Author

Topic: Swedish ASIC miner company kncminer.com - page 321. (Read 3049515 times)

legendary
Activity: 3234
Merit: 1220
I edited my post and added link to their news
Go look yourself.

I read the news, but it's simply stupid. The demand for Neptunes CAN'T be bigger than the wafers because they had a lot of refunds and people ordering now are going into Batch 3 so Batch 1 and 2 shouldn't be a problem since they knew for more than 6 months how many units they need and if we add the tons of refunds something doesn't add up. But it's KnC. Expect the worst!

The yield per wafer can be low; it's not guaranteed especially at the forefront of nm technology. They might have expected 90% usable chips from each wafer and ended up getting 50% usable chips. Plus 20nm is really just pushing the 28nm methods harder while waiting for 16nm.

Long story short, if you're waiting for batch one Neptune which you're not going to get any time soon post in my thread above.

u27

you may well find that more than half of batch 1 buyers actually refunded some time ago - around the time I was telling you to jump off the train (this is in no way a 'told you so') I was hopeful for you (and others) right to the end.

You also have to remember that only a tiny proportion of KnC's customers come from bitcointalk, if half of their customers from here refunded it probably accounts for less than 5% of their customer base.

Think about how many people actually physically post on here? Its the same old names over and over. Also, as with all sales you hardly ever hear from the people who are satisfied with their purchase or the service they receive, the noise is always from the small few who feel disgruntled or hard done by etc.

Face it, none of us have any clue about a company like KnC's true customer base, but I do remember that when I met Sam he didn't consider this forum to be of much interest in terms of where his sales come from. That might be hard for the egos of some to take in, but its probably true to a large extent.
hero member
Activity: 575
Merit: 500
This can't be true. They had A LOT of refunds so they must have a lot of Neptunes available.

Originally this batch of chips was only meant to cover the CA batch, has anyone from "B1" received their neptunes?
sr. member
Activity: 434
Merit: 250
I edited my post and added link to their news
Go look yourself.

I read the news, but it's simply stupid. The demand for Neptunes CAN'T be bigger than the wafers because they had a lot of refunds and people ordering now are going into Batch 3 so Batch 1 and 2 shouldn't be a problem since they knew for more than 6 months how many units they need and if we add the tons of refunds something doesn't add up. But it's KnC. Expect the worst!

The yield per wafer can be low; it's not guaranteed especially at the forefront of nm technology. They might have expected 90% usable chips from each wafer and ended up getting 50% usable chips. Plus 20nm is really just pushing the 28nm methods harder while waiting for 16nm.

Long story short, if you're waiting for batch one Neptune which you're not going to get any time soon post in my thread above.

u27

you may well find that more than half of batch 1 buyers actually refunded some time ago - around the time I was telling you to jump off the train (this is in no way a 'told you so') I was hopeful for you (and others) right to the end.
sr. member
Activity: 266
Merit: 250
I edited my post and added link to their news
Go look yourself.

I read the news, but it's simply stupid. The demand for Neptunes CAN'T be bigger than the wafers because they had a lot of refunds and people ordering now are going into Batch 3 so Batch 1 and 2 shouldn't be a problem since they knew for more than 6 months how many units they need and if we add the tons of refunds something doesn't add up. But it's KnC. Expect the worst!

The yield per wafer can be low; it's not guaranteed especially at the forefront of nm technology. They might have expected 90% usable chips from each wafer and ended up getting 50% usable chips. Plus 20nm is really just pushing the 28nm methods harder while waiting for 16nm.

Long story short, if you're waiting for batch one Neptune which you're not going to get any time soon post in my thread above.

u27
legendary
Activity: 1904
Merit: 1007
I edited my post and added link to their news
Go look yourself.

I read the news, but it's simply stupid. The demand for Neptunes CAN'T be bigger than the wafers because they had a lot of refunds and people ordering now are going into Batch 3 so Batch 1 and 2 shouldn't be a problem since they knew for more than 6 months how many units they need and if we add the tons of refunds something doesn't add up. But it's KnC. Expect the worst!
hero member
Activity: 882
Merit: 1003
Just wondering if Batch 1 or 2 have started receiving their Neptunes yet.
No and no more Neptunes shipping !



From KNC last news (link) :
Quote
Each and every production-ready Neptune processor we’ve been able to produce has been shipped out to customers
[...]
the demand for Neptunes has been bigger than the successful yield of 20nm wafers has been able to satisfy.

That means they don't have any 20nm chips !
That means no more Neptunes will ship !
It's going to take weeks before receiving new 20nm chips !

More BS from KNC.  They only tell you what they want you to HEAR.
hero member
Activity: 539
Merit: 517
Just wondering if Batch 1 or 2 have started receiving their Neptunes yet.
No and no more Neptunes shipping !



From KNC last news (link) :
Quote
Each and every production-ready Neptune processor we’ve been able to produce has been shipped out to customers
[...]
the demand for Neptunes has been bigger than the successful yield of 20nm wafers has been able to satisfy.

That means they don't have any 20nm chips !
That means no more Neptunes will ship !
It's going to take weeks before receiving new 20nm chips !

This can't be true. They had A LOT of refunds so they must have a lot of Neptunes available.
I edited my post and added link to their news
Go look yourself.
legendary
Activity: 1904
Merit: 1007
Just wondering if Batch 1 or 2 have started receiving their Neptunes yet.
No



No more Neptunes shipping !
From KNC last news (link) :
Quote
Each and every production-ready Neptune processor we’ve been able to produce has been shipped out to customers
[...]
the demand for Neptunes has been bigger than the successful yield of 20nm wafers has been able to satisfy.

That means they don't have any 20nm chips !
That means no more Neptunes will ship !
It's going to take weeks before receiving new 20nm chips !

This can't be true. They had A LOT of refunds so they must have a lot of Neptunes available.
hero member
Activity: 539
Merit: 517
Just wondering if Batch 1 or 2 have started receiving their Neptunes yet.
No and no more Neptunes shipping !



From KNC last news (link) :
Quote
Each and every production-ready Neptune processor we’ve been able to produce has been shipped out to customers
[...]
the demand for Neptunes has been bigger than the successful yield of 20nm wafers has been able to satisfy.

That means they don't have any 20nm chips !
That means no more Neptunes will ship !
It's going to take weeks before receiving new 20nm chips !
sr. member
Activity: 386
Merit: 250
I spent 24 hours at 500Mhz and 3.3Th/s.
I changed my room and very poor result Sad. Back in the first room.
For the moment, its always stable at 500Mhz and low HW. (0.09% !!).
(5s):3.328T (1m):3.293T (5m):3.269T (15m):3.273T (avg):3.214Th/s
 A:7611392  R:6144  HW:6909  WU:44915.7/m

http://www.image-heberg.fr/files/14042894687622.png
http://www.image-heberg.fr/files/14042894687639.png

PSU is generic (LDLC.COM) but stable and 80+. (2250 watts electric meter).
I ordered 6 x Corsair AX860 Modulaire - 860W to avoid having 5 PSU...
I try with an Corsair AX850 but PSU has broken down... (Only one Asic and controller).

I did not understand the functioning of voltage and I have not changed.
Maximum temperature for the day : 101°
I'm going to lower frequencies for the next days, 34° for thursday...
They will be installed in datacenter Monday.
Wow...  I also was thinking about the 850 for a single box and controller... not so much now... lol.
So, at night, when ambient temps cool down for the day...   what you get then?
I bet you'll be fine come monday...

Everything power related will effect performance, including different circuits in the house... (Happened to me with Jupiters)
making coffee or toast on the same circuit is obviously not a good idea.
Short 10/2 wire from outlet to breaker helps too if you're savvy.
The more pure and unhindered the current supply is, the smoother your asics will run.


I ran single cube and controller on AX760 for days without issue.
I ran about an hour with 2 cubes and controller on AX760.
EDIT I only ended the 2 cube test cause I don't like to run PSU at max output.

I used only first PCIe connector on double PCIe cables FWIW.

YMMV
Smiley
legendary
Activity: 938
Merit: 1000
LIR DEV
I spent 24 hours at 500Mhz and 3.3Th/s.
I changed my room and very poor result Sad. Back in the first room.
For the moment, its always stable at 500Mhz and low HW. (0.09% !!).
(5s):3.328T (1m):3.293T (5m):3.269T (15m):3.273T (avg):3.214Th/s
 A:7611392  R:6144  HW:6909  WU:44915.7/m

http://www.image-heberg.fr/files/14042894687622.png
http://www.image-heberg.fr/files/14042894687639.png

PSU is generic (LDLC.COM) but stable and 80+. (2250 watts electric meter).
I ordered 6 x Corsair AX860 Modulaire - 860W to avoid having 5 PSU...
I try with an Corsair AX850 but PSU has broken down... (Only one Asic and controller).

I did not understand the functioning of voltage and I have not changed.
Maximum temperature for the day : 101°
I'm going to lower frequencies for the next days, 34° for thursday...
They will be installed in datacenter Monday.
Wow...  I also was thinking about the 850 for a single box and controller... not so much now... lol.
So, at night, when ambient temps cool down for the day...   what you get then?
I bet you'll be fine come monday...

Everything power related will effect performance, including different circuits in the house... (Happened to me with Jupiters)
making coffee or toast on the same circuit is obviously not a good idea.
Short 10/2 wire from outlet to breaker helps too if you're savvy.
The more pure and unhindered the current supply is, the smoother your asics will run.
newbie
Activity: 5
Merit: 0
I spent 24 hours at 500Mhz and 3.3Th/s.
I changed my room and very poor result Sad. Back in the first room.
For the moment, its always stable at 500Mhz and low HW. (0.09% !!).
(5s):3.328T (1m):3.293T (5m):3.269T (15m):3.273T (avg):3.214Th/s
 A:7611392  R:6144  HW:6909  WU:44915.7/m

http://www.image-heberg.fr/files/14042894687622.png
http://www.image-heberg.fr/files/14042894687639.png

PSU is generic (LDLC.COM) but stable and 80+. (2250 watts electric meter).
I ordered 6 x Corsair AX860 Modulaire - 860W to avoid having 5 PSU...
I try with an Corsair AX850 but PSU has broken down... (Only one Asic and controller).

I did not understand the functioning of voltage and I have not changed.
Maximum temperature for the day : 101°
I'm going to lower frequencies for the next days, 34° for thursday...
They will be installed in datacenter Monday.
legendary
Activity: 938
Merit: 1000
LIR DEV
is  .0366 volts normal? What are they supposed to come set at?
hero member
Activity: 784
Merit: 504
Dream become broken often
I have two Neptune and its instable...
I tried different combination but it's complicated.

Single Asic totally crash after 15min :

http://www.image-heberg.fr/files/14042894688281.png
http://pastebin.com/hvDg9Hkc

5 Asic, 5 x 450Watts connected to a UPS (5000KVA, double conversion online).
Two brutally slow down with many errors (exemple : KnC: 5.1.100 disabled for 300 seconds due to repeated hardware errors)
http://www.image-heberg.fr/files/14042894686910.png

Good for 2 Asic but last, Asic 6 slight decrease
http://www.image-heberg.fr/files/14042894686893.png


Other test with 3 ASIC :

Boot : http://www.image-heberg.fr/files/14042894688980.png (ok)
5 min : http://www.image-heberg.fr/files/14042894689633.png (decrease)
15 min : http://www.image-heberg.fr/files/14042894689784.png (more errors than results)

Actually i try with 5 ASIC, 5 x 450Watts, no UPS, in another room.
At 475Mhz, Asic 2 and 3 crash : http://www.image-heberg.fr/files/14042894686910.png
I reboot Asic 2 and 3, cgminer crash and auto-restart and i set at 500Mhz, for the moment, its stable.
 (5s):3.356T (1m):3.300T (5m):3.285T (15m):2.944T (avg):3.107Th/s
 A:1618944  R:0  HW:5399  WU:43465.0/m

Between delay and instabilities, it is a BIG disappointment for the moment...

imho Sounds to me like HEAT is your main enemy by the times posted

i don't think it's heat...Loup can you watch the vrms temp...do alot of them get upto 100c? Loup, maybe try lower the volts n lower clock to see if they can indeed hash at full power? maybe check for metal shavings? maybe something is shorting out...oct. jups known for having them...knc loves to throw in extra parts to mess with ppl Smiley if they do hash at lower power then maybe it is your psu's...these units are pulling close to 300w and with a 450w...i dunno if the rail can handle it...also don't know if it's single or multi rail either...do you have a bigger psu you can test with?
legendary
Activity: 938
Merit: 1000
LIR DEV
I have two Neptune and its instable...
I tried different combination but it's complicated.

Single Asic totally crash after 15min :

http://www.image-heberg.fr/files/14042894688281.png
http://pastebin.com/hvDg9Hkc

5 Asic, 5 x 450Watts connected to a UPS (5000KVA, double conversion online).
Two brutally slow down with many errors (exemple : KnC: 5.1.100 disabled for 300 seconds due to repeated hardware errors)
http://www.image-heberg.fr/files/14042894686910.png

Good for 2 Asic but last, Asic 6 slight decrease
http://www.image-heberg.fr/files/14042894686893.png


Other test with 3 ASIC :

Boot : http://www.image-heberg.fr/files/14042894688980.png (ok)
5 min : http://www.image-heberg.fr/files/14042894689633.png (decrease)
15 min : http://www.image-heberg.fr/files/14042894689784.png (more errors than results)

Actually i try with 5 ASIC, 5 x 450Watts, no UPS, in another room.
At 475Mhz, Asic 2 and 3 crash : http://www.image-heberg.fr/files/14042894686910.png
I reboot Asic 2 and 3, cgminer crash and auto-restart and i set at 500Mhz, for the moment, its stable.
 (5s):3.356T (1m):3.300T (5m):3.285T (15m):2.944T (avg):3.107Th/s
 A:1618944  R:0  HW:5399  WU:43465.0/m

Between delay and instabilities, it is a BIG disappointment for the moment...

imho Sounds to me like HEAT is your main enemy by the times posted
Keep those VRM's cool
legendary
Activity: 3234
Merit: 1220
I have two Neptune and its instable...
I tried different combination but it's complicated.

Single Asic totally crash after 15min :

http://www.image-heberg.fr/files/14042894688281.png
http://pastebin.com/hvDg9Hkc

5 Asic, 5 x 450Watts connected to a UPS (5000KVA, double conversion online).
Two brutally slow down with many errors (exemple : KnC: 5.1.100 disabled for 300 seconds due to repeated hardware errors)
http://www.image-heberg.fr/files/14042894686910.png

Good for 2 Asic but last, Asic 6 slight decrease
http://www.image-heberg.fr/files/14042894686893.png


Other test with 3 ASIC :

Boot : http://www.image-heberg.fr/files/14042894688980.png (ok)
5 min : http://www.image-heberg.fr/files/14042894689633.png (decrease)
15 min : http://www.image-heberg.fr/files/14042894689784.png (more errors than results)

Actually i try with 5 ASIC, 5 x 450Watts, no UPS, in another room.
At 475Mhz, Asic 2 and 3 crash : http://www.image-heberg.fr/files/14042894686910.png
I reboot Asic 2 and 3, cgminer crash and auto-restart and i set at 500Mhz, for the moment, its stable.
 (5s):3.356T (1m):3.300T (5m):3.285T (15m):2.944T (avg):3.107Th/s
 A:1618944  R:0  HW:5399  WU:43465.0/m

Between delay and instabilities, it is a BIG disappointment for the moment...


What PSU's are you using?
newbie
Activity: 5
Merit: 0
I have two Neptune and its instable...
I tried different combination but it's complicated.

Single Asic totally crash after 15min :

http://www.image-heberg.fr/files/14042894688281.png
http://pastebin.com/hvDg9Hkc

5 Asic, 5 x 450Watts connected to a UPS (5000KVA, double conversion online).
Two brutally slow down with many errors (exemple : KnC: 5.1.100 disabled for 300 seconds due to repeated hardware errors)
http://www.image-heberg.fr/files/14042894686910.png

Good for 2 Asic but last, Asic 6 slight decrease
http://www.image-heberg.fr/files/14042894686893.png


Other test with 3 ASIC :

Boot : http://www.image-heberg.fr/files/14042894688980.png (ok)
5 min : http://www.image-heberg.fr/files/14042894689633.png (decrease)
15 min : http://www.image-heberg.fr/files/14042894689784.png (more errors than results)

Actually i try with 5 ASIC, 5 x 450Watts, no UPS, in another room.
At 475Mhz, Asic 2 and 3 crash : http://www.image-heberg.fr/files/14042894686910.png
I reboot Asic 2 and 3, cgminer crash and auto-restart and i set at 500Mhz, for the moment, its stable.
 (5s):3.356T (1m):3.300T (5m):3.285T (15m):2.944T (avg):3.107Th/s
 A:1618944  R:0  HW:5399  WU:43465.0/m

Between delay and instabilities, it is a BIG disappointment for the moment...
legendary
Activity: 1316
Merit: 1014
ex uno plures
Corsair IS Seasonic.

No.

Seasonic make some PSUs for Corsair, which Corsair then rebrands, but Corsair IS NOT Seasonic.


When in doubt, buy Seasonic. Or those great EVGA 1300G2 if you want to save some $$$
full member
Activity: 146
Merit: 100
Just wondering if Batch 1 or 2 have started receiving their Neptunes yet.
legendary
Activity: 3234
Merit: 1220
Neptune is very good miner
Had perfect chip

But need better extra cooling

I will make some without dremmel the case

I'm glad you are finally happy jelin Smiley

BTW your english is probably a million times better than anyone elses attempt at whatever language you speak so don't worry about it Smiley

Now looks like I might hopefully only be a day or so away from my refund, then I can pay off some debts.
Jump to: