Topic: CoinTerra announces its first ASIC - Hash-Rate greater than 500 GH/s - page 46. (Read 231002 times)

Well if they never deliver then refunding and getting your money now would be the best bet, maybe front run another price spike.  What is rational about trusting  a company that has lied to their customers already, and screwed them all over in the process?  You're putting your trust in a company that has already duped you, i don't understand why anyone would trust that they would deliver now.  Ill sell you five early batch january for 110 biticoin, thats a 50% discount on what i got them for.  LMK

Hi aerobatic

Since u seem to have bought the whole gamut of ASICs - KNC, HF, CT etc, it will be interesting to hear your views about how they compare, once you have both the HF and CT machines delivered and are then able to compare all of the performances side by side. I hope you will continue to stay neutral in your commentary as that's what makes your posts worth reading, tho I seem to have detected of late a bit of CT favouritism?

plato14 -

people make rational decisions based on their own self-interests.

asking for a refund, when you can't buy anything better, cheaper, nor faster... is what we call cutting off your nose to spite your face.    customers don't do things like cancel their order, just to fuck with the manufacturer.   especially when the manufacturer won't be affected and the customer loses out.   thats not useful nor well thought out advice.

cancelling and asking for a refund probably won't hurt cointerra either.  the boxes could easily be resold.  they're still the lowest cost $/gh of anything available.

so...  I'm not advising to cancel (my own personal opinion is to take delivery and there's upside potential if they improve performance with tweaks over time).

... BUT, despite the fact i don't think its a good idea... (shameless self promotion admitted here), if anyone is actually seriously considering cancelling their january cointerra order, please pm me as i would like to see if we can work out a deal to buy your miner(s).  price depends on many factors.. like where you are in the line, and how many you've got on order, but it will most likely be more than you paid for it (in dollar terms of course)

Cointerra is the best Bitcoin mining hardware manufacturer out there; and the name calling is completely unjustified and unnecessary. It is obvious by now that they are aware of how their delays and hardware performance issues will affected some of their customers. And they are taking action to improve relations with their current and future customers.

So Congratulations to Cointerra for their accomplishment in bringing the best Bitcoin mining hardware to the market in terms of value and engineering. I am always happy to see other fellow engineers push the limits and accomplish great things.

I will be waiting for Cointerra's next generation hardware - the 14nm 3D-structured designed ASICs. Now that would be something worth investing thousands of BTCs in.

i think all January batch customers should just get refunds, i don't think theyre going to be on time,  why not just fuck them over and flood them with refund requests.  Either that or sue for misleading customers.

Way too many words to repeat what I said in a single sentence.


 I expect they will be able to improve on the board design and capture both their original target performance and substantially improve upon the power consumption.

They didn't really compensate the December batches more than you, if anything December units are the worst off because they cost $14,000 which is $2000 more than two January units and we don't even get the second unit until March.  The people least screwed are probably the January units lol

we're way off topic.  suggest you ask your questions directly in the pkr forums if you want to take this any further as this has nothing to do with bitcoin (let alone cointerra)

as for who licenses pkr...   it holds gaming licenses in the UK, Alderney, France, Italy, Denmark and a few other places.   Each one has different rules and we follow them all.  its actually difficult to have one company that follows so many different rules, and this takes up a lot of our time.  all the executives hold licenses as well.  and shareholders too (like me).  each person individually investigated and licensed. 

http://www.pkr.com/en/support/licensing-and-integrity/licensing-and-regulation/

i know this is quite simple, but if we consider the four big things in their box...  there's some things they can tweak and some that they simply can't...

1.  the asics.

2.  the boards.

3.  the cooling system

4.  the power supplies.


so, if we look at each thing in turn...

1... the asics.   these aren't going to change.  they are what they are.  it costs millions and too much time to change them so lets assume thats never going to happen.
   but we've heard ct say they ran faster in the lab when they gave them more power, and that they expect that with board and software tweaks they will go faster in the production boxes... so i think we're good on the asic front.

2.  the boards.  they said they're doing a full board redesign to improve power conversion efficiency.   ergo, the power consumption at the wall probably won't increase by much, but the power delivery to the asics will increase when they can improve conversion efficiency... thus the asics could potentially go faster with little to no impact on the rest of the stats.    thus if they hit 2 TH with 2KW... it will then have reasonable power consumption in the same ballpark as their peers.  no asic except an under clocked bitfury gets less than 1W/GH, and under clocking anything isn't commercially justified, which is why no one does that.

3.  the cooling system.  this seems very rugged and powerful. I'm not expecting they need to do anything to improve the cooling system (except perhaps make the fans temp sensitive in the future to reduce noise).  though recognising that this system draws 2 KW, so it has 2 KW of heat to exhaust.  no getting around that.  A 400 GH babyjet or 600 GH jupiter is gonna be quieter cos its got a fraction of the wattage (and thus heat) to exhaust.

4.  the power supplies.  the ones they're using are quite compact.  I'm not aware of any other similarly sized power-supplies (1100 watts in a 1U server psu, twice - is  lot of power in a small space)... thus if you want to feed the asics any more power than these can muster, you will have to forgo the internal power supplies and do what knc does and use external atx psus.

lol rape victims don't willingly submit themselves to the process.

where did you get the picture from?  

[edit.. ah, ok, its on ct's web site.. sorry!]

[another edit:   you also need to take account that they have more than one factory making the boxes]

Pre-orders leave the customer vulnerable to being cheated.  That doesn't justify cheating the customer.

Do you spend a lot of time talking to rape victims about what they were wearing when they were attacked?

I'd extend the analogy further. The car is designed to be involved in a permanent competitive race, where you can enter as many cars as you like to win more prize money from the fixed overall pot. I'd want my underspecced car sooner. But I'd also order from the 2nd production run of a prototype design, unless that prototype was for a new deign paradigm.

If you're upset with your own investment decision, then welcome to the world of buying pre-orders for custom mining hardware. You have to be lucky or judge carefully to make a good call, if so many people weren't making quick decisions then there wouldn't be so much wailing and bawling in these threads. You can see the wailing and bawling for yourselves, why put yourself through that? The only choice you have is to accept any loss quickly with dignity, or drag it out for as long as possible with the accompanying misery and frustration.

You're missing one very important variable in this analogy TIME, if the longer time went by for them tweaking the car more useless it would become.

If wishes were horses...

Unfortunately the power consumption and performance are coupled together.  Cointerra targeted a very high clock rate, and designing boards for that is not easy.  I expect they will be able to improve on the board design and capture both their original target performance and substantially improve upon the power consumption.  But that is going to take time, and it makes no sense to hold off on shipping a functional design to gain 20% when difficulty goes up by that much every 10 days.

I still see them as the most professional organization in the market today.  Organizations are tested when things don't go to plan.  What Cointerra does in the next few weeks will define their reputation forever.  Andy Grove saved Intel's reputation by making the tough decision and recalling the CPUs with FDIV defects.  Ravi is going to have to make a similar decision here.

I only wish the reduced rate had come with better than anticipated power usage instead of the reverse.

I think people who has only 1 order should be ok with 1.6TH until they improve the speed, but think of people who has 4 or more orders, loss of 400GB per unit equals to 1.6TH that's alot.

 I can see approximately 250-300 units. That's about 1/3 of the December order queue that we can (sort of) confirm physically exist. They pushed the button to start production sometime last week, so perhaps the last December orders will have to wait another week or two. I'm hoping I'm wrong, but that's a pretty good worst-case guess given the image.

I'm Batch 2 and I just got screwed by the compensation. Cointerra seems more like " CONterra"
For this to be fair it has to be just as Dec orders we should be offered a TerraMiner II from a later batch like April or May, since we are getting -20% less Hashing and 20% more power usage. 

No they did not, is this an offer?  They simply gave me a March order as well so if they are offering conversions to January then they owe me about 2000 usd since I paid 14k for a December unit which will apparently be underperforming