CoinTerra announces its first ASIC - Hash-Rate greater than 500 GH/s

[Minor Miner]

The Jupiters are around 1W per GH/s at the wall (more or less depending on the firmware used, ambient temperature, etc.).

Big deal.
KnC, 28nm, 1w per GH/s.
Bitfury, 55nm, .9w per GH/s.
KnC's chip is nothing to brag about.

How does anyone calculate KNC at 1W per GH/s?
I have serveral.   They are all hashing between 640 and 700 but average around 650GH/s.   They all are drawing 850-900W on average.   I would be VERY happy if they would draw 1W per GH/s because then the 850W power supplies I was told to buy would have worked instead of getting told two weeks later to go buy 1200W PSUs and paying to RMA the 850W ones.

[DPoS]

Do you lack basic reading comprehension skills or basic engineering skills? He clearly mentioned the chips are overclocked, thus they will never be as efficient as an underclocked one. The Jupiters are around 1W per GH/s at the wall (more or less depending on the firmware used, ambient temperature, etc.).

that is correct - they can also be underclocked for when efficiency and lower heat is desired (summertime, higher diff) or simply just running their old .95 firmware puts them at lower than 1w/gh

My point was that cointerra's customers blindly sat in the dark accepting all their fantasies while an honest group like KNC got raked over the coals for every word they said

Cointerra stressed that these are designed from the ground up for the 'long haul' with all these pro-team credentials and I've seen this type of bullshit forever on whiteboards and come up woefully short over and over again, then just change their expectations and launch and cheer themselves on.

wish some of you listened when I was barking earlier in the thread... demand real compensation is what you should do now, not listen to them try to continually pat themselves on the back with revised expectations

oh well

[aerobatic]

The Jupiters are around 1W per GH/s at the wall (more or less depending on the firmware used, ambient temperature, etc.).

Big deal.
KnC, 28nm, 1w per GH/s.
Bitfury, 55nm, .9w per GH/s.
KnC's chip is nothing to brag about.

How does anyone calculate KNC at 1W per GH/s?
I have serveral.   They are all hashing between 640 and 700 but average around 650.   They all are drawing 850W on average.   I would be VERY happy if they would draw 1W per GH/s because then the 850W power supplies I was told to buy would have worked instead of getting told two weeks later to go buy 1200W PSUs and paying to RMA the 850W ones.

same with the bitfury's... my bitburner bitfurys are hashing at 80 GH.. and consume more than 1watt/GH.   will come back with an exact reading later.

[Minor Miner]

The Jupiters are around 1W per GH/s at the wall (more or less depending on the firmware used, ambient temperature, etc.).

Big deal.
KnC, 28nm, 1w per GH/s.
Bitfury, 55nm, .9w per GH/s.
KnC's chip is nothing to brag about.

How does anyone calculate KNC at 1W per GH/s?
I have serveral.   They are all hashing between 640 and 700 but average around 650.   They all are drawing 850W on average.   I would be VERY happy if they would draw 1W per GH/s because then the 850W power supplies I was told to buy would have worked instead of getting told two weeks later to go buy 1200W PSUs and paying to RMA the 850W ones.

same with the bitfury's... my bitburner bitfurys are hashing at 80 GH.. and consume more than 1watt/GH.   will come back with an exact reading later.

Thanks, I was going to mention the bitburners too but the KNC was the obvious BS post by people who have no idea what they are talking about.   
Con has also posted that HF is just over 1W too.  1480W and 1.4TH/s.
I actually think this likely makes HashFast the most energy efficient miner currently until CT can redesign their board.  Still NOONE below 1W per GH/s at the wall.   (sorry D&T, I know W is not the correct way to measure).

[kendog77]

The Jupiters are around 1W per GH/s at the wall (more or less depending on the firmware used, ambient temperature, etc.).

Big deal.
KnC, 28nm, 1w per GH/s.
Bitfury, 55nm, .9w per GH/s.
KnC's chip is nothing to brag about.

How does anyone calculate KNC at 1W per GH/s?
I have serveral.   They are all hashing between 640 and 700 but average around 650.   They all are drawing 850W on average.   I would be VERY happy if they would draw 1W per GH/s because then the 850W power supplies I was told to buy would have worked instead of getting told two weeks later to go buy 1200W PSUs and paying to RMA the 850W ones.

same with the bitfury's... my bitburner bitfurys are hashing at 80 GH.. and consume more than 1watt/GH.   will come back with an exact reading later.

Thanks, I was going to mention the bitburners too but the KNC was the obvious BS post by people who have no idea what they are talking about.  
Con has also posted that HF is just over 1W too.  1480W and 1.4TH/s.
I actually think this likely makes HashFast the most energy efficient miner currently until CT can redesign their board.  Still NOONE below 1W per GH/s at the wall.   (sorry D&T, I know W is not the correct way to measure).

October Jupiters that are not overclocked use ~1W/GH at the wall. November Jupiters are less power efficient and use ~1.3W/GH at the wall.

It's kind of sad that both Hashfast and Cointerra missed their power targets, and makes what Knc and Bitfury were able to deliver in October even more amazing...

[Syke]

I actually think this likely makes HashFast the most energy efficient miner currently until CT can redesign their board.  Still NOONE below 1W per GH/s at the wall.   (sorry D&T, I know W is not the correct way to measure).

A standard clocked Bitfury rig from MegaBigPower runs 400 GH/s at 350 watts at the wall.

[mikieb]

They are underperforming and they are taking a much larger power draw than was anticipated. I'm glad they're are trying to get them out as soon as possible, but they still need certification for next week and that is not a for sure thing. If you're a new inquiring customer please read my post that I wrote here http://forum.cointerra.com/threads/attention-to-all-inquiring-new-cointerra-customers.324/

I recently looked on their site to see what batch they were taking pre orders from and they are still claiming 2THs performance with the power draw that has not been changed, I think that this is completely misleading and unethical now that they have some tested devices in. Also the fact that they now are taking in May pre orders is ridiculous considering the fact that that is five months out and they don't even have any December 2013 batches out now, they need to focus on getting out the product to the people who took the most risk on them and somehow compensating them for their late, underperforming machines with a much higher power draw than was anticipated.

[JoseSan]

I'm concerned about the hash rate but way more concerned about the power draw, this higher power target makes it very difficult to power these-2100 watts each?!?  Especially for larger purchasers the difference in power bills will be astronomical.  The worst part is they originally targeted 1200w and then raised the goal to 1600w and we're still way off.  As someone with several units on order this is very concerning, now even the highest density colocation cabinets running three phase won't supply enough power...I'm going to have to redo all our power plans, this is awful.  Apparently I should have purchased the stupid terramine hosting package when it was $3500 but who knew they were going to miss the 0.6W/GH goal by this much become 1.2W/GH.  This is insanely stressful, sigh, this is almost as bad as the issues people had powering minirigs.

The problem is that power rates where we are will run 0.25 per kWh so this is insanely expensive to run now

I agree with testerx that the combination of underperformance *with* overpowered operation is quite alarming. I did plenty of preparation for the electrical setup so that day one would be hiccup-free (e.g. UPSs/Surge protectors). On their FAQ they have that "We recommend a UPS which can handle at least 900W per circuit, or 1800W total." Not only am I going to have to scramble to ensure fuses are isolated or upgraded, but also return the UPSs to get even beefier ones.

In the engineers' defence, they all have probably worked on chips that have a more dynamic, bursty power response (e.g. mobile chips). These chips are basically all gates switching, all the time, which is probably pretty distinctive in the 28nm world. Nevertheless, I'm quite annoyed at the latest news.

[Supercomputing]

I'm concerned about the hash rate but way more concerned about the power draw, this higher power target makes it very difficult to power these-2100 watts each?!?  Especially for larger purchasers the difference in power bills will be astronomical.  The worst part is they originally targeted 1200w and then raised the goal to 1600w and we're still way off.  As someone with several units on order this is very concerning, now even the highest density colocation cabinets running three phase won't supply enough power...I'm going to have to redo all our power plans, this is awful.  Apparently I should have purchased the stupid terramine hosting package when it was $3500 but who knew they were going to miss the 0.6W/GH goal by this much become 1.2W/GH.  This is insanely stressful, sigh, this is almost as bad as the issues people had powering minirigs.

The problem is that power rates where we are will run 0.25 per kWh so this is insanely expensive to run now

Yes, I am very disappointed also. This is going to be a very big issue for some enterprise customers who may now have to under volt and under clock these devices to maximize the use of their data centers. Cointerra has been going in the wrong direction since their initial estimate of 1.2 kW or better at the wall per TerraMiner IV. Also, they have not yet broken the $3 GH/s barrier, not unless they are refunding the difference.

As for me personally, the biggest mistake Cointerra made was bragging about their ability to perform within a small time frame before delivering a single TerraMiner IV. From one top Electrical/Electronics engineer to another, you should always let your customers do the bragging for you. I give them credit for having the best value in the market and I hope that they continue to have the best value.

So hopefully, some of you will receive your TerraMiner IV Bitcoin miners some time in February, unless of course, certification takes longer.

[newguy05]

i can't browse a website without tripping over cointerra's google ads, luring newbies to order may batch...just like the old bfl tactics.

This company is such a disappointment, behind all that hype, being a pro team, and polished newsletters. If you look at the actual content it's all bullshit, for the last 1 month they basically haven't done anything to improve the performance but drag time. Whatever they are releasing now is the same hardware that was already on the market last year, 28nm or no.  The whole point of having 28nm is the ability to hit 2+ TH and in one standard 20A circuit without tripping the breakers or burn down your house.  Cant believe those guys can say with a straight face - yeah just plug it into 2 separate circuits, no shit sherlock.

Fail on both watts + hash rate.

You might as well just ordered some antminer/bitfury and started mining 3 months ago, by the time the difficulty gets to the level where watt/gh difference would matter materially between cointerra and the 55nm, the boxes will be all obsolete anyway. And you would mined a lot more bitcoins with the 55nm if started 3 months ago, instead giving cointerra a 0% interest loan while they did the dog and pony show.

[aerobatic]

The Jupiters are around 1W per GH/s at the wall (more or less depending on the firmware used, ambient temperature, etc.).

Big deal.
KnC, 28nm, 1w per GH/s.
Bitfury, 55nm, .9w per GH/s.
KnC's chip is nothing to brag about.

How does anyone calculate KNC at 1W per GH/s?
I have serveral.   They are all hashing between 640 and 700 but average around 650.   They all are drawing 850W on average.   I would be VERY happy if they would draw 1W per GH/s because then the 850W power supplies I was told to buy would have worked instead of getting told two weeks later to go buy 1200W PSUs and paying to RMA the 850W ones.

same with the bitfury's... my bitburner bitfurys are hashing at 80 GH.. and consume more than 1watt/GH.   will come back with an exact reading later.

Now have accurate measuremens. My bitburner bitfury is hashing at 85 GH/s and drawing 125 watts at the wall. Not including the raspberry pi which is powered separately.  Total is 1.47 watts/GH

[plato14]

Cointerra has failed to live up to the expectations that they set for themselves and everyone else.  They need to ship immediately, so they don't fuck us over even more than they have now.   

Also I believe they need to change their product specs immediately.  I can't believe theyre taking delivery for the May batch now, hope no one is dumb enough to buy one of these.
http://cointerra.com/product/terraminer-iv-2ths-networked-asic-miner-may-batch/



The bitcoin mining community is the jaded girlfriend that keeps coming back for more.  Lesson learned...again, don't believe the pre-order bullshit hype.

[Easy2Mine]

i can't browse a website without tripping over cointerra's google ads, luring newbies to order may batch...just like the old bfl tactics.

This company is such a disappointment, behind all that hype, being a pro team, and polished newsletters. If you look at the actual content it's all bullshit, for the last 1 month they basically haven't done anything to improve the performance but drag time. Whatever they are releasing now is the same hardware that was already on the market last year, 28nm or no.  The whole point of having 28nm is the ability to hit 2+ TH and in one standard 20A circuit without tripping the breakers or burn down your house.  Cant believe those guys can say with a straight face - yeah just plug it into 2 separate circuits, no shit sherlock.

Fail on both watts + hash rate.

You might as well just ordered some antminer/bitfury and started mining 3 months ago, by the time the difficulty gets to the level where watt/gh difference would matter materially between cointerra and the 55nm, the boxes will be all obsolete anyway. And you would mined a lot more bitcoins with the 55nm if started 3 months ago, instead giving cointerra a 0% interest loan while they did the dog and pony show.

+10

I couldn't have say it better
There is no such a pro team, anyone that can pull it off designing an ASIC has some kind of experience in the semi conductor industry.
All teams have still a lot to learn from each others.

[Powell]

The real "pros" are usually quiet and come with everything ready to go and to prove itself.  It has always been that way in the car community and they build there customer base on "past experience" on other platforms or they have ONE thing they did that was never able to replicate, etc.  In the car industry I like being that quiet one but sadly no way I could ever build ASIC chips, could fund it but could not built.  Soldering is still a challenge for me LOL!

It cracks me up when all their bandwagon people have NO explanation as to why it didn't succeed.  Then out of left field the quiet giant rises and brings the pain with them.

Don't get me wrong Cointerra came out with a final product that does have some performance BUT it was nothing like it was originally hyped up to be.

[testerx]

The real "pros" are usually quiet and come with everything ready to go and to prove itself.  It has always been that way in the car community and they build there customer base on "past experience" on other platforms or they have ONE thing they did that was never able to replicate, etc.  In the car industry I like being that quiet one but sadly no way I could ever build ASIC chips, could fund it but could not built.  Soldering is still a challenge for me LOL!

It cracks me up when all their bandwagon people have NO explanation as to why it didn't succeed.  Then out of left field the quiet giant rises and brings the pain with them.

Don't get me wrong Cointerra came out with a final product that does have some performance BUT it was nothing like it was originally hyped up to be.

Yeah, if they had set expectations lower this would be fine but they were saying they would have far lower power consumption than KnC, which seems not to be the case.  Seems like bitfury did a really good job optimizing power use considering the 55nm nature.

[DPoS]

The other question will be how stable these Cointerra rigs will be and if you will be able to overclock/underclock

they have been dead silent and you'd think they would of put up a link to whatever pool they are hashing on to show results
and how about the heat?  with all the higher watts, was their airflow design robust enough to handle more?

[newguy05]

The real "pros" are usually quiet and come with everything ready to go and to prove itself.  It has always been that way in the car community and they build there customer base on "past experience" on other platforms or they have ONE thing they did that was never able to replicate, etc.  In the car industry I like being that quiet one but sadly no way I could ever build ASIC chips, could fund it but could not built.  Soldering is still a challenge for me LOL!

It cracks me up when all their bandwagon people have NO explanation as to why it didn't succeed.  Then out of left field the quiet giant rises and brings the pain with them.

Don't get me wrong Cointerra came out with a final product that does have some performance BUT it was nothing like it was originally hyped up to be.

Yeah, if they had set expectations lower this would be fine but they were saying they would have far lower power consumption than KnC, which seems not to be the case.  Seems like bitfury did a really good job optimizing power use considering the 55nm nature.

Bitfury underclocked their chips - higher clock more watts, just like gpu oc we all mess around with when young. This means  cointerra is already pushing their chips to the edge as 28nm(unless horribly designed) should have significantly lower watts than 55nm running normal. Yet it is almost on par and the hash rate is still only 1.7, they probably won't even achieve 1.5 if the chips are ran at normal clock.

Really not sure what is going on here, but from the newsletter with vague mention to pcb redesign for next release, my opinion is they fubar the current design horribly.   

[aerobatic]

The real "pros" are usually quiet and come with everything ready to go and to prove itself.  It has always been that way in the car community and they build there customer base on "past experience" on other platforms or they have ONE thing they did that was never able to replicate, etc.  In the car industry I like being that quiet one but sadly no way I could ever build ASIC chips, could fund it but could not built.  Soldering is still a challenge for me LOL!

It cracks me up when all their bandwagon people have NO explanation as to why it didn't succeed.  Then out of left field the quiet giant rises and brings the pain with them.

Don't get me wrong Cointerra came out with a final product that does have some performance BUT it was nothing like it was originally hyped up to be.

Yeah, if they had set expectations lower this would be fine but they were saying they would have far lower power consumption than KnC, which seems not to be the case.  Seems like bitfury did a really good job optimizing power use considering the 55nm nature.

Bitfury underclocked their chips - higher clock more watts, just like gpu oc we all mess around with when young. This means  cointerra is already pushing their chips to the edge as 28nm(unless horribly designed) should have significantly lower watts than 55nm running normal. Yet it is almost on par and the hash rate is still only 1.7, they probably won't even achieve 1.5 if the chips are ran at normal clock.

Really not sure what is going on here, but from the newsletter with vague mention to pcb redesign for next release, my opinion is they fubar the current design horribly.   

No question the BitFury design is very special.  But it's not just that it's in 55nm.. It's special because it was full custom and designed for very low power operation.  No other chip available has been designed in full custom before nor since for Bitcoin mining.

From what cointerra has said in their update. The asics can run at full rate but the power circuit needs a redesign to supply more power to them.  That doesn't say the chips are a disappointment.. If by redesigning a new board in a month or two they can reach the orig goals then the asics must be fine..   Right?

Meanwhile a t least they are going to ship to customers with what they have, despite it being less than the target.  Whereas some other ASIC companies decided to delay the product for months while they tweaked e design.  I know which one I'd prefer!

Anyway.. Im happy to take mine as is, as soon as possible.  And if you don't want your jan terraminer may I suggest you ask Cointerra to give u a refund.. Or better yet, I will happily buy it off you...!  Yes, seriously!  These are really good value even with a 15-20% underperform.. They're still the fastest and lowest cost miner you can get... Right?

[DPoS]

a lot easier to redesign expectations than to achieve them!!  Where is the 20% refund?  and why still false advertise on the site?

[Supercomputing]

@aerobatic
Please see my previous post. I like your optimism but you are now coming off as an opportunist.

@DPoS
I gree with you and I am one of Cointerra's biggest supporters. They will meet their goals some time in the future, but for now they will need to correct the misleading advertisement.

It is true that Cointerra has the best value in the market, but we still do not know when they are going to ship their first batch. They have not said anything new in the last three updates, not really. But if you've waited this long, the best thing to do is to wait a little bit longer.