[CLOSED]R5: KnC Jupiter, BELOW-COST+Host! $78+Bonuses! Hashing, 1st payouts sent - page 6.

philipma1957

legendary

Activity: 4256

Merit: 8551

'The right to privacy matters'

Quote from: Miramyn on October 24, 2013, 09:55:10 AM

I am a big proponent of the "If it ain't broke, don't fix it" school of thought.

I'd rather have unit working and working at 30% over spec (520 instead of 400) than try to squeeze another 5% out of it and end up like some of the reports where they are hashing at 0GH or 1GH/s

Look lets say the btc lost comes to .52BTC over the next 13 days that is about 104 usd. I donated an 8 pack of ups batteries to the coop they cost 109 usd.

http://www.rakuten.com/pr/product.aspx?sku=246155454

I am having these dropped shipped as a donation to the coop on me free gratis this will make up for the "missing" btc for the next 13 days.

With hashrate having some importance uptime is the most important thing since diff just rockets. So the coop is not getting f''d as it will now have 2 extra battery packs for the UPS.

At bobsag3 I am sending a pm with tracking for the donation 8 pack. best regards Phil

Miramyn

full member

Activity: 169

Merit: 100

I am a big proponent of the "If it ain't broke, don't fix it" school of thought.

I'd rather have unit working and working at 30% over spec (520 instead of 400) than try to squeeze another 5% out of it and end up like some of the reports where they are hashing at 0GH or 1GH/s

bobsag3

hero member

Activity: 546

Merit: 500

Owner, Minersource.net

Quote from: -Redacted- on October 24, 2013, 09:48:31 AM

Quote from: Tiennou74 on October 24, 2013, 09:34:44 AM

I saw this.
Hope it helps :

from my personal experoence: flashed 96 had to restart cgminer two times and then went good with 274Ghs. flashed 97 had to restart cgminer one time and then 284Ghs. i now think cgminer must be restarted at least one time after upgrading firmwares

Still super happy with our GB!
Thanks

Thanks! We're really trying to keep everything transparent and above-board here.

Quote

.96 does seem to be better than .97 on good jupiters. I haven't tried it on a bad jupiter yet.. so far bad jupiters hash better on .94 'overclocked' running bfgminer but the VRMs/chips do have to take a beating.. perhaps heatsinks?

to 'cure' modules, remove the asic_test file and install .96, run enable cores and few times and reboot. It should do its 1-2 min asic tests and then reboot itself. you should have more enabled cores again

There is so much information flying back and forth about the various firmwares that , at the moment, it's difficult to know what's what. it's probably better to hold off on updates in the future until they're at least a week or so old. Might as well let OTHER people figure stuff out before we do something that might adversely affect our machines....

Dunno. Some people say .94 works better, others .95, .96, .97 - some say .97 breaks their machines, so until things calm down for a few days, there isn't much point in doing anything else - lest we keep screwing around and end up with everything hashing at 450 instead of 520...

Im not going to touch .97 with a 10ft pole until I determine weather or not it had a hand in me having a large issue with a non-GB jupiter.

-Redacted-

hero member

Activity: 574

Merit: 501

Quote from: Tiennou74 on October 24, 2013, 09:34:44 AM

I saw this.
Hope it helps :

from my personal experoence: flashed 96 had to restart cgminer two times and then went good with 274Ghs. flashed 97 had to restart cgminer one time and then 284Ghs. i now think cgminer must be restarted at least one time after upgrading firmwares

Still super happy with our GB!
Thanks

Thanks! We're really trying to keep everything transparent and above-board here.

Quote

.96 does seem to be better than .97 on good jupiters. I haven't tried it on a bad jupiter yet.. so far bad jupiters hash better on .94 'overclocked' running bfgminer but the VRMs/chips do have to take a beating.. perhaps heatsinks?

to 'cure' modules, remove the asic_test file and install .96, run enable cores and few times and reboot. It should do its 1-2 min asic tests and then reboot itself. you should have more enabled cores again

There is so much information flying back and forth about the various firmwares that , at the moment, it's difficult to know what's what. it's probably better to hold off on updates in the future until they're at least a week or so old. Might as well let OTHER people figure stuff out before we do something that might adversely affect our machines....

Dunno. Some people say .94 works better, others .95, .96, .97 - some say .97 breaks their machines, so until things calm down for a few days, there isn't much point in doing anything else - lest we keep screwing around and end up with everything hashing at 450 instead of 520...

Tiennou74

sr. member

Activity: 546

Merit: 250

I saw this.
Hope it helps :

from my personal experoence: flashed 96 had to restart cgminer two times and then went good with 274Ghs. flashed 97 had to restart cgminer one time and then 284Ghs. i now think cgminer must be restarted at least one time after upgrading firmwares

Still super happy with our GB!
Thanks

-Redacted-

hero member

Activity: 574

Merit: 501

Quote

The upgrade was not a smooth process like other ones. I had the problem of getting 1GH/s after the upgrade and followed the advise in http://forum.kncminer.com/forum/main...0074-erm-1gh-s - Thanks eXceed.
Downgrade to 0.95.
Before restarting device, ssh to box and copy all *.factory files to original filenames.
Restart device.
Upgrade EnableCores.bin.
Restart.
Profit.
I did a hard reset instead of doing the copy *.factory files as I believe this does the same step.

I then upgraded to 0.97 directly from 0.95 and hey presto working again.

I've only been going 10 minutes and my average is already up by 5GH/s. In the past I've had to give it at least 8 hours to get to this level, so looking good so far.

The so called "loss" is .0389 BTC, or .000389 BTC per share - about 7 cents per share per day over a ten day period.

So, how about figuring how much "gain" there has been at the current hashrate above 400 Gh/s (which was the promised hashrate of the group buy.)

Quote

on the third day when diff jumps to 380 mill the loss is .039 btc for 1 day so 11 days at that difficulty is .429 btc plus the .1134 btc

that comes .5424 over the next 13 days

or .0052 per share over the next two weeks = $0.99.

Now calculate how much gain there has been on a minimum of 120 Gh/s over the promised 400 Gh/s since the GB started hashing.

I'm not trying to be a smart-ass here, but you are calling something a "loss" when it is not - it is slightly less of a gain over the expected payout for 400 Gh/s == hashing at 520 Gh/s versus possibly hashing at 550 Gh/s.

And we ARE looking at the issue to see what we can do.

-R-

philipma1957

legendary

Activity: 4256

Merit: 8551

'The right to privacy matters'

Quote from: Ashitank on October 24, 2013, 07:39:10 AM

Thank you philipma1957 , its just that 0.3 BTC x 15 days = 1.5 btc lost , a single restart wont loose as much.

But after reading what Redacted posted , I think we should just go back to 0.95 & not touch the machines there after , I am fully confident that organizers of this group buy are genuine good people & will take care of things fine.

that is not the loss. the loss is .389 btc in 10 days.

30gh is lost and it could be due to a lessor chip.

30gh is .0567 btc today 1 day.

30gh is .0567 btc tomorrow 2 days. that comes to .1134 btc

on the third day when diff jumps to 380 mill the loss is .039 btc for 1 day so 11 days at that difficulty is .429 btc plus the .1134 btc

that comes .5424 over the next 13 days

jungle_dave

full member

Activity: 200

Merit: 100

Thanks Adeel!

It looks like this firmware upgrade is only from .95 to .97 (woops) even so, some people are only getting 250 GH/s (jupiter) as an improvement .....
This reeks of badly debugged device driver!
Anyone want to look at the code?

Adeel06

full member

Activity: 217

Merit: 100

Quote from: jungle_dave on October 24, 2013, 07:42:58 AM

Ive been in contact with some other KnC operators, everyone is having similar problems.
Thats the price you pay for bleeding edge delivery dates.

you might want to check this

http://forum.kncminer.com/forum/main-category/main-forum/10016-97-firmware-feedback

Jungle_dave, you are the man, helping others out all the time. Thanks again for the help you provided me through PM's

!

jungle_dave

full member

Activity: 200

Merit: 100

Ive been in contact with some other KnC operators, everyone is having similar problems.
Thats the price you pay for bleeding edge delivery dates.

you might want to check this

http://forum.kncminer.com/forum/main-category/main-forum/10016-97-firmware-feedback

Ashitank

hero member

Activity: 560

Merit: 500

Thank you philipma1957 , its just that 0.3 BTC x 15 days = 1.5 btc lost , a single restart wont loose as much.

But after reading what Redacted posted , I think we should just go back to 0.95 & not touch the machines there after , I am fully confident that organizers of this group buy are genuine good people & will take care of things fine.

philipma1957

legendary

Activity: 4256

Merit: 8551

'The right to privacy matters'

Quote from: DyslexicZombei on October 24, 2013, 03:01:07 AM

I've been in the line of work of Test & Evaluation and R&D for 10 years at mostly low TRL numbers, and I can tell you that these miners are all essentially prototypes.

There was no long term testing so it's natural to have variability in chip yield and board performance. Firmware and software are still in a huge state of flux.

It sucks, but that's the kinds of risks we all take going in. IIRC, Jupiters were sold as 400GH/s units so looking at it from a glass half full perspective would probably be more realistic. In this case as we're talking about the world's first 28nm chips in a super short time frame to get to market, bumps and weirdness along the way are to be expected.

YOU ARE WRONG THE GLASS IS 52/55 FULL! LOL

At Ashitank :

The current mining difficulty is 267,731,249. The next difficulty is estimated to change by 42.16% to 380,608,553 on block #266,111. That should be 1 days 18 hours 7 mins 18 seconds from now.

from https://bitclockers.com/calc

I am 56 years old and have worked with computers since I am 18. They vary, they vary a lot.

I have used more then 500 intel cpu chips.

When overclocking they vary, they vary a lot.

If you want to f with our 520 miner trying to get it to be as good as our 550 miner I get that. I have toyed around with pc gear tweaked it to make it bigger, faster ,stronger . Sometimes it works sometimes it does not work.

So risking 520 to get 30 more does not appeal to me.

520 is .97 btc a day. 30gh is .056 btc a day I do not like those odds.

in a day difficulty goes to 387 mill (est) 520 will earn 6.744btc in 10 days 30 will earn .389btc in 10 days .

So down time would be a bad thing you are chasing after .4 coins and using 6.8 coins to get it.

The next 2 diff adjustments will be big 99% up time for the next 20-25 days is very important.

Most of us took a shot that diff would be a little less then 40% each jump. I for one do not want to send "good money after bad"

a 550 machine is 5% faster then a 520 machine. the next jump is 40% while the 5% would be nice we could create weeks of down time chasing the 5% just not worth it.

-Redacted-

hero member

Activity: 574

Merit: 501

Quote

Quote from: FeedbackLoop on Today at 03:58:54 AM

0.97 also shows a continuous drop of hashrate for me from 500 to 470 (3 hour average eligius).

Actually.. it doesn't seem to be coming back in any previous firmware. Last value was 450 and that's where it stayed. I read compatible reports on KNC's forum. I fear 0.97 might damage Jupiters... Careful upgrading yours.

So - it may well be the case that following the suggestion to upgrade the firmware to .97 has damaged the Jupiter at least semi-permanently. This is why I never attempt to fix things that aren't broken. And please recall that these units were sold as by KnC as guaranteed to hash at "at least 400 Gh/s", so anything above and beyond that is a bonus.

And you might want to also consider the recent lucky (for us) announcement that HashFast shipping will be delayed for "a few weeks" - meaning we get at least two extra weeks of hashing in before the difficulty increases due to HashFast boxes shipping kicks in.

-R-

Ashitank

hero member

Activity: 560

Merit: 500

Quote from: DyslexicZombei on October 24, 2013, 03:01:07 AM

I've been in the line of work of Test & Evaluation and R&D for 10 years at mostly low TRL numbers, and I can tell you that these miners are all essentially prototypes.

There was no long term testing so it's natural to have variability in chip yield and board performance. Firmware and software are still in a huge state of flux.

It sucks, but that's the kinds of risks we all take going in. IIRC, Jupiters were sold as 400GH/s units so looking at it from a glass half full perspective would probably be more realistic. In this case as we're talking about the world's first 28nm chips in a super short time frame to get to market, bumps and weirdness along the way are to be expected.

If we have chance of earning 0.3xxx btc more every day by just changing firmware why loose this opportunity , if any other investors in this group buy think other wise please share your opinion , after all it is your BTC at stake here

Sorry but a professional group buy should not be looked at as half glass full investment , not doing this for hobby.

DyslexicZombei

sr. member

Activity: 364

Merit: 250

I've been in the line of work of Test & Evaluation and R&D for 10 years at mostly low TRL numbers, and I can tell you that these miners are all essentially prototypes.

There was no long term testing so it's natural to have variability in chip yield and board performance. Firmware and software are still in a huge state of flux.

It sucks, but that's the kinds of risks we all take going in. IIRC, Jupiters were sold as 400GH/s units so looking at it from a glass half full perspective would probably be more realistic. In this case as we're talking about the world's first 28nm chips in a super short time frame to get to market, bumps and weirdness along the way are to be expected.

Ashitank

hero member

Activity: 560

Merit: 500

Thank you guys , I think we can re-group in 6 hours time to check on 24 hours BTC earned by that time pool shift e.t.c should be done ? , if 24 hours BTC earned is still lower that 1.xx BTC then we have a problem Capitan.

bobsag3

hero member

Activity: 546

Merit: 500

Owner, Minersource.net

Quote from: thomas_s on October 24, 2013, 12:17:47 AM

Quote from: Ashitank on October 23, 2013, 11:26:48 PM

Quote from: thomas_s on October 23, 2013, 11:14:20 PM

Quote from: Ashitank on October 23, 2013, 10:53:57 PM

Worker 1 Hash Rate 238.228 Gh/s
Worker 2 Hash Rate 269.533 Gh/s

Bobsage worker 1 hashrate is still not-up to the mark & also earner btc for 24 hours is still not 1.xx which it should be, I suppose we should revert to firmware 0.95 on worker 1 Jupiter only , worker 2 Jupiter is very happy on 0.97 firmware.

We are loosing out on 0.2xx ~ 0.3xxx btc due to worker 1 rather low hashrate on 0.97 firmware.

The 24 hour bitcoin is an average from BTC Guild it has to do on open shifts which one of the machines went down for a little while which lowered the average 24 hour BTC gathered, someone did post when those shifts close the 24 hour BTC will grow.

Ok Thomas , but what about worker 1 hashrate being much lower than Normal for 24 hours or more now

I've just talked with bobsag3 about it, he said that the two miners R5/R6 (worker1 and worker2) are averaging 520 / 556, its possible that the site isn't displaying the correct info its not updated in real time and the information isn't gathered from the miner.

Also not all units were created equal. The first Jupiter I got down here has been solid 550+ the entire time, the other is more like 520. + 3 small downtimes for firmware updates (.94/.95 > .96 > .97 with reboots in between) would account for the small difference.

thomas_s

hero member

Activity: 574

Merit: 500

Quote from: Ashitank on October 23, 2013, 11:26:48 PM

Quote from: thomas_s on October 23, 2013, 11:14:20 PM

Quote from: Ashitank on October 23, 2013, 10:53:57 PM

Worker 1 Hash Rate 238.228 Gh/s
Worker 2 Hash Rate 269.533 Gh/s

Bobsage worker 1 hashrate is still not-up to the mark & also earner btc for 24 hours is still not 1.xx which it should be, I suppose we should revert to firmware 0.95 on worker 1 Jupiter only , worker 2 Jupiter is very happy on 0.97 firmware.

We are loosing out on 0.2xx ~ 0.3xxx btc due to worker 1 rather low hashrate on 0.97 firmware.

The 24 hour bitcoin is an average from BTC Guild it has to do on open shifts which one of the machines went down for a little while which lowered the average 24 hour BTC gathered, someone did post when those shifts close the 24 hour BTC will grow.

Ok Thomas , but what about worker 1 hashrate being much lower than Normal for 24 hours or more now

I've just talked with bobsag3 about it, he said that the two miners R5/R6 (worker1 and worker2) are averaging 520 / 556, its possible that the site isn't displaying the correct info its not updated in real time and the information isn't gathered from the miner.

Ashitank

hero member

Activity: 560

Merit: 500

Quote from: thomas_s on October 23, 2013, 11:14:20 PM

Quote from: Ashitank on October 23, 2013, 10:53:57 PM

Worker 1 Hash Rate 238.228 Gh/s
Worker 2 Hash Rate 269.533 Gh/s

Bobsage worker 1 hashrate is still not-up to the mark & also earner btc for 24 hours is still not 1.xx which it should be, I suppose we should revert to firmware 0.95 on worker 1 Jupiter only , worker 2 Jupiter is very happy on 0.97 firmware.

We are loosing out on 0.2xx ~ 0.3xxx btc due to worker 1 rather low hashrate on 0.97 firmware.

The 24 hour bitcoin is an average from BTC Guild it has to do on open shifts which one of the machines went down for a little while which lowered the average 24 hour BTC gathered, someone did post when those shifts close the 24 hour BTC will grow.

Ok Thomas , but what about worker 1 hashrate being much lower than Normal for 24 hours or more now

thomas_s

hero member

Activity: 574

Merit: 500

Quote from: Ashitank on October 23, 2013, 10:53:57 PM

Worker 1 Hash Rate 238.228 Gh/s
Worker 2 Hash Rate 269.533 Gh/s

Bobsage worker 1 hashrate is still not-up to the mark & also earner btc for 24 hours is still not 1.xx which it should be, I suppose we should revert to firmware 0.95 on worker 1 Jupiter only , worker 2 Jupiter is very happy on 0.97 firmware.

We are loosing out on 0.2xx ~ 0.3xxx btc due to worker 1 rather low hashrate on 0.97 firmware.

The 24 hour bitcoin is an average from BTC Guild it has to do on open shifts which one of the machines went down for a little while which lowered the average 24 hour BTC gathered, someone did post when those shifts close the 24 hour BTC will grow.

Topic: [CLOSED]R5: KnC Jupiter, BELOW-COST+Host! $78+Bonuses! Hashing, 1st payouts sent - page 6. (Read 30647 times)