Bitcoin Forum

Declaration: Before physical chips are out and fully tested, all specs/dates/data are for reference only and subject to change.

Although with unexpected obstacles, timing rearrangements and delays, our chips are still on track of delivery in March. This thread for releasing information of chips to public, with the intention of getting enough (direct) customers or (indirect) resellers aware before actual chips are produced.

The information will be updated to be more and more accurate, as we are updated with our fab, engineers, as well as the mining market.

Specification
Process node: 40nm
Package type: QFN64 8mmx8mm (with another option of QFN64 7mmx7mm possible)
I/O: Standard SPI protocol with clk, in, out and cs.
Rated Hashrate: 12.8GHash/s per chip, with a wide range of overclock/downclock options
Rated Voltage: 0.72V, recommended voltage range is 0.55V-1V
Power Consumption: 0.2J/GHash low voltage, 0.35J/GHash rated voltage

What to expect with the chips
Datasheets and reference PCB design are going to be released as soon as they are ready.

Chips will be available for ordering once the functionality and yield rate are confirmed with testing on actual chips.

There will be three modes of orders: Immediate delivery, 2-month delivery and 3-month delivery, with the price getting lower one by one. For 2-month and 3-month delivery, premium + balance payment may be supported to enhance liquidity and enable larger orders.

Both USD and Bitcoin payments are supported, but the nominal price will be based on USD.

Full devices from ASICMiner are definitely going to exist on the form of samples and small batches. ASICMiner may also build new mining farms based on new chips. Unsliced wafers may also be available for sale. But the main business direction is to provide packaged chips.

Price range
0.49$/G-0.99$/G, depending on order size and delivery speed of choice.

Please bear in mind that when choosing brands of Bitcoin mining chips, the cost of making whole working mining devices, including PSUs, DC/DC, heatsinks, fans, and room cooling facilities, needs to be taken into account. With our relatively low power consumption, the customer can save more money on the devices' side.

Partners
We are especially interested in the following types of business partners:

1. Large mining operation owners who can also do massive PCB production of their own.
2. Mining device manufacturers who can improve our open-source design or make new designs from scratch.
3. Resellers with various channels and abundant liquidity in cash or Bitcoins.

Contact
Please PM to (friedcat) on bitcointalk, or mail to (cat@bitquan.com)/(fnnirvana@gmail.com) for future partnership discussions. Please do not use them for information inquiry, because all info about chips will be published here.

Update March 29, 2014: The chips passed all functionality tests.

Update April 21, 2014: The documents as well as the (verified & produced) sample design are out:

Schematics - https://mega.co.nz/#!uNMkFZoI!lnlRehlynQQzNRfQ87_UYek1RtrOUvLZ6a074XeRqlo

Reference PCB File - https://mega.co.nz/#!DZVHgS6D!6pmTsmrito8rfVJJ-etvznOtUcalbAT6vU-C8EqY0_I

datasheet - https://docs.google.com/document/d/1_dmXzyMqtr7tVZukk2DAJih2sRq5jDy5DQPkwCce570/edit?usp=sharing

bonding list - https://docs.google.com/spreadsheet/ccc?key=0Al1fvFT7Sd5bdFFZSUllNW5seVRfM01kcDVqZkJfdFE&usp=sharing

9x9 package pictures -
https://drive.google.com/file/d/0B11fvFT7Sd5bUV90cFN4cEhTNEk/edit?usp=sharing
https://drive.google.com/file/d/0B11fvFT7Sd5banZkMDZqSTBGTEE/edit?usp=sharing
https://drive.google.com/file/d/0B11fvFT7Sd5bWjh6MmJsMFpCYkU/edit?usp=sharing

8x8 package picture -
https://mega.co.nz/#!DA1mBJzB!7hFnoT8ZYQd1d30m2oHbuscD36cCQ4ou7stsr5Lo03Y

The package we are using is 9mmx9mm, but it might be partly switched to 8mmx8mm.
Also, we plan to increase the center pad size to at least 7mmx7mm, so please leave enough margin on the PCB to avoid unnecessary re-designs.

(Reserved for Updates)

a can of coke soon more pricy than 1GH/s bitcoin firepower  ;D
hopefully this is the 1st step to turning concept of pre-order and customer financing into a distant memory, and eradicating these telemarketers hashfast, cointerra, bfl etc

working B2B can provide immense value to private mining ops with plentiful supply of low cost, efficient chips with short lead time.
reliable partnerships can be built benefitting all. as well as to DIY miners of various sizes. BTC mining ecosystem will get much needed boost.

Had sent PM

Thank you for the update.

hi

i m amazed by the wattage

i m following this and contacting you

thanks

The cat is out of his den pan and more powerful than ever! Yes!!! 8)

Wattage is impresive indeed. But it's still 5W to dissipate from small 8x8 mm package. At least make it flip chip that most of heat can go thru top not bottom and PCB that is poor heat conductor.

EDIT: And please consider possibility of selling small batches of sample chips, like 10 chips.

thank you really great news, good luck.

Congrats sir!

The WPC would definitely be interested in this.

The right move. Cheaper chips.

We hope to discuss possible development of a 40nm Wasp using that chip. We will have a working prototype coming out soon and with our unique modular design and an open source design it is a perfect match. We will be sending you some information soon but we would like to get reference materials and ideally some sample chips asap.

We're definitely interest as well. Will send a PM.

Power Consumption: 0.2J/GHash low voltage, 0.35J/GHash rated voltage

That is insane. 3TH rigs using only 1kw at 40nm. Very, very impressive friedcat, I look forward to seeing them!

Sound like some really nice chips. Cant wait to see them.

Thank you! Now us long term shareholders and believers can sit back and relax. It's been scary to say the least lately, but now we know we made the right decision holding and buying low:) Friedcat is the man! (Or woman)

thank you  ;D

Just came, thanks Fried!

AM's chip design is open-source? They are sharing design with Mining device manufacturers, I'm confused with their business pattern.
I thought AM is partially a Mining device manufacturer itself. Isn't the design a big secret for them?
Who can explain this please?

Excellent!

Very interesting. If you can hit those specs it would be a huge coup.

Pm'ed! Very very impressive.

The open-source part will be the reference PCB, not the chip.

very impressive

Thanks for reply, FC, good job!

cool. I bought cheap stocks before the announcement. stock price will be higher. say if ASICMiner sell 200,000 chips. That will be 1500BTC. Huge gain for shareholder.

1500 BTC is not huge ,15,000 BTC is huge

Great news!

Afaik with flipchip there's still the majority of heat going through the PCB...

Very impressive indeed. Looking forward to get some more mining power

I'm already prepping myself to be disappointed with the initial pricing on anything that comes out of this.. haha

It is still in the air - I hope the price hits home. They seemed to have managed a good pricing scale for their past products.
*edit* one can get a 200GH miner from antminer now for around 1.8btc (1600$). Assuming 40% price decrease a month (avg diff incr. of 30%) that brings it between (560 and 900) for March, That is for a fully assembled machine, save for power supply -  while the same hash through ASICminer would cost around 200$ in chips only....
ASICMiner is listing price for chips only, I wonder what the cost per GH would be for actual miners.
(But it was assumed that they would go into the chip business rather than miner hardware resale)

I is time to buy shares? Stock exchange same as cex.io will be presented?

 ;D

you can buy direct shares at www.havelockinvestments.com

Direct shares? Is this new? O_O

Very Nice, so you are saying that Asicminer will focus on selling chips.
But also Asicminer could make some full miners just to keep in yours mining facilities?

this is not new !! I don't know where is your confusion ? you asked where to buy shares, and I told you where can you get them, otherwise this is not a new IPO, ASICMINER exist for a couple of years now, and proved to be the best business around bitcoin economy...

basically when you buy a share you will be entitled to a weekly dividend that they pay for every outstanding shares from the mining and hardware sale income, you can also trade the shares and speculate on their prices, the price has already jumped 15-18 % since the announcement...

Alright, let me be a little more blunt, shares on havelock aren't direct, though they can be converted to such.

this is what I've understood yes, really huge news, I have a feeling that AM will change the future of mining if they achieve what they aimed for..

well theoretically they are not but technically they are ;)

from havelockinvestments

QFN often has thermal pads on the underside. Connect these with enough vias to an underside copper pour without silkscreen and you've got an effective radiator. Stick a fan over that radiator and you'll suck even more heat away.

That's what I do with my QFNs, but I'm not sure about 5W dissipation. Anyone want to crunch the numbers?

For comparison, Bitfury chips use a similar package, and I figure they dissipate about 2.5W in my overvolted rigs with just high-flow fans and no heatsinks. Bitfury chips use thermal pads to dissipate most of the heat through the PCB.

Replacement boards for the ASICminer Cube?

You can put a lot more than 5W through this size of package depending upon how it's mounted. PCB via's won't give a particularly good thermal path, but it is possible to use a copper slug to connect to the die pad (with a suitable square hole in the PCB)  and then connecting to a larger heatsink. That way over 25W can be removed - assuming die size of 5.3 x 5.3mm - with a large enough sink, the trick is in coupling the die, slug and heatsink together.

Email sent.  Hoping to hear from you soon; this is an exciting announcement.  :-)

What is the frequency of the chips in normal mode?

How cool would that be!?   ;D

This is great news!

Will this price be somehow coupled to the BTC market price?
For example, the price is $0.99 when the BTC price is $1000. Now the BTC price rises to $3000, will the chip price be adjusted to $2.97 or remain at $0.99?
At an increased BTC price the chips would - of course - be worth more and should(?) undergo a correction.

Messaged! Looking forward to seeing the new chips :D

That would be too good to be true  :o

Woo hoo  ;D
Looking forward to these units

Can't wait to throw money at Friedcat for this new hotness !!!

next usb miner coming yay

Great news on the chips Friedcat, would be looking to join the team as a reseller for UK and other places.

Good to see you looking publicly for designers and company's to take on board as I'm sure people on here could come up with some pretty good designs.

That fit well into racks or stack plus PSU built in so just a 3 pin kettle would be the ideal future of bitcoin hardware delivery.

Plug and play would make the big sales - If you can get suppliers for each of the stages you will be onto a winner.

Keep us updated as always Friedcat

CroverNo

Is there a timeframe for this that I missed?

We are interested, mail sent to you. TMC

The world of Bitcoin is amazing.   From somewhere around 15 BTC a gigahash to 15$ USD a gigahash in less than a year, pretty amazing.

Which I had more capital, would love to be a reseller, I think this product will kick ass.  Good job F.C .

Yes - there is a timeframe you missed - from OP: "our chips are still on track of delivery in March"

Very interested in both OEM and resale (especially the unsliced wafers), PM sent.

finally !

Very Impressive FC.
Looking forward to it.

looking forward, hoping to get the first in line of next gen asic chips from FC :)

Woohoo.  :o

Happy new year ,Friedcat!

Are these power estimations for real??  Ridic!  I Think ill grab me a reel or two.

So, are you going to get with the plan and use stratum in your forthcoming miner, or are you still going to be stuck in the dark ages with getwork?

Fantastic news, thanks for the update Friedcat! You should be receiving a PM from me shortly.  :)

interesting! i looking for news! :o

Indeed!

Me too :)

PM and Mail sent! But no answer, please reply.

TMC

Has anyone heard from them yet?  We've sent an email directly, but received no reply.  I don't want to spam them though.  It's likely that they've received overwhelmingly good response, and are just buried, so I was curious if anyone had gotten through.  We want to build things with their chips!  ;D

Same here

+1 on interest

Nice I am in goint to be watching closely...

Hi Friedcat,
Can you sell the sample chip in small batch? we really keen to get the sample in time since Avalon next Gen also release at the same time and the Engineering module already in place. we hop to see your PCB design much earlier.

Thanks

chinese new year.                                                                                                                                                                                                                                       

+1  One of several reasons for not going ASICminer............for me anyhow  ::)

I'd bet they just wait and see how many people are interested and will get back to you when they can make a specific offer. Keep us posted when they answer!!! But it may still take some days/weeks. Maybe until tapeout when they've got their final specs.

Still in this...

Interesting news, and I have to admit I'm rather puzzled why a company that makes no secret of it's mining activities wants to make it's hardware publicly available? Is it altruism or something else? In the commercial world, despite what is said in public, companies do not 'welcome' competition - they hate it and will do almost anything to eliminate it. Anyone who thinks or believes otherwise is an idiot.

I'm also concerned that many, many miners reading this information will only see the headline $/GH figures of '0.5 to 1' and imagine that you're going to be purchasing hardware at anywhere near those figures - you're not, as the chip cost is only a part of total rig cost and unless you're going to be buying the chips and their associated pcb's, psu's and controllers in volume then the figures you more likely to be looking at are $2.0 - $2.5/GH. Yes, I know it's much better that what's available right now, but that will change soon.

Got to hand it to Asicminer though. Nice pitch and nice to see that someone has finally paid some real attention to their sha256 silicon implementation rather than just using the standard blocks and recognised that you don't need 28 or 20nm to make a very efficient hashing chip.

While this is true, see how many times JP Morgan has tried to patent e-currency to capture it through the law, I think most of us who believe in Bitcoin do it for more altruistic reasons. I know I do. Like Tyler Winklevoss recently said in the NY regulatory hearings, "Bitcoin is freedom. It’s very American.”

http://www.zerohedge.com/news/2013-12-15/jpmorgans-bitcoin-alternative-patent-rejected-175-times
http://dealbook.nytimes.com/2014/01/28/in-praise-of-bitcoin-with-little-regard-for-banks/?_php=true&_type=blogs&_r=0

No altruism I can detect. AM is just saying they have the best chip and realises they are chip experts. Why mess with the other stuff where they have no competitive edge. Best to sell chips to maximise ROI. Also AM has learnt that dominating hash is not good for bitcoin, so by distributing their chip widely they are protecting themselves by protecting bitcoin. AM has done the maths and this is a calculated move.

They can also start generating revenue as soon as they get the chips in hand. Just send them out to buyers with a good datasheet containing a reference PCB design and voila! This is what linear, microchip, TI, analog all do.

Nice... At least there is some movement

The cat delivers..  he now owns the floor

Hi Friedcat,

Can you sell the sample chip in small batch? We have factory in Shenzhen, have numerous customers and we are keen to arrange early access so that we can lockdown our reference design.

We are considering Avalon next Gen but we really would like to see your PCB design much earlier.

Thanks

LGV

>>>>>>>>>>>>>>>>>>>>

Same question, what is the smallest purchase size/quantity?

You are asking the wrong question. Read the post:


If you check the boxes on these, fine. If not, you may need to wait for a reseller.

Also, FC asked for PMs or emails, not questions in this thread :)

May be he changed of mind or someone provided already all the required money he needs.

:-)

Well, we might say that if someone can't even follow this simple instruction, it's not someone you want as a partner when huge sums of money are involved...

We are very interested in a partnership going forward.  We run an institutional mine and have quite a bit of capital to invest into a wholesale relationship for chips.

Please consider us. Thanks.

To repeat again:

"FC asked for PMs or emails, not questions in this thread"

For those who haven't seen this yet:
There has been a board meeting and friedcat answered questions we collected and submitted to him. Also a big thank you to Jutarul, who helped us getting them through and inform us about the other topics discussed during the meeting.

https://bitcointalk.org/index.php?topic=99497.msg5025133#msg5025133 (https://bitcointalk.org/index.php?topic=99497.msg5025133#msg5025133)

yeah if a 5gh bitfury chip hashes at 2.6gh in an ice fury or red fury    these should hash at 5gh-6gh

And people who can't even read simple instructions, won't likely be good partners for anything.

Out of curiosity - what makes you think that the bitfury chip should hash at 5-6GH?

it was his initial plan but it seems his chips can go more than 2.5 ghas,you can see also that he has printed 5ghs on his first packaged chip

there is some pictures available here an there

First post in bitfury thread     
https://bitcointalk.org/index.php?topic=228677.0

Just you wait :)

... perhaps BitFury purchases in non-US work out better.  From miners to miners is a bad tagline based on current pricing.

I heard a rumor Bitfury 2 would be here by March, sample chips end of February. I cannot wait to see what Punin brings to this fight! 8) It's on!

Just let AM supply the canary  with tons of 7gh usb sticks at a good price.  I will be a happy camper.

7Gh/s - how do you figure that?  shouldn't they be 10-12 Gh/s?  I can't remember if I did my math right...

There is not important "how many Ghas", if you will be able to buy chips and use open projects "to do it self", it save more than 50% from USB or Miner price

The post you're referring to was made while they were waiting for the first chips from the fab and actually didn't know for sure how were they going to perform.

This is what Bitfury said about the design of the chip:

and I'll have to look for the other post where he said that there was a bug on the die and because of that they couldn't get the double performance (7GH).

Also, there are very few people that have actually achieved even the 3.3GH as that requires very very precise conditions, and even reaching 3GH is quite a challenge. Most chips perform fine and start hitting the limits in the 2.5-2.7GH range.

I just wanted to set the record straight so that people don't have the wrong expectations.

I've done both, as I would prefer not to fall through the cracks on this.

Good idea. Emails can get eaten by spam folders ...... caused some 'missed' emails to FC recently it seems.

Note all my estimates are rough not exact!

I am thinking they will undervolt to keep a stick in the 2.5 watt range.

this was done with the antminer u-1 chip vs the the antminer s-1 chip.  same chip lower volts and lower freq settings for the chip in the sticks.

same chip the usb stick is undervolted a bit and the the freq is set lower.


the antminer  usb stick gets 2 gh pretty easy after that it is too hard on most hubs .



same is true with the cubes vs the be sticks  the am cube pull more gh per chip due to more power and better cooling.

Still a usb stick in the 6-9gh  range vs the 1.6- 2.0 range is a big boost.

You don't think they will design usb sticks for more than 2.5 watts? USB 3.0 is becoming pretty standard now and hubs are able to supply closer to the standard of 900 mA per port (4-4.5 watt). I wouldn't be surprised if asicminer provided adequet hubs for additional cost either.

I'm not sure AM will be focusing on the USB stick market anymore.

Maybe not, but a third party will probably will make one with AM chips. I'm interested in a dual chip miner though, like the bifury or twin fury usb's. A 15-20 gh/s Usb miner would sell.

unlike USB 2.0, USB3.0 hubs/ports/devices are not all adhering to the max power output per port.

If you want to do it do it right and get a blade of some sort. USB sticks are a waste of time and energy figuratively and literally.

+1 That would be awesome!

NinjaTech is interested in partnership.  :)  ninjatech.org (http://ninjatech.org)

At the risk of repeating myself, FriedCat asked for enquiries via PM or email. If you just post in this thread, chances are he won't see it.

The A1 chips you're selling were quoted as "0.35J/20GHash" too. Look how that turned out!  :P

Anyone actually contact FriedCat? Seems like 100's of PMs and emails and no one got a reply.

No answer I insist, probably he found a big investor and don't need us anymore.

How do you know he got 100's of emails and and PMs? How do you know nobody got a reply? The people he replied to are working with him now, not announcing the fact he replied to them to the world.

Easily.

No one posted they got a reply. There are at least 30 people said they PM'd and emailed. Easily 100's have tried contact him. 5 or 6 from our own collective have tried. 0 response. I think the BIG INVESTOR idea probably the result. Money talks community has to walk.

Open Source ASIC chip beginning to look like the only real viable option.

Don't panic! Really! The tapeout should occur about now. AM will have reference chips by the end of February and are going to get back to all people interested when they have the chips tested and their actual specifications determined. Friedcat is committed to give an announcement before March and I'd be puzzled if this happens before about February 25th.
Wait for the actual final offer and you'll be fine.

I got reply about share transfer this week. So Friedcat is alive. I assume he's waiting on detailed specs from tapeout. If Friedcat speaks, he's telling the facts, not bullshits like everyone else in mining market. So just be patient ...

I have to believe this is spot on. I've worked with a few ASIC companies and they either communicate well or not at all.  Friedcat has a long history in this business, I anticipate communication from him or his team in the next few weeks.

seems like friedcat is still nowhere to be found

I just got word that the next announcement is now expected to be given 'after the first week of March' and that 'everything is still on track'.
So don't hold your breath for an announcement today or tomorrow, but I feel that said announcement will be the publication of the final specifications and the details on how to order gen3 chips.

Nice timing. I will certainly be eager to get these into a Wasp.

Good! We need as many good board/miner developer as possible! I believe they'll make excellent chips, and don't forget that ASICMiner is reliable!

This seems to be around the time to start watching this thread so putting in an observant post for the ANN

Maybe this is all some form of market research ;D

 ???

I am also looking for what he is selling, except shares...  ???

He is selling chips. Have you contacted him?

to be clear, Friedcat has selled all his shares out 1 year before at the price 0.1 per share.

all the shares owner who buy at IPO price have been rewarded much.

Now , Friedcat just working hard and focus on 3-generation chips , be patient, sample chips maybe will be out this month.

Is there any thread about selling these chips? However, I will try to PM him also.

 ;) this is that thread.
But mainly for resellers and/or board manufacturers, looking to buy greater quantities.

It depends on how many you want. He made it clear he is looking for *volume* buyers. If you haven't got a reply, its because you aren't talking FC kind of numbers unfortunately.

anyone had a reply yet?

I wish they had specified their defition of high volume. It probably would have saved them a lot of PMS and emails, and given us an idea of the requirements for consideration.

I would hazard a guess and say 5 - 10,000. It really doesn't make sense to supply chips in the 10's or 100's, and if they did, it would be through a third party that is buying 1000's, i.e. a distributor, so it's extremely unlikely that you'll be able to buy 10 of these for $64 ($0.5 per gigahash) probably more like $200 ($1.5 per gigahash). That doesn't in any way make them bad value, quite the contrary, but I think that a lot of miners have expectations that they're going to be able to put together a rig for $0,5 per gigahash, and that's simply not going to happen.

Perhaps Asicminer could have worded their initial pitch slightly differently, but good on them for going against the herd and making the most of existing, cheap technologies. True Value Engineering.

Don't know if this article in cybtc.com is related to ASICMiner's new product.

http://cybtc.com/article-655-1.html

Google translate is hilarious. "baked cat talk forum"... I had to stop after that.

Honestly, standard reel sizes would still be fine.  Reel of 1500 (18PH) for $20k?  Sign me up.  It really leaves the middle-ground design teams in a state of paralysis...  Big might be too big, but small might be too small.  Anyway, hoping to hear something soon.

edit: re article and google translate.  Evidently, they can't decide how to cook the cat.  roast cat, baked cat, grilled cat...

left an email  :D

I'm hearing scuttlebutt* that a deal is in progress to supply a Chinese manufacturer.... so just wondering if anyone can point me to details, can't find anything.

I'm just eager to buy miners or at the least boards with AM's new chip.

*rumor

This one? https://bitcointalk.org/index.php?topic=99497.msg5485359#msg5485359 (https://bitcointalk.org/index.php?topic=99497.msg5485359#msg5485359)

Thanks minerpumpkin. I wonder what us smaller manufacturers have to do to get on AM's radar...

I think you have to do a song and dance

in the meantime... bitmain keep delivering kick ass products... https://bitmaintech.com/productDetail.htm?pid=00020140314135446510cxXNeYAy06E5 1th for just 3899 USD...Power efficiency: 1Watt/GH/s

friedcat needs to work harder for his shareholders this time...  ;D

And Bitmine also delivered A1 Coincraft chips so China is full of 1THs miners with even lower consumption that the one in the above link  ;)

1w/gh is a bit late at this point. AM gen3 is going to put all antminers out of business.

I honestly expected more from bitmain gen2.

and when do you predict these gen2 miners to be actually hashing???

they are not gen 2, just a under volted build of their current chip, more chips, less watts.

asicminer will be gen 3 not gen 2, gen 2 skipped production.

gen 3 will be close to 1TH 500w $2000

Not if ASICMiner continues to overprice their mining hardware. Given ASICMiner's previous track record pricing hardware so miners have no chance of ROI unless they resell, why would you expect them to do anything differently with Gen3?

I doubt ASICMiner will be able to match Bitmain pricing.

Better be, no way I'm paying more than $2 a Gh now and even that is looking highly short term.

aaand more competition... here comes avalon gen 3... http://downloads.canaan-creative.com/hardware/A3233/A3233Q48-140313-V01-EN.pdf

768 hash cores, 7 GH/s per chip, 0.75 watts per GH/s, on a 40nm full custom design, not as power efficient as they should be but still adds more pressure..

0.75/w is hardly competition. Unless these chips sell for dirt cheap I don't see how they can possibly compete with AM chips that are the same die/process size but with twice the hashrate and efficiency.

This all assumes they meet simulated specs.

If knc 20nm is actually 40% more efficient than their 28nm they might be on par with asicminers efficiency but I doubt they can match the price with 20nm being so much more expensive.

Not only that,KNC (& Hashfast,Cointerra,etc...) will/have only offer miners in the $5000-$15,000 price range,not many are offering anything for the little guys,so only the rich can get into Bitcoin mining  >:(

Thank god Bitmain came up with the S1........now if Blackarrow can get its act together with their Prospero X1.........& maybe ASICminer can offer an affordable miner too.......one can dream can't he  ::)

Don't forget when the Antminer was released the price was 4,5-4.75BTC @ a BTC price of 1000 usd. Eventual the price came down when the diff went up. I don't expect many AM gen miner with a hashrate lower as 1TH. When released in april the price need to be around 4,5-5 BTC/Th. Probably it will, because in april there is a lot of competition, Avalon gen3, Bitmain s2, Chinese A1 (stolen chips or not...). Without a reasonable price people will buy the other "less efficient" hardware.

Totally agree with Fireblade, in 2 weeks prices will be half or even less, as it was with the others miners...

I believe prices are currently set based on factoring competition and ROI which is to say there's potentially higher margins then simply bringing the product to market. At some point soon I think it will be simply who can build a miner and sell it for a profit period. The more cost effective manufacturer's will prevail. The inefficient one's will have to bow out.

At some point the only other cost cutting measures will be to remove the distribution supply chain and logistics of distributing equipment. I can see a movement to investing in a manufacturer's mining operations to eliminate all other unnecessary costs of mining.

It may soon come to a point where there isn't enough margin to keep shaving the ice cube on ASIC development efficiency and the the efforts go to running with what you got in higher numbers. Quite conceivably AM may have timed Gen3 perfectly to come in with an efficient ASIC for this very purpose and at a time where further development isn't going to provide significant return for other ASIC makers. They'd have to come up with something considerably more powerful and cost effective and I believe Gen3 was already designed with those two factors in mind.

there's also bitfury that seems to be doing pretty good job on their version 2 chip... and that's just 55nm, i think they can squeeze helluva more should they move to 40nm  ;D

https://bitcointalk.org/index.php?topic=228677.1080

punin's comments: "Our preliminary tests show 25% increase in hashrate and 25% lower power consumption. We will receive new chips tomorrow and will have samples available to developers."

If they had to pay for a whole new mask then that doesn't seem so good to me. If they paid very little or zero for the mask then it's ok :)

Bitfury is just too expensive. Way too expensive its not even funny anymore.

€4,864/$7000 for 560GH/s  ???

That's the old chips. The new chips are priced better. We'll have to wait to see the price of full systems with the new chips before we can call them overpriced.

They will need to be, there is a lot more competition around these days. They won't be able to get away with charging like wounded bulls while (allegedly) financing their own mine this time round.

All this talk about widdling down the cost/efficiency of chips to be able to achieve ROI is intersting.

Since were down the price of moving electrons, riddle me this.

Today I asked UPS how much the shipping charges were for sending me one of my U1's. I was quoted, wait for it, $338 using UPS express saver the way it was shipped. Three hundred thirty eight dollars which is incorporated in the price of the U1 and works out to half of the overall cost. Now how much margin is there in these current devices and if time is money how the heck is anyone going to compete against employing them at thier place of origin if they are paying as much to move them as the unit itself could sell for locally.

I'm hoping UPS quoted me incorrectly because this is the elephant in the room.

Buy from local resellers - drastically reduces shipping cost

Not so much "incorrect", but not relevant. People who ship lots of items get large discounts on shipping rates. You on the other hand will be quoted costly "retail" rates.

Well that's good news. It was quite disconcerting that there appeared to be such a discrepancy in the cost to mine if you didn't happen to reside where miner's are made. It did elevate my concerns over the critical requirement to look at distribution, assembly and design from a logistical perspective if your interest is in deploying your product world wide. Sending out completed individual units out from the manufacturer given the potential distances involved probably isn't the most cost effective approach.

The local resellers weren't competitive but where I am probably isn't fully defined as "local". It's a fast moving market and I have a sense it caused the imbalance that allowed for the manufacturer to supply for less even with shipping factored in.

You need to use a comparison site so they compete for your business. I'm in the UK and Australia is as far away as possible from me, with http://www.parcel2go.com/parcel-delivery/australia I can send a parcel up to 15kg/1.2m for about £24/$36USD!

https://bitcointalk.org/index.php?topic=250249.msg5751682#msg5751682

Did FriedCat make a reply and then deleted his post?!

https://twitter.com/FriedcatSays/status/445828515490193409

You are a fucking genius. What would we do w/o you?

0.6w/gh is not really comparable to 0.35w/gh and at $0.5-1/gh asicminer is by far the cheapest option available.

However their next generation 5.4th miners using 28nm will compete in efficiency but probably not in price.

"The Yukon will deliver...."

All these guys are starting to sound like a bunch of weathermen.

Proof is in delivering.

It strikes me that we will experience a repeat of the 2013 mining race, where 50% of the offerings are fraudulent, 25% are failures, 20% or OK, and 5% are actually winners (*numbers made up for illustrative purposes, final stats welcome).

The Yukon sounds good and overall, esp. power consumption wise (if at the wall), on par with AM and that supposed AM rockxie miner. May even have a very very slight edge what power consumption is concerned. But it's not blowing AM away. It's just some slightly more powerful competition the cat will have for breakfast.
The thing is - we know AM. And we know, they deliver. In Bitcoin world this is priceless.

This 'brilliant' design sounds suspiciously similar to the asicminer one in all but name (and power consumption).

Could it be the same chip? It's watts (or rather Joules) per gigaghash figure is about 2 - 3 times what the asicminer is spec'd at (0.2 - 0.3) but I am curious as to exactly how they (asicminer) managed to coax these figures out of what is clearly a standard SHA256 implementation. The Israeli one is probably closer to the truth, but even then they're having to run the device at way under it's normal core voltage to get the power down.

I'm thinking of buying around 1100 chips. Could you direct me to the PCB design as well and give me a quote on 1100?

1100 chips, that's 14 boards or so and <7kW total, given that AM's china reseller cranks out 1TH from a board with 500W.

And you want to bother with your own PCB design, getting a quote and all that?

I am not sure if FC had that in mind when he wrote the following:

You might want to investigate what The Wasp Project is up to. They seem to be very interested in getting on board with Asicminer and may be in a position to buy volume for resale to support their efforts.
https://bitcointalk.org/index.php?topic=422243.0

I believe friedcat has regional distribution (resellers) as well that may handle smaller orders. Back in September they were noted as the following :

 US: eleuthria, CanaryInTheMine, SilentSonicBoom
 Canada: teek
 EU, Switzerland and North Europe: yxt
 China, Taiwan, Hong Kong and Macau: rockxie
 Australia: asicminer@swishbits.com.au
 India: Pinwheel

We are keen on doing up a AM2 chip Wasp. Just point us to the sample chips and the detailed specification and we will design it and release the BOM and Gerbers for it.

Update March. 29, 2014: The chips passed all functionality tests.

Fantastic!!!

good news！

We all wait so long time,it sees like good news.

i am waiting for you!

F*ck me sideways and call me friedcat :)

lol

WPC notified we will be receiving samples of the chip. Thanks Friedcat we PM'd you our EE's address in Seattle.

Very good! Please let us know when you receive the chips and how their performance is!

Will do.

Sample chips received.  They are real.

Very very good! Care to provide some photos? :)
Are you able to test the chips in any way or are you "only" in for the reselling?

Yes, more details please! Are you able to test them?

Canary or me? I don't know my way around electrical engineering enough to test them, unfortunately. Plus, I'm located in Europe.

But I believe even if Canary isn't able to test them, he received them in order to encourage him - as a major AM reseller - to start a group buy. I guess the most important figures designing PCBs and miners also received samples or are in contact with AM in order to receive some. Hence I'm not worried that the relevant people haven't received any chips.

sexy talk  ;D build something and send me it so i can send someone some coins  :-*

what else is known? pictures? will you be able to test them yourself soon?

So what is the difference between the sample chips and the real thing? How far aware from the real thing does this mean we are?

Excellent!  This is great news!

This is basically "the real thing". When you design chips, you have to spend a lot of money and time before you get the first real chips and can be sure, they actually work the way they should. A rough outline showing the importance of the news: https://bitcointalk.org/index.php?topic=99497.msg5968245#msg5968245 (https://bitcointalk.org/index.php?topic=99497.msg5968245#msg5968245)
Afterwards you can order them en masse. Production of a batch takes about a month or more.

So we're still about a month out before they're ready?

Before mass production, yes.

Awesome news. Been waiting a long time, thought these chips a long were coming while back, but always knew that FC would deliver a great product when they did come at last.

let's see some numbers

FTFY

EE received chips thank you gentlemen:


Anyone interested in working with our team please contact me via pm or email.

12gh at 4w tested?

Give us the final specs big cat!

Seems right on target, 4 W/12 GH/s is just about 0.35 W/GH/s...

Just the "specs" included nothing tested yet. We will let you know when we get that far. We will probably look at a quick turnaround USB module solution for testing called "Sting" maybe.

How to get samples and documentation? Thanks.

Subscribe to the question.

I'm slightly perplexed about this whole scenario. The new chip is without doubt a good idea, especially for the DIY crowd, but what I can't fathom is why prospective buyers would want to use these chips when their manufacturer has made it abundantly clear that they want to sell huge amount of them to 'mining operations with large PCB making capacity'? Surely this is going to dilute any potential future earnings - small players can never hope to compete on economies of scale here, and remember that $0.5/GH or thereabouts is only going to be available to 'huge' volume users.

What an actual system will cost to make will be a lot more.

The net result will be that the network hashing power will go through the roof, and only the big operators will be able to survive.

Comments?

Some will have access to the appropriate tools and setup to be part of a group buy, get some chips, assemble, and get their miners online before the competition. They're betting that with the initial low cost, they will be ahead of the large scale DIY manufactures, who need to order massive amounts of parts, assemble, and deal with shipping, customer service, and etc and still make a profit.

Plus with KnC, CoinTerra, and HashFast going their current paths, what other alternatives do miners have?

The manufacturer may only sell large quantities directly but there's nothing to say a distributor won't. Will being able to purchase in large quantities be advantageous? Likely since anyone in the supply chain wants their piece of the pie.

I am interested to buy chips, is there a minimum quantity required ?

Good points, my worry would be that by the time the smaller guys get sorted out, the big ones will be going full steam ahead. As for KNC and co, it will be interesting to see how (if) they respond.

Friedcat, we are waiting for you.

I LKE FIRECAT :) :) :)

In Bitcoin, You wait for friedcat.

Not sure if this has been posted here yet:

Update

The documents as well as the (verified & produced) sample design are out:

Schematics - https://mega.co.nz/#!uNMkFZoI!lnlRehlynQQzNRfQ87_UYek1RtrOUvLZ6a074XeRqlo

Reference PCB File - https://mega.co.nz/#!DZVHgS6D!6pmTsmrito8rfVJJ-etvznOtUcalbAT6vU-C8EqY0_I

datasheet - https://docs.google.com/document/d/1_dmXzyMqtr7tVZukk2DAJih2sRq5jDy5DQPkwCce570/edit?usp=sharing

bonding list - https://docs.google.com/spreadsheet/ccc?key=0Al1fvFT7Sd5bdFFZSUllNW5seVRfM01kcDVqZkJfdFE&usp=sharing

9x9 package pictures -
https://drive.google.com/file/d/0B11fvFT7Sd5bUV90cFN4cEhTNEk/edit?usp=sharing
https://drive.google.com/file/d/0B11fvFT7Sd5banZkMDZqSTBGTEE/edit?usp=sharing
https://drive.google.com/file/d/0B11fvFT7Sd5bWjh6MmJsMFpCYkU/edit?usp=sharing

8x8 package picture -
https://mega.co.nz/#!DA1mBJzB!7hFnoT8ZYQd1d30m2oHbuscD36cCQ4ou7stsr5Lo03Y

The package we are using is 9mmx9mm, but it might be partly switched to 8mmx8mm.
Also, we plan to increase the center pad size to at least 7mmx7mm, so please leave enough margin on the PCB to avoid unnecessary re-designs.

what's the power consumption ?

Hurrah! Fixed the first 2 links

Important stuff:
QFN package
The typical hash power is ~12GH/s in rated mode.
There is a pin to initiate turbo clock ranges:
bs decides the operating mode. When bs=0, the range of core clock frequency is 200MHz-400MHz. When bs=1, the range of core clock frequency is 375MHz-750MHz.
Core Vdd, 0.55v~0.88v

Any news on the actual consumption of the chip and availability date/quantity/price? When will these figures be available?

+1

The actual consumption will depend on the PCB design that the miner will implement and the DC to DC power circuit (buck regulator, MOSFETs, etc). The only thing we know today is the info from Rockxie's first tests with the test board he designed. He had some component problems unrelated to BE200 as far as I know.

I assume the AM BE200 does not support a "OneString" design?

Excellent!  This is great news

For all you AM fans, this might be slightly disturbing:

https://bitcointalk.org/index.php?topic=387533.msg6358781#msg6358781

Seems AM are going to build some 8 million chips by June. That's a lot of wafers, and a lot of capacity. It will take them a good few months to deploy it, but when they do it's going to drastically affect earnings, I'd estimate by September that the network rate will be 200PH. At that level, a TH of power might earn you $195 a month after you've paid for your power, and they'll want to add the same again, no doubt.

I just don't understand your enthusiasm for your soon-to-be biggest competitor? (albeit indirectly)

They're not deploying that power for self mining. They're selling those chips. Still the same increase, but that's simply the way things go.

I'm having troubles downloading that file ... has anyone else been able to get it?

AM's goal is to reclaim the lead in chips and keep it.  To avoid the centralization debate, the emphasis is on selling chips to anyone who wants to create their own miners.
in the long run, we should see something like the Intel/AMD (maybe few others will join Bitcoin ASIC chip producers club) as they relate to PC/server manufacturing today.  i.e. AM makes the chips, but other companies build miners to satisfy different market segments.

also, very soon, there's no reason why Samsung TVs, DVD players etc... shouldn't have a swap-able card that runs Bitcoin mining chip/s.  that device may be "free" to you if you provide electricity/internet and Samsung keeps mined btc?  (substitute Samsung for any other maker of some appliance).

it's possible that the entertainment industry can use Bitcoin to "solve" the piracy "issue" when they finally see the light and get some forward thinking, technology understanding board members at Hollywood & Co.

and 3d printing can benefit by using Bitcoin to deliver one time use/print designs for an item you print at home or local Kinko's (former or new 3D-Kinkos).  no need to go to store to buy a cup for ex.  (also, no need for another country to manufacture it and ship it in a container overseas).  Your payment for a printed item, can be made to multiple parties in the new supply chain: the designer of the item and the raw material manufacturer.  the 3d printer could have a chip to help process printing requests and handle instant payments to all involved (no need to have NET 30 terms when I buy a cup, everyone gets paid instantly).  this is a high level, simple example using a cup, let your imagination run wild.

devices/systems that utilize Bitcoin (yes, upper B) network may need/want chips for uses besides just mining for bitcoins.  AM is positioning to take the lead and keep it.

This is an interesting concept, Canary.  I can imagine this might spur a startup or two--is anyone talking about something like this I could read more about?

there are already systems under development (near release) from the 3d printing leaders to offer pay per use 3d printing.  its also been proposed doing licensing for printing popular likenesses and characters etc, as well as in-store printing of your own figures etc.  this is coming to the retail and home environment, regardless of bitcoin and isn't specifically a bitcoin related business.

http://www.3dsystems.com/press-releases/staples-partners-3d-systems-launch-store-3d-printing-experience-new-york-and-los

its not clear how bitcoin itself helps avoid piracy... although in general - business model-wise - if only the movie industry would offer its content in a form, and at a time when the customers wanted it (and pay for it) maybe there'd be a lot less piracy.  these staggered windowed releases and staggered territory releases are whats hurting them.. yet again, bitcoin or not.

How are you suggesting that AsicMiner's bitcoin mining chips can be useful outside of mining for bitcoins?  you mean mining for some other SHA-256 coin?, or are you referring to non-mining uses (are there any?)

In the medium to long term, why not put a few ASIC chips into a device that uses power to produce heat (electric hot water heater, oven, space heater) and earn some satoshi while running?  You could recuperate some of the electricity cost that way.

not aware of specific startups but in today's world, when you get an idea, it's probably not unique anymore... wouldn't surprise me if these concepts are being worked on already.  I mentioned one of the ideas a while back on Reddit, maybe someone picked up on it already.

I'm actually designing a house right now and I'm very tempted to use bitcoin miners instead of boilers, to generate heat (heaters, hot water etc)... and sure, I'm paying for expensive electric heating, but they're earning coins at the same time, so it means free heating !

Thats what I did to partially heat my house this winter.  I bought some AM cubes in November, positive ROI, and kept the house warm while mining.

well, as with proof of work and elimination of double spending (Bitcoin), proof of consumption can be done so that you don't "get" a second copy for "free"?  if you print that cup, it was actually printed beyond any doubt and when you watched that movie, you actually did.
of course with proof of consumption ideas, the price per item should come down significantly, that cup may cost you pennies, so should the movie.  you pay per "proved" use/consumption?

what you've described isn't per se a use for AM's chip.  you're describing a new coin entirely.. that is somehow related to the production and printing process... i.e., the 3d printing machine would have to be linked internally to the '3d printing coin'... so that they are inextricably linked...

We should be able to use Bitcoin (not btc) instead of re-inventing the wheel?

Now if someone could just invent a BTC miner that pumped out cold air! My A/C is already running non-stop from all of the excess heat.  :(

That would be great but I think it would violate the 2nd Law of Thermodynamics

AM's goal is to make a lot of money, and they seem to be determined to out produce everyone else, in the process they'll make mining completely uneconomical for everyone who doesn't have their economies of scale. Do you really think a DIY miner is going to be able to make their rigs at a cost that will allow them to compete? Even if third parties produce systems or subsystems you're not going to get hashing power at anywhere near the headline $0.5/GH in AM's initial release, it you think otherwise you're going to be disappointed. They are not your friends or saviours, they should be viewed as your mortal enemies, financially at any rate.

As for Intel and AMD, financially to them Bitcoin is a minor sideshow. If Intel wanted to, it could make a chip that would make previous efforts look like amateur hour, and it could do so in less than a few months. But why bother when it can make huge margins off it's server cpus which are sole source?

2 phase Immersion cooling might be the solution.

No heat. No sound. Much less space required.

Damn you Physics!   ;D

It's called an absorption fridge. http://en.wikipedia.org/wiki/Absorption_refrigerator

Yeah, but a lot of the concepts you laid out above are basically re-inventing the wheel, albeit lopping BTC on top for no good reason.  The goal of most manufacturers is to produce things cheaper and better than anyone else while getting them out to the door sooner so they can charge a premium before the market catches up.  Sound familiar?  BTC is small fish compared to any other manu market in the world, so there is no reason for them to even care about a) sourcing chips, b) understanding the tech/sourcing techies, c) dealing with networking issues, d) dealing with already idiot consumers.  Samsung doesn't want a tech it can't control, shit, it can't manage to keep its devices online when one data center has a fire, you think they're going to rely even 1% of their business on a hashing network?

The problem with the BTC optimists is that they all remind of me of John Stewart in Half Baked.  Saying "... but have you tried owning a toaster, THAT HASHES?!" isn't a compelling argument since that would mean I now need my toaster to be network-enabled (or open up my wireless to another device, f-that), it may require tech support, the BTC chip can fry and I have to send my toaster out for repairs.  

The bigger problem is that we don't need a coin for that^tm.  The mining market has effectively done itself in because of all the big players with $$$ in their eyes (note: not BTC in their eyes, BTC is just a vehicle to cash out with) are cannibalizing one another and the bit players in the community that started it all are now irrelevant.  Now people are trying to find irrelevant uses for the technology.  I remember reading an article about a year ago talking about all the possibilities for BTC since it would provide African farmers with a sustainable currency that avoids national conflicts and thinking "These people are fucking idiots.  African farmers are more concerned with running water and basic medical care."  

Not trying to be offensive, but the mining market is insanely self-serving and relies on a wide net of scammers convincing other suckers to buy stuff that isn't worth buying.  AM is late to the game, all they're going to do is flood a saturated market and make it more likely that other vendors will pull a KNC and self-mine rather than unload cheap tech to willing suckers.  They can just unload cheap BTC to willing suckers instead and people will think they're getting a deal when BTC price is a-falling and nobody cares enough to buy it back up.

There's a strong disconnect with reality in this community.  Unless AM manages to produce (and produce for shareholders) borderline immediately their next line is likely going to be about as useful as the cubes.  Unless you're the one selling hardware.

Good grief, a man talking sense. What on earth are you doing on here?

10.9 million chips @ ~10G = 109 PH/s by June's end.

Note: Only ~3 PH/s is scheduled to be retained for self-mining, the rest is to be sold or franchised to interested parties.

Get the lambs to pay for their own slaughter then? That's smart.

Still no answer to my reply: https://bitcointalk.org/index.php?topic=99497.msg6334999#msg6334999 I don't understand how can you think that we will have 109PH deployed by AM by the end of June. Who will invest so much money?

I don't think the lack of interest in your group-buy accurately reflects the market conditions faced by AM.

I assume the chips are already paid for by large AM customers because:
(1) the delivery dates suggests that the chips are already ordered
(2) fabs are not known for giving generous credit lines
(3) I don't believe AM has ~$21.8 million in spare cash lying around (109PH@$0.20/GH wafer cost) to build stock.

So to answer the question of "who", suffice to say there appear to be large buyers out there.

Bitfair is quite correct  in his assessment, however it's extremely unlikely that anyone would be able to deploy such a large amount of equipment in such a short time let alone build it. If the chips (packaged) come out in June then it would likely take until end August at least to fully deploy them, even with tons of money behind you. My question would be: what size is the next order and when will it be delivered?

My group buy has nothing to do with it. Since you haven't red the post that i pointed to i will ask here too. What's the expected $/GH for this numbers? Assuming 1$/GH at system level, that would mean that 50PH=50M$. With the current exchange rate and with the current market i simply don't think that will happen. A lot of people in the Hardware forum are screaming that mining isn't worth anymore, but yet you expect 109M$ pouring in a matter of 2 months. I don't see that happen.

My point is that it is largely irrelevant at this point from the point of view of AM: AM appears to have collected payments for the chips already, and so the buyers of the chips are the ones taking the risk.

Anyway, my calculations show that this hardware will produce a (significant) positive return, and where there is positive return I believe there will be buyers!

Having chips but not mounted on boards and either deployed or sold means that 0 hashrate will go online. What are the buyers of the chips doing with the chips if nobody is buying them? Start deploying their massive farms?

Your argumentation has become pathetic.

I am still waiting for some kind of good and well-argumented replies not posts who bring nothing to the discussion. I see none until now.

AFAICT, you are arguing that there is no demand for these chips from miners.

Yet, AM appears to have received the payment for them: this fact unambiguously proves that there is demand for these chips!

Can you tell me again how you concluded that there is no demand? Or, in case I misunderstood the point you were trying to make, please repeat it more clearly.

Never said there is no demand. There is demand for sure. I don't argue that. I was just asking at what price do you see those chips being sold to miners when they are assembled? Until now nobody answered that even if i seem to get replied a lot.

In that case, I misunderstood you - my apologies.

A while back, FC said the chips would be priced between $0.49 and $0.99 per GH.

The price of a plug-in-ready miner is anyone's guess, depends who builds it.

If they sell chips for those prices then the end price will be higher than 1$/GH for sure and we are back to my posts/questions again. If you expect 109 PH deployed that translates to 109M$ worth of equipment sold to the miners. I just don't see that happen in such a short time.

Assuming the chips are cost $0.5/GH, ie aout $5 each, then a 2TH box will come in at a cost-to-build price of around $1600, or $0.8/GH assuming it has some form of internal heatsinking/fan arrangement and they give you a PSU with it. That's to get someone to assemble the bits, the chips take away $1000 of the cost, so if they cost $3 to make (?) then AM could in theory get systems for $400 less, around $1200 for 'internal' us. A third part supplier w0uld want to make a margin of at least 25% on selling their box, so getting anything under $1/GH is pretty unlikely.

25% margin is pretty high for such a competitive market. Using your number of a $0.8/GH total build cost for the assembler, different margins would produce different prices:
5%: $0.84/GH
10%: $0.88/GH
15%: $0.92/GH
20%: $0.96/GH
25%: $1.00/GH

So by adjusting your margin, you could easily get below $1.00/GH. And I imagine that the market is so competitive at the moment that margins should be on the low side.

Anyway, it's difficult to say what final prices would be, depends a lot on what kind of miners are built: immersion cooling blades, standalone boxes, rack mountable, air-cooling, liquid-cooling, etc. A lot of different possible design choices can be made, each with their specific costs.

Judging by the history the in stock miners were always overpriced while pre-orders had a lower price. If we add up shipping, customs and maybe PSU then the difference from 0.84$/GH to 1$/GH isn't so big as it looks when comparing 5% with 25%. Now(i know some people will be pissed)...if we look at my GB then we see that pre-orders for a company that already has launched products aren't flying off the shelf even at 0.83$/GH. Even if AM sells in stock miners for 1$/GH or less i still find hard to believe that there are people/miners/investors that are ready to spend ~100M$ for mining equipment in 2-3 months. In my view most of the money went down the hole of the fails and scams that ran for the last year. Now there aren't simply enough money, not unless the BTC price goes up a lot and very fast.

Sounds like you are judging market conditions by the interest in your pre-order group-buy.

I'm surprised your group-buy has any interest at all, since it sounds like a profoundly stupid idea to send $5 million to some guy on an internet forum in order to (maybe) receive an item in four months.

I think most people would pay more to have a miner today than have it for a cheaper price in a few months (maybe)

My group buy has nothing to do with the amount of money are ready to invest into mining machines in just 2 months starting now.

Regarding the group buy, I am only collecting reservation fees from which 70% will go to Sean's Outpost, i will only keep 10%. The "$5 million" will go to the company who has proven itself until now if we get to 1000 units. They have a very good product launched, a very supportive and active support and they delivered before the announced shipping time so i see no problem there to pre-order from them. If you have a problem with them then feel free to vote with your wallet, but don't try to change the subject of our original topic. Do you honestly believe that there are $50M waiting to be invested in just 2 months in AM gen3 miners?

Yes.

Although perhaps a little more than 2 months.

Define more than 2 months! It can be 3 months or 10 months.

I guess we'll just have to wait and see.

So you can't or you are afraid to tell a number of months. You are arguing with me because I don't believe that $50M will be invested into mining in 2 months, but can't give a specific time? Then what's the point of the argue? It's just a dead end. I gave you arguments you give me nothing. It's just because you say it so. If the money will be invested in 10 months you will tell me that you won the argument because it was a little more than 2 months while i have a strict deadline of 2 months you have no deadline. Be fair.

Don't get your panties in an uproar.

I believe that the hashing power in question will be deployed within five months.

You say you have given arguments, but I see nothing more than opinion. You haven't backed it up by any arguments. You're all like "I don't believe", "I don't think it will happen", "I find it hard to believe", etc. And that's fine, I have nothing against opinions, just don't try to misrepresent it as "fact" or "arguments".

Anyway, I'm sick of you and you have trolled me all too well. From now on, you are ignored.

Ok. No problem with that. You just prove me that i was right :) The argument started when I said that i don't think it will happen in 2 months, you think it will happen in 5 months for a whooping 250% more time. Thank you for proving me right.

I have a design ready for this chip based on the documentation provided so far.

Any clue how to get sample chips? I'd be willing to pay a nominal amount for a few chips.

thanks!

Please explain how that "proves" that you're right? From what's been posted, there isn't anything that indicates as such. Have you heard the phrase "the absence of evidence is not the evidence of absence" before? In other words, just because it cannot be proven/verified, it doesn't mean it's completely false. Your own statements say "I don't believe that $50M will be invested into mining in 2 months." At this moment, we all don't know and everything that's been posted is only conjecture and nothing more. The majority of those in favor of AM do not have facts to confirm that those funds will be present when the hardware is available, just as you do not to verify your own claims. No one has been proven right currently, only time will tell.

here's what i'm looking for in a chip: quality, stability and reliability. throwing about figures on who's got the biggest profit margin, and speculating on how much money will be spent on ASIC's in the near future, is just obscene - we are all customers and we are all miners at the end of the day and it is indeed, us all who will contribute to the coffers of the manufacturers we choose... what i'll be buying now and in the future relies solely on performance and reliability. I'll be more than happy to buy a slightly more expensive set, if need be, to have more stable and reliable rigging deployed. I look at past history and I look at design. What i'm currently seeing is a gen1 from Spondoolies, which is merely only just matched by AM's latest.
So, I know where i'm swaying towards at this point in time, and I look forwards to the future, with relish, but I have to say,  at this point in time AM are quite a way behind in the race.

Having the most efficient and powerful asic currently in existence puts you behind the race? You learn something new every day.

I have it, I have it running now and I have it mining BTC.

jimmothy -

first, that assumes asicminer's chip is the most efficient and powerful chip in existence.  that assumption is now likely to be incorrect (says rock miner who has tested it as using twice as much power as it was intended,which means it costs twice as much to run)
second, asicminer is at the samples stage and is not yet in production nor has anyone built nor shipped any systems incorporating the chip.
third, getting systems into production takes time.  there's supply chain issues (ordering parts, manufacturing, shipping etc).  more goes into a bitcoin miner than hashing chips (there's power and control circuitry too)
fourth, asicminer's new business plan assumes most of those chips are bought by other systems builders, and the public and are not intended for asicminer to use in their own mines.  thats a big assumption when bitcoin is at the current price... as sales of new bitcoin mining equipment has somewhat slowed and may not pick up til prices are higher.

if rockminer's claims are true (and certainly need validating) then asicminers new chip isn't the leap in performance or power that was promised.  its a perfectly ok chip but breaks no new ground and hasn't beaten other chips in systems that are already in production.  as others have said, it isn't any better than spondoolies last chip that is already being shipped to customers (and they're working on a better one for the summer) and of course, bitfury has recently released a pretty good V2 55nm chip and has promised to follow that up with further developments soon.  And thats ignoring all the other big players, who are also working on their next systems.

Companies normally tend to overvalue in the first instance, and the IPO for rockminer was set out to raise around 7600BTC. I'd certainly be apprehensive, until we see some kind of validation on their product. The claims are not impossible, but do give that 'too good to be true' sort of sense.

Is there another chip with less than 0.55w/gh? Spondoolies is half the GH per chip and bitfury rev2 is something like 0.7w/gh.


But if it was spondoolies you would throw your money at it right?

You seem very dedicated on spreading misinformation. Rockminer raised 1875BTC and we have already seen products sold from rockminer. They sold asicminer gen1 hardware.

IPO finished in 15 seconds flat so I'd say people disagree with you.

The new asicminer chip is nice but finished products will most likely have around ~0,8W/GH/s.

This is nothing new, bitfury chips do that since Summer 2013, spondoolies does it in a nice 1,25u case etc.

Granted Asicminer chips have more GH/chip than Bitfury, but finished devices that are rackmountable etc. will take some more time until they are developed.

What it will all come down to for ASICminer is how cheap they can sell chips, which seems to be heading in a good direction.

Bulk chip sales will most likely be the easiest way to get some quick profit, as it also requires little own funds to be invested, as chip design is already done.

Hopefully ASICminer will be able to rake in some cash with this chip and we will then see a nice chip-battle on 28nm by the leading companies in Summer/Autumn of this year.

I find it interesting that regardless of process node with all of the various ASIC makers; working independently, in this present generation not one really stands out in the crowd as being tremendously more efficient. It gives the appearance that they may be encountering similar issues with power reduction and there may be physical limitations at play. If one were to extrapolate meaning out of this assumption, it would likely equate to lower profitability in future generations of ASICs as the demand to replace ASICs would be less since although more efficient, they may cost significantly more than ones who's R&D is already covered. I'm hesitant to speculate on the profitability of a forthcoming chip achieving significantly greater efficiency in order to justify those costs in light of the physical evidence currently out there.

If one looks at what's happening in the market, there used to be a requirement to buy very large scale rigs to obtain the lowest cost/Gh. Bitmain has been able to cut the price on the miners to the extent the S1, a relatively small unit, is at the present moment arguably one of the most cost competitive miners. It demonstrates the ability to sell an inferior ASIC in volume and even incorporate it into small scale miners as long as you can afford to hit a price point on your equipment. It does so in spite of what is offered is technically inferior to what others offer.

going by my experience with them to date, yes.


I have spread no such thing. What rockminer set out to raise and what they actually raised, were two different figures. Please take proper time to read my comments before thrusting out with irrationality to the debate.
https://bitcointalk.org/index.php?topic=528464.0 (https://bitcointalk.org/index.php?topic=528464.0)
also, please accept my apologies,
I merely recalled what I had read and I recalled wrong now i revisit the thread - 7600BTC 7500BTC.
it's been a whole month since the thread.

Well, in the long term when the reward of mining will be a very low bonus on top of the cost of electricity, small changes in mining efficiency will not be enough to justify the purchase of a new miner.

Therefore, my conclusion would be, get an efficient miner now or within the foreseeable future and hold on to it, as soon "new" mining hardware, unless dirt cheap, will not ROI.

i agree with some of your points... although antminer has sold plenty of their cheap & cheerful s1's, i don't think people are buying them cos its the most efficient miner (it isn't)... i think they're simply buying them because its an affordable entry level miner (but running costs make it one of the more expensive to run in its category as its more than 2w/gh)

its not any one thing that makes one system more efficient than another.  there's a lot of variables...

So according to you

15,000 ipo shares at 0.1 per = 7500btc?

And 7gh> 12gh?

Is this not spreading misinformation?

go buy a calculator.  :D

Efficiency comes into play when you're paying for hydro as I am. I recall when selling cubes one respondent in an email saying something to the effect he was an IT guy and wasn't concerned about hydro. I can imagine there's likely more than a few of those around.  ::)

The other issue with the S1 as was the case of the cube is the cost of the PSU is deferred to the purchaser giving some skewed results. Since some 20% of the market is Ants IMHO a good marketing strategy would be to build rigs suited to the existing PSUs and continue to defer this cost in addition reducing shipping costs, UL or similar approvals costs and show a cost savings to those who are already prepared to accept the equipment. This made the S1 appealing to me when upgrading from cubes. 20% running existing PSUs is a significant number.

I said that i don't think people will invest 50M$ in 2 months. bitfair said that i'm wrong, but also that he thinks that people will invest the same amount of money, but in 5 months. How can i be wrong if he thinks that it will take 5 months when i was saying that it won't be possible in just 2 months? Nobody is claiming to be proven or verified, it's just what we think. Yes no one can be proven right right now, but if i come here and state that i don't think one thing and you come and tell me that i am wrong and then you also tell me that you think that it will be possible in 5 months then you just let me win the argument.


Are we comparing power consumption at chip level or system level? You see to mix things up. Please show me proof where AM chips need 0.55W/GH. I haven't see it until now even if they have chips for almost one month now. Very nice derailing, stating that there is no chip that has that w/GH then saying that Spondoolies chip is half GH per chip but no mention of power consumption. Very smooth, but also very noticeably. Spondoolies chip is 0.8W/GH at system level and also 0.58W/Gh at chip level. And they are hashing since March. I don't see any AM chip in the wild hashing right now.

Check the specs from rockminers testing. They have hashing samples in hand.

To be clear about the specs

SP = 4w/7gh = 0.58w/gh
AM = 6.37w/11.5gh = 0.55w/gh

Please reread what you've quoted.

Yeah and AM comes up first in the phone book.  ::)

You cannot be seriously stating that the remaining 60,000 shares have ZERO value?
we are talking about IPO value are we not..? I've reread it and unless you have some sensible reason to ZERO-value 60,000 shares in an IPO then you need to simply accept the overall value, as I have stated in the first instance, was indeed 7500BTC - please advise the VC's or company bigwigs that their remaining 60,000 shares are totally worthless, because I wouldn't dare to think it (but of course, you may be right, who am I to argue?)

You are confusing total shares with IPO shares.

wrong, I have valued remaining shares at the same value of the IPO shares, you have simply ZERO-valued them.
I seriously hope that you do not run a business.

the difference in system performance between ones that utilise 0.58w/gh asics and those for 0.55w/gh aren't significant, and the final system power consumption 'at the wall' will also depend on other factors like how efficient the power supply used, the cooling system, the dc/dc conversion circuitry, the controller etc... so they're pretty much equal at the asic level and may well be equal at the wall too (except spondoolies is already shipping).. and we know that spondoolies has done the rest system design efficiently, and we don't yet know who the system integrators for asicminer's asics will be nor how efficient their systems will be 'at the wall', but lets presume they're about the same.

Agreed. The main difference will probably be the production cost/gh. Either way these are competative chips.

Do you talk shit just for the sake of talking shit?

I'll spell it out for you; why a company does an Initial Public Offering;

The purpose of an Initial Public Offering is to raise funds and to set the value of all shares.
Public shares are given a value, relative to the full amount of shares available, and the desired overall value sought by the company.
'Private' shares may be shares held by the company, or offered to VC's, or other investors at the value relative to the share value within the IPO.
In this, these shares are part of the IPO (in actual fact, in this case, the most part)
Overall value is calculated by public shares + private shares. Both types of shares are still valid in that 'Initial Public Offering.'

I reiterate - the IPO for rockminer was set out to raise around 7600BTC 7500BTC

what part of this sentence do you find difficulty in understanding? and please, refrain from personal attacks in future.

Do you disagree that raised value =/= total company value?

'Overall value' in this case is for a product, not the company. The rockminer thread makes this quite clear (as I read it). For example, say Spondoolies do an Initial Public Offering based on their own Gen3 chip, then people buying into that IPO would be buying shares in the production of the Spondoolies-Tech Gen3 chip, not the overall company. I hope that this helps how you understand the purpose of an IPO.

Thank you. After bitfair prove all by himself that i was right with my statements now you are the second one to do the same. AM's gen3 chips have nothing special when comparing with SP-Tech SP10 chips when it comes to power consumption. We still don't know what price they will have and we still don't have measurements for a full system. So what's so important about them? The density? You are only talking about chip density since we don't have any full system available yet. Let's see if they can pack 1.4TH/s in a 1.25U box. Even if they do it they will be a couple of months later to the game.

What if:

1. Netgear / Cisco / DLink / Zyxel / TPLink/ Asus / Belkin / a bunch of router / switch manufacturers imbed a chip in their hardware, and lower the price of their products, so more people buy them. Most people leave their routers on, all the time.

This will be the norm in the future.  Anything that has an internet connection and power will have a swappable card that has a Bitcoin ASIC on it.  Btw, that appliance/hardware may be close to "free" if you let the manufacturer mine.

They just better make sure that no one hacks the hardware to mine to a different pool or a different address. Happens all the time with console boxes (Sony Playstation, Microsoft X-Box, etc.).

I'm in for a free router. Why not? As long as it's still 10 watts or something.

Don't most people get their routers from the ISPs (most as in the majority of all people that use internet, not the majority of geeky people)

I often make a joke to my wife about my cooker mining bitcoins... once you switch off the oven, the oven fan continues to run to dispel the excess heat, it sounds like a little mining rig... but in fact, it's not too far from possible, stick in a wireless transmitter, embed the f/w and there you have it, roast beef and bitcoins.  :D

Miners should come with a Teflon top so they're dual purpose. ;D They kept my basement fireplace from coming all winter. Have to find something useful for them to do with the heat in summer. If I can figure out a way for them to heat the pool I've got it covered for all seasons.

how many guys here have received the first batch of AM 3rd generation chips ? It seems so quite here and  just feel nothing happened still...

yea i believe there was some sort of non disclosure agreement between FC and buyers so that until they launch their products, no one has a clue on FC's order book depth and sales.. but bear in mind that batch 1, 2 & 3 combined are supposedly > 10 Million chips.  :o  :D

Keep in mind that those numbers where agreed on following the original chip specs.

However, currently delivered chips get a maximum of around 0,8W/GH in completed devices, so the actual order amount could be considerably lower, as the chips are way off-spec.

I don't think any chips were sold before AM had them in hand and tested.

https://bitcointalk.org/index.php?topic=622439.0

No AM gen3 complete device showed any number below 1.1 W/GHs

These numbers are derived likely running them near max capacity while unit costs are high. AM should have a long term price advantage to get more ASICs into systems when the economics for this requirement come into play. Unfortunately we haven't the full operating range data yet to see what opportunities exist in that regard.

There's a sucker born every minute, your going to have to find them to buy this under performing tech.

still waiting on dividends?  :-\

Some questions...

What is the behaviour if task address is changed while the chip is busy hashing?

What about when address 44 is changed?

How is nonce_mask to be interpreted?

Is there any safeguard against a race clearing r_ready after reading nonce(s)?
For example, the order of events:

• Host reads nonce from chip
• Chip finds new nonce
• Host clears r_ready

I think you may be underestimating the power of the locomotive that is driving this program. Based on what I know about the chip, even if it is a bit of a 'dog' in terms of power consumption, it's small enough to yield in the 90's of percent and will probably cost less than $2.00 to make, test and package in high volume. A 4TH system using 512 of them -running in near threshold mode - could probably be built at a cost of around $1700, including the PSU's. If 'someone' had access to about $15m in liquid capital, they could procure about 6000 of these machines - 3 million chips, or 50 wafer batches, rent a big warehouse in Washington state, pay a few million in setup costs and installation of a 50MVA substation then that could easily be 24PH on hand within 6 months from a standing start (which may have started 2 months ago); with cheap electricity at less than 4 cents/kWh, power consumption becomes a lot less important.

Let's say that difficulty in October is 30 x 10e9 and BTC at $450. This installation would earn about $4.5 million in it's first month, even if the rigs and their cooling gobble up power at 1.2 Joules/gigahash. The 'suckers' would break even in just over 6 months allowing for running costs, but then they might decide to buy some more capacity - they can afford to and there's plenty spare capacity in that substation (and on TSMC's 40nm lines). Once hashing capacity is built and installed and has paid back it's manufacturing, installation costs and interest it can keep on hashing until the power and running costs get too high, and that might take a year or two - think an ROI of over 500%.

Individual miners do not have these resources, even the ones who have 'free' power (whatever that is, it won't last for long once their landlords start realising why their energy bills are so high).

Summary: don't underestimate your competition.

Money don't grow on trees! Wake up!

I think you're fooling yourself if you truly believe that. However, that''s your perogative.

Someone with 15m$ in hand should be better designing their own chip than to buy the underperforming AM chip anyway.

Why bother re-inventing the wheel, spending probably $3m to get someone to do it for you and missing probably 6-9 months of earnings? It's a no brainer, even with an inefficient solution. With the Bitcoin price rebounding  (for now) it looks more and more attractive.....

Less.

1) 2) When the chip is busy hashing the values in task addresses (including 44) cannot be changed.

3) Each 1 in the nonce_mask indicates a nonce for the current job. If the number of nonces exceeds 4, the first ones will be dropped.

4) There aren't safeguards.

For 1) and 2) we suggest a higher SPI clock for better efficiency. Using the difficulty setting larger than 1 could alleviate most hashrate losses caused by 3) and 4).

Any chance I get an answer to my mails and PM's about sample chips , datasheet , pricing etc.

Hmm, is there any way to tell the chip to abort processing a work then?
Or we just need to wait it out?

The soft reset can be triggered no matter if the chip is working or not. After that you need to reconfigure the PLL though, which will take about 0.2ms.

Will these chips ever be available to the public; in reasonable quantities ? 
I'm sure www.digikey.com/ or www.jameco.com and others, could sell a fair number.
Any resellers planned for this summer ?

it is planned

canary, if thats not too indiscreet, from whom did you already ordered AMgen3 based products? Rockminer? will you also be reselling AM chips ???

rockminer.  chip sales channels are being developed. ofcourse, a large order can be placed with ASICMiner directly at any time.

I got tracking  so sample chips are on the way to me :D
Some info about chip sales chanel  in the range of 100 , 500, 1k,  and so would be useful

WE are finishing our design of a 16 chip mining board with ASIC miner BE200 chips

specs
Our own design HEX 16E miner main specs:
 - 16 BE200 chip board in 2 rows
 - 16 bit Microchip controller with usb control
  - Voltage  to be adjustable by command in the PIC
 - Multi channel power supply
 - power connector PCI-e
- capable to run Overclocked

http://s7.postimg.org/5my63bimj/Placed.png


*greetings BICK :P

Here is hex16E ( AM Be200) 16 chip board populated.
Let the sleeples nights with tests begin :D

http://s28.postimg.org/56yhs3j4t/20140702_165943.jpg

is there a chance for beautyfull miner from technobit finally?

That looks like quality work marto74!

I hope the different BE200 orientation per row is intentional.

(did you notice the placement of many caps is V style?)

Yes "V style" because it is a test board and we left room on the stencil for bigger 1210 caps .
At the moment there are 0603 470nF ones, that can be exchanged with 1206 1uF or 1210 upt to 220uf if needed.
Usualy on the boards we put a lot of pads that way we can exchange components if there is trouble during the test or in production if the there is no stock.
Same way is with part of ICs we can put 2 different footprints of transistor drivers or 2 different footprints of 3.3 and 5 V regs.
The when the production starts we make anew stencil only for the components we are about to place

Test Board
hashing
http://s30.postimg.org/47cqmkm0h/160_GHs_200_W.png

This ID is in charge of the sales of AM-GEN3-MINERS.

Our official-miners are coming soon.

All resellers interested in it please msg us for more info.

congratulations!

Phasebird,

Are you an official employee of bitfountain/Asicminer?  Are you the community relations person we have been waiting patiently (well, mostly patiently) for?

If so, please introduce yourself on the main thread... We have tons of questions.

He is.

https://bitcointalk.org/index.php?topic=715382.msg8079873#msg8079873

Perhaps I could have worded my post more clearly.  Shareholders have been asking for Friedcat to appoint a community liaison for a long time.  It would appear that we may finally have been granted one, but yet it is not widely known and that liaison has not introduced himself to the main thread.  It would appear that not many know he is even here.

This is not the thread for this discussion so I will end by simply suggesting that Phasebird make his position known and give us some idea of the scope of his responsibilities.  It is possible that he is simply as friedcat said, responsible for the re-seller communication and questions regarding the meetup.