking forward to Q3/Q4 to see these new ASIC integrators show their stuff. I can wait til then to part with my $$.

I wonder whether BFL even considered this possibility in their risk management assessments.  This might not even be the only company which has developed chips which are near ready for release to the B2B market.  For all we know, there's another company out there we haven't yet heard of which has end user units almost ready for release, too.

This could force BFL to go to a smaller process sooner than they'd originally planned, or to drop the prices of their current units earlier than they'd hoped.  It's definitely going to be interesting as more and more players enter the ASIC market.

I doubt it was considered. And it might not make much difference either way. What I mean is, there's really zero downward pressure on prices... assuming that the new offering isn't orders of magnitude better (and a move from 65 to 28 really isn't going to show that much gain imo). You'll dirty the water by having multiple board designers... which maybe help with the prices being lower...

But the long and short of it is... if they don't absolutely destroy BFL as far as power draw and shiny packaging they aren't really competition long term. And even if they are... there's no reason they wouldn't have pricing thats identical or very close to BFLs.

[Magnate]

Mr Teal- I hope your prediction is more correct than mine!

Firefop- from a business prospective you are correct. Why would you charge half the price of the competition when you could charge about the same and make a similar amount of sales. If they have speed and power advantages they could make the claim of "we are better you pay more" just like BFL are doing currently.

I don't think the BTC community is that worried about fancy cases. They/we are used to playing with computer boards and cables, and plenty had FPGAs which had no cases at all. Yes the BFL cases look nice, but I was planning on getting mine without the case to increase airflow- function over form for me, and when it comes to making ROI I'm sure I wouldn't be the only one!!

helveticoin - I'm desperate for some teaser specs off your reference test board. If you want people to hold out and not spend their $$ on BFLs and Avalon before you and your integrators launch then give us a reason to postpone!! What clock speed speed have you reached with your current cooling method, and what power and hash results did you get? Yes it will vary depending on what the integrators do for cooling etc, but it would give us something to ponder when considering a future purchase.

Hmm one chip as fast as other vendors multi chip systems- if an integrator used a standardised heat sink mounting pattern it would be killer for aftermarket cooling and even off the shelf water cooling (please please please!!)

[helveticoin]

update added to OP.

A lot of people are sending requests through the forum message system, and some are requesting our email.
The email is stated in the first post: helveticoin at. gmail dot com and I would greatly prefer you use that because I have to fill out captcha's to reply to forum messages.

I understand there is a lot of interest and speculation about our time to market. As a principle, we do not want to pre-announce our customers announcements. However, bitcoin asic market is somewhat unique, and gives a huge and unfair advantage to companies announcing early even if they deliver late. Since we want your suppliers to compete on service and technical merits, rather than on the date they announce or start accepting preorders, we will synchronize their product announcements. A date for this is yet to be set, but I hope doing this in May will also give our new partners sufficient time to prepare and we should be sure of our ability to deliver by then. If our partners agree, we will put out a common statement much earlier giving end users a better idea of what to expect and when to expect it.  This could happen in a few weeks time.

[invader]

It maybe sounds funny, but did you ever considered supplying a limited amount of chips to those individuals who can make their own PCB and want to use them in DIY miner with ARM-core ? no ASIC manufacturers doing this yet, you can be first.
I think it's a great possibility to create a miner-on-a-chip with some general purpose interfaces for extended use.
It's a good idea to make miner core as a secondary option, as even when mining not actual any more, such board can still be used for other purposes in a manner like raspberry-pi.