Author

Topic: Swedish ASIC miner company kncminer.com - page 1513. (Read 3050071 times)

legendary
Activity: 1456
Merit: 1018
HoneybadgerOfMoney.com Weed4bitcoin.com
October 01, 2013, 05:17:08 AM
Shocked ...I guess it'S not real and I'm dreaming....DA..FUQ!!...I'm blown away


   Hashrate Average   Weighted Shares
3 hours   457.13 Gh/s   1149489
22.5 minutes   718.62 Gh/s   225878
256 seconds   793.56 Gh/s   47300
128 seconds   820.62 Gh/s   24456

Could be multiple units on burning maybe not we don't know if they are doing batches of machines or whatnot.
sr. member
Activity: 476
Merit: 250
let's have some fun
October 01, 2013, 05:08:38 AM
 Shocked ...I guess it'S not real and I'm dreaming....DA..FUQ!!...I'm blown away


   Hashrate Average   Weighted Shares
3 hours   457.13 Gh/s   1149489
22.5 minutes   718.62 Gh/s   225878
256 seconds   793.56 Gh/s   47300
128 seconds   820.62 Gh/s   24456
sr. member
Activity: 356
Merit: 250
Dock.io
October 01, 2013, 05:07:24 AM

Well, if you purchased today:

Bitfury, 400 GH/s, $8000, delivered in Oct.
KnC, 500 GH/s, $4995, delivered in Nov.

This is somewhere between 1 day and two months difference, based on exact delivery dates. Hard to choose

Well, October-delivery h-boards have been shipping already and they've been delivered today. Yeah, in September.. In September though they're marked as October delivery. Cheesy

500 GH/s in November is going to be pissin' in the wind it looks like...
hero member
Activity: 574
Merit: 500
1.21 GIGA WATTS
October 01, 2013, 05:07:16 AM
anyone with more than a dozen jupiters shoud srsly consider donating one to ckolivas, even if its only 'temporary' while he irons out the kinks...

if anyone can get us in the 4 figure hashrate, its him and or kano.

but then my post will probably be modded away so the real shitwaffle barons' crud doesnt get sidetracked...



Which country are they located..?  is one of them in Australia..?
sr. member
Activity: 1176
Merit: 265
October 01, 2013, 05:06:36 AM
If I were there making  that video I'd be awfully tempted to turn round and reveal Josh's immortal words about KnC on the back of my lab coat.
Turned out to be the highest praise didn't it?

Dunno if it's the room it's in being so big or the time it's been running not being too long...but it's pleasantly quiet running for a box of fans isn't it?
hero member
Activity: 574
Merit: 500
1.21 GIGA WATTS
October 01, 2013, 05:06:00 AM
If I find a block mining in a pool, is the bock that my miner found solely work submitted by my miner only, or is it a combined effort with all pool miners.?

I'm asking because I have been mining in a pool with 30GH/s for 34 days and my miner has found 2 blocks.  Is this combined effort or just my miner.?

legendary
Activity: 1859
Merit: 1001
October 01, 2013, 05:02:29 AM
anyone with more than a dozen jupiters shoud srsly consider donating one to ckolivas, even if its only 'temporary' while he irons out the kinks...

if anyone can get us in the 4 figure hashrate, its him and or kano.

but then my post will probably be modded away so the real shitwaffle barons' crud doesnt get sidetracked...

hero member
Activity: 574
Merit: 501
October 01, 2013, 04:55:02 AM
500GH but 20% of hardware errors. It's better to lower to original 400GH and have more efficiency at 1% HW errors.

Actually according to the video they are running at diff=127 so about 18mil accepted shares and less than 1% errors.
Those hardware error share are also diff=127. So accepted 18mil shares and 3.6mil shares are hardware error. Let's be honest...

Ummm - no.  Hardware errors are not diff 127...
Yes they are. HW errors are those share that's not validate on CPU but are reported by hardware. And in this case all shares reported are diff=>127.

The mining software determines the diff shares reported back to the pool.  The hardware has no idea what diff is, it hashes nonce ranges.  A hardware error not is a diff anything...
If so, it ruins whole idea about vardiff, reducing bandwith (between device and software, at least). I didn't look at cgminer code, but I would like to hear from ckolivas about that matter...


Vardiff is about controlling the number of answers you are returning to the pool each minute.  At diff-1 you return all of them, at diff-2 you return approximately 1/2 of them, of those about 1/2 are diff-4 or better, etc.  There are multiple nonces in the range of work assigned to you by the pool, how many of them you return depends on how many zeroes are in the answers for that assignment as your hardware searches the range - the diff setting the pool has assigned.   The mining software makes this determination.
copper member
Activity: 2898
Merit: 1465
Clueless!
October 01, 2013, 04:53:38 AM
I think KNC's achievement is great. The extra hashrate is more than welcome given the recent spike in difficulty.

BUT....

Why not keep up the good communication? They clearly didn't ship yesterday so we are past the date they promised. The least one would expect is an update on the new expected delivery date. At this point the uncertainties should be much smaller so they should be able to give the dates for the different orders (day 1, day 2 etc).

Please acknowledge that your customers are taking actions based on this information, e.g. booking flights, taking days off etc.


Then again we could just chill. It's only btcs  Cheesy


er they did mention and bitcoinorama mentioned that they we're at the mnfg facilities...he said with security it was a pain to get stuff out when he had permission..that may have had something to do with it...that and the fact they we're rushed FYI....

Searing
legendary
Activity: 1260
Merit: 1008
October 01, 2013, 04:50:26 AM
Sorry if this is a repost havent had time to get through old posts yet -

Hashing @ 500GH - http://www.youtube.com/watch?v=BgS6JBb4r1g

And shipping today!

sorry for bashing but if you don't have time to skim the last few pages just don't post

Last few pages... try 25 some of us do have jobs and cant spend every minute on here, as much as I would like too... I also figured that since the video was only uploaded 24mins before I posted it was pretty fresh... turns out it was a duplicate video with a different name!

don't take it personally, really.

I was just trying to maximize thread's SNR. Anyway I know that not everybody has the luxury to waste time to follow the thread too closely, but posting redundant info like you did, make it even harder and more time consuming.
legendary
Activity: 1029
Merit: 1000
October 01, 2013, 04:43:17 AM
500GH but 20% of hardware errors. It's better to lower to original 400GH and have more efficiency at 1% HW errors.

Actually according to the video they are running at diff=127 so about 18mil accepted shares and less than 1% errors.
Those hardware error share are also diff=127. So accepted 18mil shares and 3.6mil shares are hardware error. Let's be honest...

Ummm - no.  Hardware errors are not diff 127...
Yes they are. HW errors are those share that's not validate on CPU but are reported by hardware. And in this case all shares reported are diff=>127.

The mining software determines the diff shares reported back to the pool.  The hardware has no idea what diff is, it hashes nonce ranges.  A hardware error not is a diff anything...
If so, it ruins whole idea about vardiff, reducing bandwith (between device and software, at least). I didn't look at cgminer code, but I would like to hear from ckolivas about that matter...
newbie
Activity: 53
Merit: 0
October 01, 2013, 04:36:36 AM
Sorry if this is a repost havent had time to get through old posts yet -

Hashing @ 500GH - http://www.youtube.com/watch?v=BgS6JBb4r1g

And shipping today!

sorry for bashing but if you don't have time to skim the last few pages just don't post

Last few pages... try 25 some of us do have jobs and cant spend every minute on here, as much as I would like too... I also figured that since the video was only uploaded 24mins before I posted it was pretty fresh... turns out it was a duplicate video with a different name!
legendary
Activity: 1260
Merit: 1008
October 01, 2013, 04:35:01 AM
And what about the 5V connector.

Should I buy a 230V to 5V DC converter?


yeah there's a lot of think that has to be sorted out imho, let me summarize my own list:

- Is a 850W PSU really enough for a Jupiter? (in the 'orama video (1) I saw the power consumption at wall was 843W)
- Do we need a 5V connector?
- What is the delivery ETA for day 1? Or looking at it in another way how long does it take to do the last tweaking to the machine?

(1) http://www.youtube.com/watch?feature=player_detailpage&v=CiQRxJuoRR8#t=25
hero member
Activity: 574
Merit: 501
October 01, 2013, 04:34:25 AM
500GH but 20% of hardware errors. It's better to lower to original 400GH and have more efficiency at 1% HW errors.

Actually according to the video they are running at diff=127 so about 18mil accepted shares and less than 1% errors.
Those hardware error share are also diff=127. So accepted 18mil shares and 3.6mil shares are hardware error. Let's be honest...

Ummm - no.  Hardware errors are not diff 127...
Yes they are. HW errors are those share that's not validate on CPU but are reported by hardware. And in this case all shares reported are diff=>127.

The mining software determines the diff shares reported back to the pool.  The hardware has no idea what diff is, it hashes nonce ranges.  A hardware error not is a diff anything...
sr. member
Activity: 288
Merit: 250
October 01, 2013, 04:30:45 AM
I do hope the hashing speed showing now on that test account is a result of further tweaking/oc . It's showing 650GH/s pretty constant now, with spikes @ 700+

Edit: showing 900+ now, so more than Jupiter hashing there :-)
legendary
Activity: 1442
Merit: 1000
Antifragile
October 01, 2013, 04:29:14 AM
500GH but 20% of hardware errors. It's better to lower to original 400GH and have more efficiency at 1% HW errors.

Guys, the units were just turned on and not tuned yet. 20% errors is unacceptable and they know that. But the units have to be tuned first.
They are not shipping units with 20% errors.
hero member
Activity: 635
Merit: 500
October 01, 2013, 04:28:36 AM
And what about the 5V connector.

Should I buy a 230V to 5V DC converter?
legendary
Activity: 1029
Merit: 1000
October 01, 2013, 04:27:43 AM
500GH but 20% of hardware errors. It's better to lower to original 400GH and have more efficiency at 1% HW errors.

Actually according to the video they are running at diff=127 so about 18mil accepted shares and less than 1% errors.
Those hardware error share are also diff=127. So accepted 18mil shares and 3.6mil shares are hardware error. Let's be honest...

Ummm - no.  Hardware errors are not diff 127...
Yes they are. HW errors are those share that's not validate on CPU but are reported by hardware. And in this case all shares reported are diff=>127.
hero member
Activity: 574
Merit: 501
October 01, 2013, 04:24:41 AM
500GH but 20% of hardware errors. It's better to lower to original 400GH and have more efficiency at 1% HW errors.

Actually according to the video they are running at diff=127 so about 18mil accepted shares and less than 1% errors.
Those hardware error share are also diff=127. So accepted 18mil shares and 3.6mil shares are hardware error. Let's be honest...

Ummm - no.  Hardware errors are not diff 127...
hero member
Activity: 826
Merit: 1000
'All that glitters is not gold'
October 01, 2013, 04:21:33 AM
The real story here isn't "ROI or no ROI?  Ship 9/30 or 10/3?"  The celebration and the sour grapes, the sore losers and the ecstatic supporters, the assholes and the gentlemen, that's all fun to read.  But IMO opinion that's not the story.

The real story is the presentation of a functioning IC entailing leading edge technology in a period of time far less than they'll teach you in school is even feasible.  

Every time some information came out, it was met with armchair experts who pontificated on the folly of it all.  It's all in the 600+ pages of this thread:  "The package is ridiculously large; OrSoC must be idiots."  "The power consumption will be way off the scale."  "They can't come to market without the test sequence. It's going to be a disaster of BFL proportions."  "They're too small to produce in size."  Best of all:  they're Swedish, how can they possibly do high tech?

Spending a significant part of my life in venture capital analysis, I have to say:  while there were no red flags for me from a business or technical standpoint, I didn't think it would happen this close to schedule; too many moving parts and supply chain issues to be certain.  

There are still red flags, I'd say, in both CoinTerra and HashFast, but I'm not going to discuss them.  They will be what they will be.  Instead, I'm just sayin' to KnC's professionals:

Spectacular performance!


+1
Congratulations KNC !
Jump to: