Pages:
Author

Topic: CoinTerra announces its first ASIC - Hash-Rate greater than 500 GH/s - page 51. (Read 231002 times)

hero member
Activity: 702
Merit: 500
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

hero member
Activity: 1316
Merit: 503
Someone is sitting in the shade today...
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.
sr. member
Activity: 278
Merit: 250
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.
member
Activity: 117
Merit: 10
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.
newbie
Activity: 1
Merit: 0
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.
legendary
Activity: 3878
Merit: 1193
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.
hero member
Activity: 742
Merit: 500
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...
legendary
Activity: 2478
Merit: 1020
Be A Digital 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).
hero member
Activity: 702
Merit: 500
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.

sr. member
Activity: 462
Merit: 250


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
legendary
Activity: 2478
Merit: 1020
Be A Digital 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.
legendary
Activity: 1428
Merit: 1000
https://www.bitworks.io
I am quite surprised that after missing there performance and power target by so much they didn't attempt to do anything. For example suggest the redesigned board will be provided at no cost to all units when it is available, something to make people feel like they may not have been cheated quite as much..

I wonder if anyone is going to attempt to slap them with a lawsuit given that it's treading on FTC territory with false advertising. I haven't bothered looking at the fine print so perhaps Cointerra has themselves covered..

hero member
Activity: 608
Merit: 500
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
hero member
Activity: 1316
Merit: 503
Someone is sitting in the shade today...
Quote
1.63TH/s and 1.72TH/s with power draw at the wall between 1900W and 2100W – typically 20% higher than our anticipated target of 1650W but still sufficient to operate in a typical US home (110V) plugged into two outlets on separate 15A circuits

What a joke...not even compensation

They know the buyers are so desperate for shipping they will take anything.

member
Activity: 64
Merit: 10
LTS Scrypt Asic Preorder :P
Email just received from CT.

Quote

DATELINE 1/24/2014

Today we wanted to communicate our plans as we move from pre-production into production. As soon as the FCC/CE/UL certification process is completed we will be able to commence shipping the very first production TerraMiners and incrementally ramp up to volume capacity over the coming days.

On the hardware front, we have implemented circuit board modifications which have improved the efficiency of the power conversion circuitry from the time of our initial board bringup and we have revised the firmware on the BeagleBone Black controller card to take advantage of this.

Having achieved a good level of stability and based on feedback from our customers, we have decided to push the button to release our first production TerraMiners IV’s.

Our production lines are assembling units and we have submitted for certification simultaneously. The production units that we have assembled and tested so far are running at between 1.63TH/s and 1.72TH/s with power draw at the wall between 1900W and 2100W – typically 20% higher than our anticipated target of 1650W but still sufficient to operate in a typical US home (110V) plugged into two outlets on separate 15A circuits, or in a typical European home (230V) on a single circuit.

The first-batch of TerraMiner IV units shipping to customers

    Is easily the most powerful high-performance Bitcoin miner available today. It smashes the 1TH/s barrier by a huge margin
    Is the lowest cost per gigahash of any Bitcoin mining rig on the market
    Achieves the highest density of Bitcoin hashing power in its compact 4U enclosure

We will continue tweaking the firmware and any updates will be posted to the CoinTerra website.

Additionally, we have begun a board redesign which will enter production when fully tested and approved. It is anticipated that the board revision will offer further hardware optimizations and power efficiencies and based on the initial simulations it will permit us to hit our initial performance target of 2TH/s.

We expect to post the next update as soon as we’ve received certification early next week and begin delivery.

http://cointerra.com/update-engineering-production-status/
legendary
Activity: 3878
Merit: 1193
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.
sr. member
Activity: 486
Merit: 262
rm -rf stupidity
that's a terrible effort by the PROS OF ASIC with a design FROM THE GROUND UP!!  (not just some slapped together KNC rush job)

I run two jupiters with two expansion boards overclocked mildly ~1.5Th @ 1800watts at the wall
so the big scary clownterra is doing basically the same as my rigs but 4 months after KNC 


 Tongue CLOWNS  Tongue

Compare both to Bitfury's 55nm chip well under .9 watts per GH/s at the wall and the C in KnC stands for clown also.

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.).

Also, the clowns are those over-promising and under delivering. And Bitfury are also clowns in my books because they'll sell to anyone without giving a damn as to how that will affect their smaller customers. Buy only from the companies that only sell to end users and not to big mining consortiums.

What? You are seriously mad at bitfury for selling miners to mining groups? What do you think knc is doing all these months? What miner do you see at this farm?
http://www.nytimes.com/video/business/100000002607442/mining-for-bitcoins-in-iceland.html

Hundreds upon thousands of USB miners but I may have been alt+tabbed and guessing.
sr. member
Activity: 434
Merit: 250
that's a terrible effort by the PROS OF ASIC with a design FROM THE GROUND UP!!  (not just some slapped together KNC rush job)

I run two jupiters with two expansion boards overclocked mildly ~1.5Th @ 1800watts at the wall
so the big scary clownterra is doing basically the same as my rigs but 4 months after KNC 


 Tongue CLOWNS  Tongue

Compare both to Bitfury's 55nm chip well under .9 watts per GH/s at the wall and the C in KnC stands for clown also.

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.).

Also, the clowns are those over-promising and under delivering. And Bitfury are also clowns in my books because they'll sell to anyone without giving a damn as to how that will affect their smaller customers. Buy only from the companies that only sell to end users and not to big mining consortiums.

What? You are seriously mad at bitfury for selling miners to mining groups? What do you think knc is doing all these months? What miner do you see at this farm?
http://www.nytimes.com/video/business/100000002607442/mining-for-bitcoins-in-iceland.html
full member
Activity: 173
Merit: 100
that's a terrible effort by the PROS OF ASIC with a design FROM THE GROUND UP!!  (not just some slapped together KNC rush job)

I run two jupiters with two expansion boards overclocked mildly ~1.5Th @ 1800watts at the wall
so the big scary clownterra is doing basically the same as my rigs but 4 months after KNC 


 Tongue CLOWNS  Tongue

Compare both to Bitfury's 55nm chip well under .9 watts per GH/s at the wall and the C in KnC stands for clown also.

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.).

Also, the clowns are those over-promising and under delivering. And Bitfury are also clowns in my books because they'll sell to anyone without giving a damn as to how that will affect their smaller customers. Buy only from the companies that only sell to end users and not to big mining consortiums.
hero member
Activity: 784
Merit: 1000
I knew the certification would take longer than 3 days.  While a different industry, my experience with UL & CE approval wait times took months.
Pages:
Jump to: