Bitfury: "16nm... sales to public start shortly"

[sidehack]

From what I understand from Bicknellski's group, they had one engineer who blew through about $20k in materials designing something that was basically impossible to actually manufacture or even prototype, and everyone else sat around and talked for months. I forget how many hundred or thousand hours of archived design discussions they had.

Just throwing this out there, and it did take quite a bit longer than expected in part due to budgeting problems, but I started from a bare BM1384 and got the Compac up to production-ready with a materials budget of about $1000. Some of us are used to working for free.

If someone had accurate chip info (pinouts and protocols) a basic miner shouldn't be that hard to cook up. With a couple grand in budget for rapid prototyping it gets better. I'm not saying this is a great idea, but we've talked with the WASP guys and they pretty much ran things the opposite of how I would.

[Mitak]

What is this WASP thing

[smracer]

Hi punin, thanks for your engagement here.

Just to be totally clear, you are saying that, at this point in time, there are no machines anywhere hashing with your new chip?

I can't help noticing that your share of total hash power increased sharply recently too.

That is correct. We have just a handful of chips right now and all are used for R&D and profiling including occasional rapid unscheduled disassembly as we look for the limits of the chip . We are expecting the engineering lots in a few weeks time. Luckily our schedule was not affected by the recent Taiwan earthquake.

Our total hash power has increased due to our Gldani immersion cooling datacenter getting to full capacity.

So the immersion facility is using 28nm ATM?  Ugh.

[Prelude]

Hi punin, thanks for your engagement here.

Just to be totally clear, you are saying that, at this point in time, there are no machines anywhere hashing with your new chip?

I can't help noticing that your share of total hash power increased sharply recently too.

That is correct. We have just a handful of chips right now and all are used for R&D and profiling including occasional rapid unscheduled disassembly as we look for the limits of the chip . We are expecting the engineering lots in a few weeks time. Luckily our schedule was not affected by the recent Taiwan earthquake.

Our total hash power has increased due to our Gldani immersion cooling datacenter getting to full capacity.

So the immersion facility is using 28nm ATM?  Ugh.

Going to hurt like hell when 16nm chips replace all of those 28nm. Yikes.

[dogie]

From what I understand from Bicknellski's group, they had one engineer who blew through about $20k in materials designing something that was basically impossible to actually manufacture or even prototype, and everyone else sat around and talked for months. I forget how many hundred or thousand hours of archived design discussions they had.

Just throwing this out there, and it did take quite a bit longer than expected in part due to budgeting problems, but I started from a bare BM1384 and got the Compac up to production-ready with a materials budget of about $1000. Some of us are used to working for free.

If someone had accurate chip info (pinouts and protocols) a basic miner shouldn't be that hard to cook up. With a couple grand in budget for rapid prototyping it gets better. I'm not saying this is a great idea, but we've talked with the WASP guys and they pretty much ran things the opposite of how I would.

I only swooped past them once briefly, but got caught in some reply all email string. There were a few engineers each working on individual aspects, with most of them being paid solely in shares (that turned out to be worthless). They tried making reference boards for maybe 4 chips at once in entirely unusuable form factors just to 'prove' something. Then they were planning on making proper boards to actually use - it was very confusing. Yeah then about 10 people talking and having stupid meetings while never actually having any idea what they were doing, how to sell anything, what they were selling etc.

Not unexpectedly the engineers soon revolted for getting paid nothing with no end in sight and demanded others buy their shares off them. Bick was really the only one putting money in and when he stopped it crumbled and everyone left. That's why he lodged some weird crusade against me - he was strangely bitter that I hadn't joined and that I told them they were going about it wrong, and that I didn't invest (and lose) alongside him. He also generally hated Bitcoin after that because it was the industry's fault he lost out :/.

I don't know, some people just aren't cut out to win.

[sidehack]

Yep. If Novak and I had even a third of their budget we'd have been selling BM1384 full-scale miners five or six months ago. What we worked on was a basic flexible concept where the power and controls could be reused with very few changes for pretty much any chip (within certain practical boundaries, of course, which kinda rules out Spondoolies anything) and using standard bus for everything for compatibility. Complexity just gets in the way. And designing a board around massive digital multiphase controllers with 0402 parts worked out well for the Habanero guys but I'm certainly not gonna attempt it. My pick-and-place isn't comfortable with stuff that small.

[Mitak]

Just out of interest what PnP you are using
I have one of these in the shop for prototypes and short runs

http://www.elinform.ru/data/news/Image/2010/pubimage283_2.jpg
It is with linear drives
and up to 0402 works just fine

[sidehack]

Neoden TM245P - not great, but decent, affordable and gets the job done relatively fast.

[cryptotore]

Hi punin, thanks for your engagement here.

Just to be totally clear, you are saying that, at this point in time, there are no machines anywhere hashing with your new chip?

I can't help noticing that your share of total hash power increased sharply recently too.

That is correct. We have just a handful of chips right now and all are used for R&D and profiling including occasional rapid unscheduled disassembly as we look for the limits of the chip . We are expecting the engineering lots in a few weeks time. Luckily our schedule was not affected by the recent Taiwan earthquake.

Our total hash power has increased due to our Gldani immersion cooling datacenter getting to full capacity.

Thank you for being active in the community punin, it's appreciated.

I got a small question to ask, are you using Novec 7000 fluid in your immersion tanks?
If yes, then awesome.
If-else, then what?

Thanks again.

[jstefanop]

Yea they either use 7000 or 7100. Depending on what temps they want to run the chips at.

Here is my "testing" immersion tank running 7100  https://www.youtube.com/watch?v=p5-BaH8_SAk

[cryptotore]

Yea they either use 7000 or 7100. Depending on what temps they want to run the chips at.

Here is my "testing" immersion tank running 7100  https://www.youtube.com/watch?v=p5-BaH8_SAk

! Thats exactly what I want to make!  But.. 3M won't sell me the fluid! Can you hook me up with Novec 7000?

How are your temps with the 7100? It boils at 61C, so I guess they are high?
Wouldn't 7000 be better for high powered ASIC chips, as it boils around 32-34C, if I remember correctly?

[kilo17]

Hi punin, thanks for your engagement here.

Just to be totally clear, you are saying that, at this point in time, there are no machines anywhere hashing with your new chip?

I can't help noticing that your share of total hash power increased sharply recently too.

That is correct. We have just a handful of chips right now and all are used for R&D and profiling including occasional rapid unscheduled disassembly as we look for the limits of the chip . We are expecting the engineering lots in a few weeks time. Luckily our schedule was not affected by the recent Taiwan earthquake.

Our total hash power has increased due to our Gldani immersion cooling datacenter getting to full capacity.

So my information was correct 6 pages ago that you do not have a reference board built yet.  

https://bitcointalk.org/index.php?topic=1291890.msg13810862#msg13810862

I would think not having the engineering samples "for a few weeks" that an end of March product release for home miners is an aggressive (and likely impossible) task.   Having said that, I think if you haven't ran a full batch of chips then you are expecting everything to go without a hitch with Batch 1 chips.  I think that is also a very ambitious expectation for a new 16nm chip.

I am going to go out on a limb and say end of April or beginning of May is a better bet.

[el_rlee]

Let's say you are using Novec 7000, and have a total power outage... in summer... at noon.... what happens?

I guess they are using something that boils a little higher, afaik 3M can also manufacture the boiling point to your needs. Magnificent stuff.

[NotFuzzyWarm]

Yea they either use 7000 or 7100. Depending on what temps they want to run the chips at.

Here is my "testing" immersion tank running 7100   https://www.youtube.com/watch?v=p5-BaH8_SAk

! Thats exactly what I want to make!  But.. 3M won't sell me the fluid! Can you hook me up with Novec 7000?

How are your temps with the 7100? It boils at 61C, so I guess they are high?
Wouldn't 7000 be better for high powered ASIC chips, as it boils around 32-34C, if I remember correctly?

Won't sell you any? Then I assume you are not a registered company. Most industrial use fluids/chemicals/etc. will not be sold to the general public unless the manufacturer specifically allows it. Stuff like this is not only expensive but in many ways a hazardous substance so they need some assurance that if something happens they are not the 1st target in a cleanup cost liability suit.

[baazju]

Hi punin, thanks for your engagement here.

Just to be totally clear, you are saying that, at this point in time, there are no machines anywhere hashing with your new chip?

I can't help noticing that your share of total hash power increased sharply recently too.

That is correct. We have just a handful of chips right now and all are used for R&D and profiling including occasional rapid unscheduled disassembly as we look for the limits of the chip . We are expecting the engineering lots in a few weeks time. Luckily our schedule was not affected by the recent Taiwan earthquake.

Our total hash power has increased due to our Gldani immersion cooling datacenter getting to full capacity.

So my information was correct 6 pages ago that you do not have a reference board built yet.  

https://bitcointalk.org/index.php?topic=1291890.msg13810862#msg13810862

I would think not having the engineering samples "for a few weeks" that an end of March product release for home miners is an aggressive (and likely impossible) task.   Having said that, I think if you haven't ran a full batch of chips then you are expecting everything to go without a hitch with Batch 1 chips.  I think that is also a very ambitious expectation for a new 16nm chip.

I am going to go out on a limb and say end of April or beginning of May is a better bet.

OR it is all a load of crap and nothing can really be trusted. I think everyone knows by now that it is quite a rare thing to have anyone give the information straight, without a twist or flat out lie to help further their own agenda.

[punin]

Thank you for being active in the community punin, it's appreciated.

I got a small question to ask, are you using Novec 7000 fluid in your immersion tanks?
If yes, then awesome.
If-else, then what?

Thanks again.

Yeah it's Novec 7xxx.

[punin]

Yea they either use 7000 or 7100. Depending on what temps they want to run the chips at.

Here is my "testing" immersion tank running 7100  https://www.youtube.com/watch?v=p5-BaH8_SAk

Awesome setup!

[punin]

Hi punin, thanks for your engagement here.

Just to be totally clear, you are saying that, at this point in time, there are no machines anywhere hashing with your new chip?

I can't help noticing that your share of total hash power increased sharply recently too.

That is correct. We have just a handful of chips right now and all are used for R&D and profiling including occasional rapid unscheduled disassembly as we look for the limits of the chip . We are expecting the engineering lots in a few weeks time. Luckily our schedule was not affected by the recent Taiwan earthquake.

Our total hash power has increased due to our Gldani immersion cooling datacenter getting to full capacity.

So my information was correct 6 pages ago that you do not have a reference board built yet.  

https://bitcointalk.org/index.php?topic=1291890.msg13810862#msg13810862

I would think not having the engineering samples "for a few weeks" that an end of March product release for home miners is an aggressive (and likely impossible) task.   Having said that, I think if you haven't ran a full batch of chips then you are expecting everything to go without a hitch with Batch 1 chips.  I think that is also a very ambitious expectation for a new 16nm chip.

I am going to go out on a limb and say end of April or beginning of May is a better bet.

Will you take 1 BTC bet?

[punin]

OR it is all a load of crap and nothing can really be trusted. I think everyone knows by now that it is quite a rare thing to have anyone give the information straight, without a twist or flat out lie to help further their own agenda.

You might want to check out our 28nm designs that we just released under CC-BY-SA license.

[Mitak]

any info on 28 nm chip that is used in this design