Topic: Swedish ASIC miner company kncminer.com - page 1183. (Read 3049528 times)
DHL is odd. It got to the states fast. Now its came to a craw..
When I first got my Saturn I removed the heatsinks, cleaned off whatever thermal grease that was there ( it was very thin stuff ) and put my own Artic Silver high density stuff on. Since then no matter if the temps are cool or hot I've always hashed on average 280-285. Top off, top on, fans on or on...always the same. Maybe the thermal grease or whatever they used was just inferior low quality stuff? After the grease dries up a little it should thicken up...when I received my miner the grease coming out on the sides was still almost wet. Maybe if you can't get your miner to hash up no matter what you do, try to clean off that crappy thermal grease that's there and re-do it yourself. AND REMEMBER, all you need is just a little eraser sized grease and not to over tighten. And I'll still not break even probably.
Did they do worse before you changed the thermal paste? I doubt that it's much of a problem, it seems like most people are getting 280 saturn /560 jupiter without doing that. Some people are having issues, but that's probably not related to heat.
I never trusted the grease that was on there to begin with, so I only powered up for about 20 minutes to make sure everything worked then changed the grease over to what I wanted. It could have been the same, but my Dad who has a Saturn also has great running speeds also and I changed his thermal grease as well. It can't hurt I guess. The stuff that I saw there just didn't look right. Very bright silver stuff that was almost wet. That's what made me change it to begin with.
This asics actually do like it warm, this does not mean you should remove the cooling. But when i put in 2 tornado fans which gave a drop from 62 c to 28 c on the asic cores it made them slow down.
The hash speed ended up being on avg 210 GH at that moment.
Before anyone asks NO the psu is not a weak one its a 1200 watt Antec so for this simple saturn even a bit overpowered with its 6 x 30 amp rails.
Now when i removed the tornado fans the hash went back up to 265 Gh on avg.
You guys are gonna think I'm totally nuts... remember how I removed the ASIC fans??
Well, my basement is so cold that I'm now experimenting with freezer bags covering
most of the heatsinks. lol Temps went up about 5C and confirming more hash and WU.
Power draw is about 343W. Temps are in the mid-40Cs. I'm thinking maybe the ASICs
need to be warm, but the rest of the components should be kept cool. I put rubber bands
towards the top to keep them in a steady position after the pic. And check-out my peak
number!!! FM 9.8 and enablecores. I have not done a restart after initially putting the
bags on less than they are now.
Well, my basement is so cold that I'm now experimenting with freezer bags covering
most of the heatsinks. lol Temps went up about 5C and confirming more hash and WU.
Power draw is about 343W. Temps are in the mid-40Cs. I'm thinking maybe the ASICs
need to be warm, but the rest of the components should be kept cool. I put rubber bands
towards the top to keep them in a steady position after the pic. And check-out my peak
number!!! FM 9.8 and enablecores. I have not done a restart after initially putting the
bags on less than they are now.
Also, if you hate editing files while in SSH you can always do the lazy mans way. Download WinSCP and log into your miner that way. It's all GUI for those of us that hate text based stuff. You'll get a good feel of the file hierachy and it is easier to move around, at least in my opinion it is. I'll just download a file, edit it locally and then upload it. I hate VI. Also of course, you can download all your files and make a nice backup locally.
http://winscp.net/eng/index.php
http://winscp.net/eng/index.php
When I first got my Saturn I removed the heatsinks, cleaned off whatever thermal grease that was there ( it was very thin stuff ) and put my own Artic Silver high density stuff on. Since then no matter if the temps are cool or hot I've always hashed on average 280-285. Top off, top on, fans on or on...always the same. Maybe the thermal grease or whatever they used was just inferior low quality stuff? After the grease dries up a little it should thicken up...when I received my miner the grease coming out on the sides was still almost wet. Maybe if you can't get your miner to hash up no matter what you do, try to clean off that crappy thermal grease that's there and re-do it yourself. AND REMEMBER, all you need is just a little eraser sized grease and not to over tighten. And I'll still not break even probably.
Did they do worse before you changed the thermal paste? I doubt that it's much of a problem, it seems like most people are getting 280 saturn /560 jupiter without doing that. Some people are having issues, but that's probably not related to heat.
I never trusted the grease that was on there to begin with, so I only powered up for about 20 minutes to make sure everything worked then changed the grease over to what I wanted. It could have been the same, but my Dad who has a Saturn also has great running speeds also and I changed his thermal grease as well. It can't hurt I guess. The stuff that I saw there just didn't look right. Very bright silver stuff that was almost wet. That's what made me change it to begin with.
3 hour steady at 140. slowlllllly increasing
http://eligius.st/~wizkid057/newstats/userstats.php/12kiuiVdTiYUA5ridr9S2GoWuaGUKVZQnP
http://eligius.st/~wizkid057/newstats/userstats.php/12kiuiVdTiYUA5ridr9S2GoWuaGUKVZQnP
confirmed slight performance increase at the pool for 0.98
With 0.98 installed, everytime I restart CGMiner via the web interface, and then quickly start a screen session so I can see the startup, I see the warning "Error in configuration file, partially loaded.". If I start cgminer by hand with '-T -D', I see this error:
Looking at the config file saved by the KnC WebUI, I see this line:
as the last line in the file, before the closing brace. Looks ok to me...
Any one else seeing this?
It doesn't matter if remote management is enabled or not, it still causes the error. It still starts and runs, so it's not a show-stopper, but I don't understand the error.
Code:
[2013-10-29 23:20:43] Invalid config option --api-network: Invalid value
Looking at the config file saved by the KnC WebUI, I see this line:
Quote
"api-network": false
as the last line in the file, before the closing brace. Looks ok to me...
Any one else seeing this?
It doesn't matter if remote management is enabled or not, it still causes the error. It still starts and runs, so it's not a show-stopper, but I don't understand the error.
Thank you KnC for the 0.98 firmware!
Edit: Also thank you ckolivas for your work! Tip incoming!
You're welcome, and thanks! Edit: Also thank you ckolivas for your work! Tip incoming!
I've just created a kncminer branch on my github and have imported their most current code and built my first binary for this. It is not a full firmware and the only thing users may note is the hashrate won't lie with this one, as it won't count hardware errors as hashrate. Also my hardware error count is significantly lower with this on the Saturn dev machine I was sent.
Here's a direct link to the binary:
http://ck.kolivas.org/apps/cgminer/kncminer/cgminer
You'll have to ssh into the kncminer and copy and use that file directly. Hopefully the libraries all match between my dev environment and the default kncminer operating system which means the binary should just work if you run it like on any other OS.
Here's a direct link to the binary:
http://ck.kolivas.org/apps/cgminer/kncminer/cgminer
You'll have to ssh into the kncminer and copy and use that file directly. Hopefully the libraries all match between my dev environment and the default kncminer operating system which means the binary should just work if you run it like on any other OS.
Sorry, but what would be the benefit on doing this? I'm not trying to be an ass, I really don't know.
wtf ?!
they send you untested broken crap for several $k and you have to pay for sending it back ?!
does this even comply with laws concerning B2B ?
+111111111111111
My thoughts exactly. Why the shit do I need to pay for shipping on a dead unit they sent me?
It's not right but would you prefer to wait for them to get round to sorting it out ? Tomorrow means something entirely different in Sweden apparently.
A bit like manana does in Spain.
So I was one of the lucky ones that got my Jupiter which came with 0.95 and flashed it to 0.96 and got it hashing at 555 Gh/s on the miner and 555 Gh/s on the pool (Slush)
0.97 dropped my performance down to 525 Gh/s on the miner and sub 510 average on the pool. I left it a few days to get up to speed but it never did.
0.98 showed no better, if anything out was worse with cgminer disabling and enabling cores all the time. So in desperation I flashed back to 0.96 and everything is back to normal speed.
Weird.
Oh and my miner still shows 1.1.2000 after reboots and never change to the correct date.
Otherwise everything is good
0.97 dropped my performance down to 525 Gh/s on the miner and sub 510 average on the pool. I left it a few days to get up to speed but it never did.
0.98 showed no better, if anything out was worse with cgminer disabling and enabling cores all the time. So in desperation I flashed back to 0.96 and everything is back to normal speed.
Weird.
Oh and my miner still shows 1.1.2000 after reboots and never change to the correct date.
Otherwise everything is good
I've just created a kncminer branch on my github and have imported their most current code and built my first binary for this. It is not a full firmware and the only thing users may note is the hashrate won't lie with this one, as it won't count hardware errors as hashrate. Also my hardware error count is significantly lower with this on the Saturn dev machine I was sent.
Here's a direct link to the binary:
http://ck.kolivas.org/apps/cgminer/kncminer/cgminer
You'll have to ssh into the kncminer and copy and use that file directly. Hopefully the libraries all match between my dev environment and the default kncminer operating system which means the binary should just work if you run it like on any other OS.
Here's a direct link to the binary:
http://ck.kolivas.org/apps/cgminer/kncminer/cgminer
You'll have to ssh into the kncminer and copy and use that file directly. Hopefully the libraries all match between my dev environment and the default kncminer operating system which means the binary should just work if you run it like on any other OS.
Sorry, but what would be the benefit on doing this? I'm not trying to be an ass, I really don't know.
So far with fw 0.9.8 my saturn went from 135GH/s to 203GH/s at the pool. That is the first time I've seen it hit 200GH/s. I wish I could get to 270+ like some of you guys are getting.
Hey, have you used bertmod to see what is going on in there? Post those stats, maybe give you some ideas. Hurry!
and enablecores.bin
I wasn't able to get bertmod to work with the 0.9.8 fw, I don't know if someone else was able to.
While I was on 0.9.4, I saw that one of my boards has a dead vrm and a bad vrm, while my other board has one bad vrm.
enablecores.bin didn't make any changes. It went up to 220GH/s on the CGMiner for a minute, but slowly settled back down to 212GH/s at the CGMiner and 210GH/s at the pool.
Ask KnCMiner for RMA, you'll get a replacement.
Just keep in mind not to send both boards at once, if possible keep one hashing
I've been telling them that since Friday, but since they don't work weekends, I got a reply yesterday to wait for 0.9.8, which I've done today, and still not fixed. I told them it was a hardware issue, but I think they just expect 0.9.8 to be a cure-all. I emailed them again today letting them know it's still not fixed. Gotta wait till tomorrow for a response.
Thank you KnC for the 0.98 firmware!
Thank you as well, I encountered no problems with 0.98 and it increased hashing speed.
Did anyone make a note of the running processes on previous versions, done a top on 0.98 and see a process called monitordcdc.
Was that there before ?
Yes, that process was there before.Was that there before ?
Hello all,
I was having problems with my Saturn where it would only run on 0.95 firmware. 0.96 and 0.97 would only start the first ASIC.
I contacted support and they asked that I apply 0.98 which came out today. I did and the Saturn is humming along nicely. cgminer is now 3.6.6 and extremely stable. The Saturn has been running the same PID for cgminer for the last 2 hours, HW errors are down.
I am getting 209 Gh/s
Jupiter Firmware 0.97 to 0.98 hash rate graph from BTC Guild (each data point is 1 hour):
0.97: about 320 Gh/s, 540 Watts at the wall
0.98: about 530 Gh/s, 640 Watts at the wall
For my troubled Jupiter, 0.98 handles the bad cores MUCH better... 200 Gh/s improvement over 0.97!
Thank you KnCMiner! I recommend trying firmware 0.98, particularly if you have some bad cores.
0.97: about 320 Gh/s, 540 Watts at the wall
0.98: about 530 Gh/s, 640 Watts at the wall
For my troubled Jupiter, 0.98 handles the bad cores MUCH better... 200 Gh/s improvement over 0.97!
Thank you KnCMiner! I recommend trying firmware 0.98, particularly if you have some bad cores.
For the record guys, I'm at 550 with all Jupiters on FW 98 on Eligius. 1 Jupiter has 2 bad boards which I'm RMA'n today- but the 2 good boards are hashing ~270
Man I'm wondering if I just need to powercycle my unit to fix the Eligius issue but now that it's been holding at 560GH/s at BTCguild I'm afraid to touch it!
Well, worst case you'd just set it back up to BTCGuild. (if you power cycled and Eligius still doesnt work).
The steps that fixed mine didn't fix yours.. which is weird. I even cycled 1 of mine just to verify that the quickfix indeed still worked.
So when you're on the Miner tab and do the following steps, what happens after you hit Restart CGMiner? It takes about a min but it grabs the correct URL and starts mining. Err, it should.
After a 30 second power cycle it appeared to work as expected for a few minutes but the Eligius pool started throwing weird errors again:
Code:
[2013-10-29 22:53:07] Rejected 04645fca Diff 58/1 KnC 0 pool 1 (unknown-work)
[2013-10-29 22:53:08] Pool 1 stratum share submission failure
Guess I will stick with BTCGuild for now tired of messing with it
Did anyone make a note of the running processes on previous versions, done a top on 0.98 and see a process called monitordcdc.
Was that there before ?
Was that there before ?
I am at disbelief at the level of misinformation and false claims repeatedly made by KNC support staff.
And I wonder who benefits from such a way of conducting business
It's definitely not me.
I was told yesterday noon (28.10) that my miners are ready and waiting to be picked up same day.
It turns out they will most probably not going to be picked up today either.
And as usual nobody is picking up the phone.
And I wonder who benefits from such a way of conducting business
It's definitely not me.I was told yesterday noon (28.10) that my miners are ready and waiting to be picked up same day.
It turns out they will most probably not going to be picked up today either.
And as usual nobody is picking up the phone.
You have "some weird expectation and all" you do "is keep demanding them" to give you a tracking number...
I feel sorry for you, really...
I've been getting all the response I want via email. Bitching in this forum is the worst way to get answers.
Haha exactly. The guy has some weird expectation and all he does is keep demanding them here. It's like trying to talk to the band, while screaming at a concert
Jump to: