Syke
Legendary
Offline
Activity: 3878
Merit: 1193
|
|
June 19, 2014, 12:07:42 AM |
|
KnC delivered on spec and on time. Which other company did that?
Quite a few. Avalon, BitFury, ASICMiner, Bitmain, Spondoolies.
|
Buy & Hold
|
|
|
samsonn25
|
|
June 19, 2014, 12:11:32 AM |
|
We are assuming 1 board per box. Perhaps there are 2 boards. Or by some amazing chance 2 connectors but they worded the announcement incorrectly.
|
|
|
|
xstr8guy
|
|
June 19, 2014, 12:25:40 AM |
|
We are assuming 1 board per box. Perhaps there are 2 boards. Or by some amazing chance 2 connectors but they worded the announcement incorrectly.
http://forum.kncminer.com/forum/news-and-anouncments/46242-the-first-20-nanometer-chip-devoted-to-bitcoin"The Neptune is a 1440-core 55x55 mm package tailor-made for the Bitcoin market. Each Neptune consists of 5 chips and comes in new sturdier enclosure and packaging. Using the speed offered by a true 20nm die process each Neptune churns out more than 3 TH/s while consuming only 2.1 KW at the wall. That works out to 0,7 watts per GH/s or less." Wow, that was hard to find! And a little bit more research and you find this... https://www.kncminer.com/news/news-89Notice the single PCI power connector?
|
|
|
|
samsonn25
|
|
June 19, 2014, 01:19:16 AM |
|
Yes. But that's the whole discussion. How can that 1 connector power up 450 + watts
|
|
|
|
Soros Shorts
Donator
Legendary
Offline
Activity: 1617
Merit: 1012
|
|
June 19, 2014, 01:33:27 AM |
|
Yes. But that's the whole discussion. How can that 1 connector power up 450 + watts
I believe the 6-pin Mini-Fit Jr can go up to 13A per circuit if the largest spec pins are used with the appropriate gauge wires. That is 12V x 13A x 3 circuits = 468 watts. Of course, that is pushing it and the PSU side of the connector might just burn out if not up to spec.
|
|
|
|
Bogart
Legendary
Offline
Activity: 966
Merit: 1000
|
|
June 19, 2014, 01:38:55 AM |
|
Yes. But that's the whole discussion. How can that 1 connector power up 450 + watts
It can't. At least not for very long. It's a seriously retarded design. The connectors don't always make perfect contact either. I have had to replace several of these 6-pin connectors on my AntMiner S1s after they fried themselves, and they only move around 190W each.
|
"All safe deposit boxes in banks or financial institutions have been sealed... and may only be opened in the presence of an agent of the I.R.S." - President F.D. Roosevelt, 1933
|
|
|
xstr8guy
|
|
June 19, 2014, 01:58:03 AM Last edit: June 19, 2014, 08:09:12 AM by xstr8guy |
|
Yes. But that's the whole discussion. How can that 1 connector power up 450 + watts
I was just attempting to correct your misconceptions. You didn't appear to have read the news updates so I thought I'd link/quote them. I didn't say what KNC is attempting is logical or feasible. In fact it might be their biggest mistake yet since they seem to be purposely releasing miners into the wild that could potentially be extremely hazardous.
|
|
|
|
faetos
|
|
June 19, 2014, 02:23:57 AM |
|
Yes. But that's the whole discussion. How can that 1 connector power up 450 + watts
I was just attempting to correct you misconceptions. You didn't appear to have read the news updates so I thought I'd link/quote them. I didn't say what KNC is attempting is logical or feasible. In fact it might be their biggest mistake yet since they seem to be purposely releasing miners into the wild that could potentially be extremely hazardous. BBQ Miners part deux! Maybe you all are getting some of these:
|
|
|
|
crashoveride54902
|
|
June 19, 2014, 02:33:49 AM |
|
And in other news: KNC is still screwing ppl...bought some hosted miners shipped to me from someone off forum...he says he has official statements and announcements from them that said they were upgrading them to nov. boards free of charge...and what do i receive...yep you guessed it Oct. jups oh and you'd think by now they would at least secure the flipping fan shrouds on after the first shipment hazard...nope...even some of the heatsinks where crooked...typical stuff that has been all over the forum for months now and with the neptombs...good luck with those if they didn't put another connector on the asic...if oct. jups OC'd can melt psu pcie connectors....you'll insta start a fire with these lol...well maybe not cause it'll probably just trip the psu before it would go poof...but then if they require 420w thur 1 pcie 6pin...how the heck did they test this? and why would they think this would work for the general public? I wonder if they'll ship with their famous pcie extenders on just to increase the odds of fire?
|
Dreams of cyprto solving everything is slowly slipping away...Replaced by scams/hacks
|
|
|
rograz
|
|
June 19, 2014, 02:50:08 AM |
|
Yes. But that's the whole discussion. How can that 1 connector power up 450 + watts
I was just attempting to correct you misconceptions. You didn't appear to have read the news updates so I thought I'd link/quote them. I didn't say what KNC is attempting is logical or feasible. In fact it might be their biggest mistake yet since they seem to be purposely releasing miners into the wild that could potentially be extremely hazardous. If they are only using 1 connector for 450W they better pull a approved PSU list out of their asses, the cables should be fine on most high end psus but the quality of the pci-e connectors can vary quite a bit as well. I might add that 450w is way out of the specs of the connector itself from molex, from what i've heard it's 11A per pin (not talking about the pci-e specs that are even lower), but if they are good quality they should be able to handle it (on a good day) I have personally run close to 400w trough them (but only short term when doing rather extreme OC)
|
|
|
|
dropt
Legendary
Offline
Activity: 1512
Merit: 1000
|
|
June 19, 2014, 03:40:07 AM |
|
Yes. But that's the whole discussion. How can that 1 connector power up 450 + watts
I was just attempting to correct you misconceptions. You didn't appear to have read the news updates so I thought I'd link/quote them. I didn't say what KNC is attempting is logical or feasible. In fact it might be their biggest mistake yet since they seem to be purposely releasing miners into the wild that could potentially be extremely hazardous. If they are only using 1 connector for 450W they better pull a approved PSU list out of their asses, the cables should be fine on most high end psus but the quality of the pci-e connectors can vary quite a bit as well. I might add that 450w is way out of the specs of the connector itself from molex, from what i've heard it's 11A per pin (not talking about the pci-e specs that are even lower), but if they are good quality they should be able to handle it (on a good day) I have personally run close to 400w trough them (but only short term when doing rather extreme OC) Table 4.3 - http://www.molex.com/pdm_docs/ps/PS-45558-001.pdfand http://www.molex.com/webdocs/datasheets/pdf/en-us/0457320001_PCB_HEADERS.pdf
|
|
|
|
RoadStress
Legendary
Offline
Activity: 1904
Merit: 1007
|
|
June 19, 2014, 04:34:18 AM |
|
It's not like you can get a 3Th/s Neptune now. If you buy a Neptune now you will get it in September too. There is nothing on stock right now.
Do you really think they will be shipping in september? They said they expect to ship batch 1 and 2 before the end of this month.. Right they will ship 2 batches in 1 week. You only need 0.028btc to afford datatank/giant mining operation rates.
It's a fair comparison because your customers need to spend $5,300 + $0.15/kwh where as datatank customers spend $17 for 18GH. Please show me hosting at $0.025/kwh available to your customers. I doubt you can even find below $0.05/kwh.
Yes someone could probably build an air cooled DC but even at $0.025/kwh immersion cooling would save ~$200,000/year in electricity costs alone. What will 0.028btc get you in order to afford datatank mining? 200k$ per year saved in cooling electricity costs right? What is the cost for the Datatank?
|
|
|
|
rograz
|
|
June 19, 2014, 04:40:43 AM |
|
The first one is just a copy of the the pci-e express power standards iirc, if we were limited by those standards the world would be on fire by now! Second one states 8A however so it's even worse than the info I had, could be that there are different pin configurations however with different ratings. Better pull out some LN2 and cool those connectors.
|
|
|
|
fubly
|
|
June 19, 2014, 05:07:33 AM |
|
|
each time you send a transaction don't forget to use a new address, each time you receive one also!
|
|
|
xstr8guy
|
|
June 19, 2014, 05:54:16 AM |
|
Be sure to stack dry kindling on and around those PCI cables to keep them cool. If you don't have any kindling on hand, oily rags will do in a pinch. Also be sure to have a full gas can venting into the surrounding area. Err, I mean... DON'T!
|
|
|
|
Collider
|
|
June 19, 2014, 05:57:15 AM Last edit: June 19, 2014, 04:27:32 PM by Collider |
|
Isn´t KNC late on their shipment?
Didn´t they say the "customer appreciation" batch would ship one month before end of Q2?
Also, 0.7W/GH on 20nm is a waste of Silicon, you can get that on 40nm or even 55nm.... It just makes for unreasonably high NRE.
|
|
|
|
xstr8guy
|
|
June 19, 2014, 06:07:16 AM |
|
Isn´t KNC late on their shipment?
Didn´t they say the "customer appreciation" batch would ship one month before end of Q2?
Also, 0.7W/GH on 20nm is a waste of Silicone, you can get that on 40nm or even 55nm.... It just makes for unreasonably high NRE.
Yeah, they said a lot of things. Most of them untrue. And for gods sake I hope they didn't use silicon e. That would indeed be a waste! I'm pretty sure that isn't semiconductor material, lol.
|
|
|
|
The Avenger
|
|
June 19, 2014, 06:21:46 AM |
|
Be sure to stack dry kindling on and around those PCI cables to keep them cool. If you don't have any kindling on hand, oily rags will do in a pinch. Also be sure to have a full gas can venting into the surrounding area.
lol Of course, that will void your knc 30 second warranty
|
"I am not The Avenger" 1AthxGvreWbkmtTXed6EQfjXMXXdSG7dD6
|
|
|
rograz
|
|
June 19, 2014, 06:21:51 AM |
|
Also, 0.7W/GH on 20nm is a waste of Silicone, you can get that on 40nm or even 55nm.... It just makes for unreasonably high NRE.
For the customer yes, but unless the KNC design is horrible they are using A LOT less die area building a 3TH machine than you would with 40/55nm due to performance density from node shrinking (each full node shrink is eta 50% higher transistor density iirc) add increased frequency headroom and lower power usage (comparatively, example running chip x on 28nm it uses y power at z frequency, shrink x to 20nm and power usage y at frequency z should decrease with expected -20%+). The fact that KNC are running these chips at 0,7W/GH could mean a number of things. 1. Another hard copy rather than custom design (worse density/effiency) 2. 20nm has leakage problems 3. The ships scales very well and tolerate high speeds, in other words they are clocked for speed rather than efficiency making it cost less for KNC to provide the ordered performance. It simply makes no sense for them to use up to 50% extra chips so they can ship at <0,5W/GH at this point.
|
|
|
|
samsonn25
|
|
June 19, 2014, 06:36:22 AM |
|
The downsize with die shrinking is more heat when the frequency is run up
|
|
|
|
|