|
601
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 14, 2015, 07:35:36 PM
|
Alright, do perform a git pull n test it out.
Basically, if I have it set correctly, Ive modified KNC's loop to see if either DCDC are supplying current, if so then it doesnt kick it over to my loop(the custom resetting loop)
If all is correct, it will only kick over to my custom reset loop when both dcdc's on a die are registering current below the threshold of 5 amps.
ok, I'll check it now looks like it is still going into that loop, I haven't had a failed soft reset yet -- otherwise it's running good sometimes it scrolls "KNC 4b[5:1:55]: Got nonce for unknown work in slot 0" before the reset, other times not (always been like that)
[2015-08-14 13:52:24] Die 5-1 requires restartAttempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439578355,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Detected AMPS DCDC0 0 Detected AMPS DCDC1 0 [2015-08-14 13:53:05] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 13:58:24] Die 5-1 requires restartAttempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439578716,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Detected AMPS DCDC0 0 Detected AMPS DCDC1 0 [2015-08-14 13:59:06] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 14:04:24] Die 5-1 requires restartAttempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439579075,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Detected AMPS DCDC0 0 Detected AMPS DCDC1 0 [2015-08-14 14:05:06] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 14:10:19] Die 5-1 requires restartAttempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439579427,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Detected AMPS DCDC0 0 Detected AMPS DCDC1 0 [2015-08-14 14:10:57] Die 5-1 restarted Um.... weird, because the current registered through the DCDC's is 0 for both ... yet YOU ARE SURE its hashing?? Thats crazy! its impossible! LOL Let me add a current measure output to the KNC loop then paste the output of script. wait up. |
|
|
|
|
602
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 14, 2015, 06:28:43 PM
|
|
Alright, do perform a git pull n test it out.
Basically, if I have it set correctly, Ive modified KNC's loop to see if either DCDC are supplying current, if so then it doesnt kick it over to my loop(the custom resetting loop)
If all is correct, it will only kick over to my custom reset loop when both dcdc's on a die are registering current below the threshold of 5 amps.
|
|
|
|
|
603
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 14, 2015, 05:41:41 PM
|
GenTarkin -- looks like after one soft reset fail it is going straight to hard reset not the loop.
btw how can I delete this log and start over?
[2015-08-14 08:39:44] Die 5-1 requires restart
Attempting softreset of ASIC# 5 DIE# 1
{ "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } }
STATUS=S,When=1439559594,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0|
[2015-08-14 08:40:24] Die 5-1 restarted
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4
Moving on with dead die test, no manual disabled die found
[2015-08-14 08:45:38] Die 5-1 requires restart
Attempting softreset of ASIC# 5 DIE# 1
KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } }
Soft reset failed, initiatng hard reset
Stopping bfgminer.
Power cycling ASIC# 5
INFO: Attempt to power down dc/dc
INFO: Attempt to power UP dc/dc
Starting bfgminer.
[2015-08-14 08:46:45] Die 5-1 restarted
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4
Moving on with dead die test, no manual disabled die found
[
Thats intentional, if it fails a soft reset issuing the waas -s command, it will immediately perform hard reset. The reason is, from watching my titan, in all cases where it failed waas -s it could never be brought back unless a hard reset was performed. Its working as I designed it =) ... Remember, the waas -s command should work even if there is pool issues, therfore like I said, waas -s failing needs a hard reset. KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } ^ means waas -s failed. I guess, if it really is problematic for your rig, I could have it just skip the waas -s failed / success checking and simply rely on current readings from the DCDC's to measure if soft resets worked on the die. But that could be potentially time loss because its running the loop for 5 minutes. Can you confirm if just soft reset spamming this die actually brings it back, 100% of the time? If so, Ill upload a mod for ya that disables waas -s success/fail checking. Also, regarding logs, I just looked at the log directory, the monitordcdc.log is not getting large for me. So, you shouldnt have to worry bout space consumption(of course this depends on how misbehaving ur dies are in a given timeframe =P. I think they get cycled out too though =). If 'catting' the log file is too large for you then just use tail to return n lines from the end of it like tail -n 25 /var/log/monitordcdc.log spamming it with soft resets sometimes does bring it back, that is the default way that KNC handles it so I have seen them come back, surprisingly Also the soft resets are occurring about 6-7 min apart -- it's like this up the whole log, everywhere I checked. Seems odd to be that regular, isn't 6-7 min about the time it takes to run thru the loop?? thx for the tip: tail -n 25 /var/log/monitordcdc.log [2015-08-14 08:10:24] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439557835,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:11:05] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:16:22] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439558190,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:17:00] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:22:09] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439558537,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:22:47] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:28:01] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439558890,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:28:41] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:33:52] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439559239,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:34:29] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:39:44] Die 5-1 requires restart I just thought of something based on your log... You know what?! .... KNC's main loop that tests DCDC current does a DCDC 0 / 1 OR comparison as well. In your case KNC's loop is sending your 1 failed DCDC on that die to my loop. So that particular die is ALWAYS gonna get at least a soft reset. If I change KNC's loop to do an AND comparison like my loop then that die should work just fine & dandy at 150 w/o a soft loop ever being needed until both DCDC's drop below the threshold. The question is... what affect would that cause on other peoples titans or KNC's original loop intent. I cant think of any side affects atm. We know that, so far, both DCDC's register near 0 current when the die needs a soft or hard reset. I think in EVERY case, if a die was humming along just fine and one of the DCDC's stopped for w/e reason ... the other DCDC would drop because the die wouldnt be hashing at all in that case. So ... I THINK doing an AND comparison of the DCDC's in KNC's loop "should" work. Thoughts? Thank you SO MUCH!!! for the donation man! I greatly greatly appreciate it =) I couldn't think of a reason why it would be going thru your loop, however that makes a lot of sense. Let's try it, I have 3 good cubes, a cube with 2 dead dies, and the soft reset cube on this rig, so we should be able to test it pretty good. there's also a hard reset cube on this rig every few days. I am going to make a back up image in case it goes haywire, but I doubt I'll need it Alright, Ill let ya know when to do a git pull =) |
|
|
|
|
604
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 14, 2015, 04:27:57 PM
|
GenTarkin -- looks like after one soft reset fail it is going straight to hard reset not the loop.
btw how can I delete this log and start over?
[2015-08-14 08:39:44] Die 5-1 requires restart
Attempting softreset of ASIC# 5 DIE# 1
{ "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } }
STATUS=S,When=1439559594,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0|
[2015-08-14 08:40:24] Die 5-1 restarted
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4
Moving on with dead die test, no manual disabled die found
[2015-08-14 08:45:38] Die 5-1 requires restart
Attempting softreset of ASIC# 5 DIE# 1
KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } }
Soft reset failed, initiatng hard reset
Stopping bfgminer.
Power cycling ASIC# 5
INFO: Attempt to power down dc/dc
INFO: Attempt to power UP dc/dc
Starting bfgminer.
[2015-08-14 08:46:45] Die 5-1 restarted
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4
Moving on with dead die test, no manual disabled die found
[
Thats intentional, if it fails a soft reset issuing the waas -s command, it will immediately perform hard reset. The reason is, from watching my titan, in all cases where it failed waas -s it could never be brought back unless a hard reset was performed. Its working as I designed it =) ... Remember, the waas -s command should work even if there is pool issues, therfore like I said, waas -s failing needs a hard reset. KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } ^ means waas -s failed. I guess, if it really is problematic for your rig, I could have it just skip the waas -s failed / success checking and simply rely on current readings from the DCDC's to measure if soft resets worked on the die. But that could be potentially time loss because its running the loop for 5 minutes. Can you confirm if just soft reset spamming this die actually brings it back, 100% of the time? If so, Ill upload a mod for ya that disables waas -s success/fail checking. Also, regarding logs, I just looked at the log directory, the monitordcdc.log is not getting large for me. So, you shouldnt have to worry bout space consumption(of course this depends on how misbehaving ur dies are in a given timeframe =P. I think they get cycled out too though =). If 'catting' the log file is too large for you then just use tail to return n lines from the end of it like tail -n 25 /var/log/monitordcdc.log spamming it with soft resets sometimes does bring it back, that is the default way that KNC handles it so I have seen them come back, surprisingly Also the soft resets are occurring about 6-7 min apart -- it's like this up the whole log, everywhere I checked. Seems odd to be that regular, isn't 6-7 min about the time it takes to run thru the loop?? thx for the tip: tail -n 25 /var/log/monitordcdc.log [2015-08-14 08:10:24] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439557835,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:11:05] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:16:22] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439558190,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:17:00] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:22:09] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439558537,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:22:47] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:28:01] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439558890,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:28:41] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:33:52] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439559239,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| [2015-08-14 08:34:29] Die 5-1 restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-14 08:39:44] Die 5-1 requires restart I just thought of something based on your log... You know what?! .... KNC's main loop that tests DCDC current does a DCDC 0 / 1 OR comparison as well. In your case KNC's loop is sending your 1 failed DCDC on that die to my loop. So that particular die is ALWAYS gonna get at least a soft reset. If I change KNC's loop to do an AND comparison like my loop then that die should work just fine & dandy at 150 w/o a soft loop ever being needed until both DCDC's drop below the threshold. The question is... what affect would that cause on other peoples titans or KNC's original loop intent. I cant think of any side affects atm. We know that, so far, both DCDC's register near 0 current when the die needs a soft or hard reset. I think in EVERY case, if a die was humming along just fine and one of the DCDC's stopped for w/e reason ... the other DCDC would drop because the die wouldnt be hashing at all in that case. So ... I THINK doing an AND comparison of the DCDC's in KNC's loop "should" work. Thoughts? Thank you SO MUCH!!! for the donation man! I greatly greatly appreciate it =) |
|
|
|
|
605
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 14, 2015, 02:50:01 PM
|
|
So, I got a question for yall, in the status page.
What is the "Temperature" column? on my titan its substantially lower than DCDC temp average.
Is it the ASIC temp or some sort of ambient temp?
I dont see how it would be the ASIC itself.... since mines so cool and I would expect the chips to run quite hot.
If its the actual ASIC temp, I *may* be able to code in a temperature protection for the chip. But if its not then, I wouldnt be able to.
|
|
|
|
|
606
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 14, 2015, 02:33:44 PM
|
GenTarkin -- looks like after one soft reset fail it is going straight to hard reset not the loop.
btw how can I delete this log and start over?
[2015-08-14 08:39:44] Die 5-1 requires restart
Attempting softreset of ASIC# 5 DIE# 1
{ "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } }
STATUS=S,When=1439559594,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0|
[2015-08-14 08:40:24] Die 5-1 restarted
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4
Moving on with dead die test, no manual disabled die found
[2015-08-14 08:45:38] Die 5-1 requires restart
Attempting softreset of ASIC# 5 DIE# 1
KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } }
Soft reset failed, initiatng hard reset
Stopping bfgminer.
Power cycling ASIC# 5
INFO: Attempt to power down dc/dc
INFO: Attempt to power UP dc/dc
Starting bfgminer.
[2015-08-14 08:46:45] Die 5-1 restarted
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2
Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4
Moving on with dead die test, no manual disabled die found
[
Thats intentional, if it fails a soft reset issuing the waas -s command, it will immediately perform hard reset. The reason is, from watching my titan, in all cases where it failed waas -s it could never be brought back unless a hard reset was performed. Its working as I designed it =) ... Remember, the waas -s command should work even if there is pool issues, therfore like I said, waas -s failing needs a hard reset. KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } ^ means waas -s failed. I guess, if it really is problematic for your rig, I could have it just skip the waas -s failed / success checking and simply rely on current readings from the DCDC's to measure if soft resets worked on the die. But that could be potentially time loss because its running the loop for 5 minutes. Can you confirm if just soft reset spamming this die actually brings it back, 100% of the time? If so, Ill upload a mod for ya that disables waas -s success/fail checking. Also, regarding logs, I just looked at the log directory, the monitordcdc.log is not getting large for me. So, you shouldnt have to worry bout space consumption(of course this depends on how misbehaving ur dies are in a given timeframe =P. I think they get cycled out too though =). If 'catting' the log file is too large for you then just use tail to return n lines from the end of it like tail -n 25 /var/log/monitordcdc.log |
|
|
|
|
607
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 10:11:26 PM
|
Hi GenTarkin can you please list all the changes from the original img from KNCMINER 2.0 to your latest image? Please list it here or preferably on the github?
Thanks
Vegas
Well, for my latest changes including the smarter soft / hard reset ... I havent made a new release out of those. But if you go through all the releases on my github, I have notes for all the changes made since the previous release. At some point when the current test is confirmed working Ill make a new release and post change notes to it =) |
|
|
|
|
608
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 09:37:36 PM
|
If GenTarkin or anyone else could provide input, I would appreciate it.
I have a two titans that were working solid, one was 330 the other 320, 1 die dead on one 2 dies dead on the 2nd. After shipment, there are 4 dead die's, 1 half working die @ 200mhz, and 2-3 dies that work when reset, but slowly gain in amperage from standard area around 38-42 (300-325) all the way up to around 52(and over 100C) when they cutoff and everything goes 0.00. I am unsure how this occurred as everything was working fine prior to shipping. I thought at first it was the diff brand PSU's being used, but they are excellent brand/models and should not be causing any issues.
Is the die's slowly gaining amperage a possibility of heatsinks getting knocked loose in shipping, or do you suspect something else?
I will be able to test your fw on the units tomorrow and provide results. In the mean time I would like to brainstorm possibilities for resolutions.
My firmware is not designed to fix anything that is related to actually damaged dies. You can give it a shot but any dies which misbehave please flag as OFF or to the speed they actually do run. It sounds like the one thats climbing in temp maybe needs HSF reseated, that is of course if only its the dies that are overheating. |
|
|
|
|
609
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 07:59:27 PM
|
here's the screen shot, trying your latest commit now  Wait, this is while bfgminer says its configured successfully?! ... cuz according to screenshot its not configured successfully at all, its not hashing ... thats why it got hard reset. yeah it an odd one, it is hashing -- that die never displays -- if I turn off all the good dies and let that die run I get aprox 10 mh/s on that bad die btw that latest commit fixed it, the soft resets are random as needed -- before they occurred every 37 secs or so while in that loop I bet whats going on is one of the DCDC's for that die is completely shot and the other is barely limping along. The shot one would of course fail the threshold test. Now I have the script doing an AND comparison of threshold results for both DCDC's in the pair for the die. If both are fail the threshold test then it schedules the hard reset. I think dies that need a hard reset will have the DCDC's not putting any current out anyways so thats good, they would be scheduled for hard reset as desired. Dies such as urs still have a bit of current through at least one DCDC so we wouldnt want a hard reset on it. For those miners who have completely dead dies, they should be set to OFF in the webgui that way my script wont constantly try to do hard resets. Also, keep in mind this will still schedule a hard reset in the event that there has been no work from pool in over 6mins or so(once it enters the die reset loop). Hopefully this will work as expected for needed hard resets, thats what Im waiting to test next on my rig haha! Also, eventually I will scale back the log output during that loop, once we confirm stuff works so the log doesnt get huge. |
|
|
|
|
610
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 07:50:49 PM
|
https://github.com/GenTarkin/Titan/releases/tag/v.93New release published!! ---details--- Login through SSH & webgui is now: admin/admin (should be anyways LOL! Hope I updated it correctly) =P test it peoples! =) If soft die reset fails then initiate hard reset sequence (power off cube, restart bfgminer) Instead of setting Dies w/ overheating DCDC's to OFF, now scales down 25mhz each check until DCDC's are under temp threshold.. If goes to 100 then sets die to OFF. Added more temps to temp threshold setting, including all numbers between 90 & 95 *default DCDC temp monitoring settings are: ENABLED / 90C ---details--- Please test everyone, Ill fix as issues arise. Please Please donate =) Helps fuel my motivation to continue improving upon stuff =) instead of burning the new img file I tried doing a git pull: cd knc-asic git stash save --keep-index (didn't need this line on all rigs) git pull cd ./update-webgui.sh seems to work but did I miss anything?? If I wanted to update my pi to your firmware without win32diskimager flashing the full image, what would be my commands? git pull ./update-webgui.sh thats it? How do I specify your git repository? Im a little confused as TXSteve added that 'didnt need this line on all rigs' comment so not sure if I need that git stash line or not. I need to upgrade a remote box to test with this fw but dont have physical access. Im assuming this will work? Yeah, ull want to at least do a full flash to new SD card once, then can do git pulls =) |
|
|
|
|
612
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 06:06:38 PM
|
cool, I'll check it out
It looks good, however I did see one small issue. I raised the MHz on the flaky soft reset die to stress it and bfgminer wouldn't load. It got stuck in a hard reset loop when the die failed. I was able to change the advanced setting to get out of the loop, but maybe need to delay the hard reset routine for a few min to let bfgminer load when there is a bad die that's overclocked or not turned off ok I just noticed something ... If I get a bunch of "Got nonce for unknown work in slot xx" errors scrolling then it goes straight to hard reset ... (this may tie in with the bgfminer not loading loop). other than that everything is working fine I'll get another cup of coffee & watch this for awhile, see if I can notice anything else that might help you troubleshoot it ...... ok this time the "Got nonce for unknown work in slot xx" errors didn't trigger an instant hard reset and it reset successfully, must have been coincidence earlier ok looks like just getting random hard resets instead of running thru the soft reset loop. Most of the time it runs thru the loop. If I disable the flaky die there aren't any problems btw, that die is set at 150 MHz so it has less amp draw So, for the die thats at 150mhz ... are either of the DCDC's less than 5amp output for more than a few mins? If so, thats why its carrying out the hard reset. Paste contents of /var/log/monitordcdc.log if ya could =) thanks If it turns out that one of the DCDC's is putting out current but the other is not, I may just change the comparison to DCDC1 AND DCDC2 less than threshold rather than OR. Because both will be less than 5 if the die is not working at all. I bet that's it, in adv settings one die is showing zero amps, the other die isn't even listed -- I know it's hashing because I turned off the good dies and got 10 mhs in bfgminer for that bad die, about half what it's suppose to be here's the end of the log, it's getting pretty big Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "200" } } STATUS=S,When=1439479313,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "200" } } STATUS=S,When=1439479352,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "200" } } Soft reset failed, initiatng hard reset Stopping bfgminer. Power cycling ASIC# 5 INFO: Attempt to power down dc/dc INFO: Attempt to power UP dc/dc Starting bfgminer. [2015-08-13 15:23:46] Die 5- restarted Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 2 Manually disabled die detected, skipping dead die detection. ASIC# 4, DIE# 4 Moving on with dead die test, no manual disabled die found [2015-08-13 15:28:51] Die 5-1 requires restart Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439479743,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439479781,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439479818,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439479857,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439479897,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439479936,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } STATUS=S,When=1439479973,Code=92,Msg=PGA 0 set OK: Die setup Ok; asic 5 die 1 cmd RECONFIGURE,Description=bfgminer 5.2.0| Attempting softreset of ASIC# 5 DIE# 1 KnC: Frequency change FAILED! { "asic_6_voltage": { "die2": "-0.0366" }, "asic_6_frequency": { "die2": "150" } } Soft reset failed, initiatng hard reset I was scrolling thru the log and saw this "Failed multiple soft reset attempts, performing hard reset" -- above it there were 10 soft reset attempts like above -- however there aren't any die configuration failed error messages in bfgminer .... ok I just confirmed in bfgminer -- 10 successful soft restarts and the 11th goes straight to hard restart :| ... it's not detecting a successful restart because of the low/zero amp issue?? Can you try my latest commit? I just uploaded it. Let me know what happens, also could u paste a screenshot of that 150mhz die from advanced screen config? Thanks |
|
|
|
|
614
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 01:26:23 PM
|
cool, I'll check it out
It looks good, however I did see one small issue. I raised the MHz on the flaky soft reset die to stress it and bfgminer wouldn't load. It got stuck in a hard reset loop when the die failed. I was able to change the advanced setting to get out of the loop, but maybe need to delay the hard reset routine for a few min to let bfgminer load when there is a bad die that's overclocked or not turned off ok I just noticed something ... If I get a bunch of "Got nonce for unknown work in slot xx" errors scrolling then it goes straight to hard reset ... (this may tie in with the bgfminer not loading loop). other than that everything is working fine I'll get another cup of coffee & watch this for awhile, see if I can notice anything else that might help you troubleshoot it ...... ok this time the "Got nonce for unknown work in slot xx" errors didn't trigger an instant hard reset and it reset successfully, must have been coincidence earlier ok looks like just getting random hard resets instead of running thru the soft reset loop. Most of the time it runs thru the loop. If I disable the flaky die there aren't any problems btw, that die is set at 150 MHz so it has less amp draw So, for the die thats at 150mhz ... are either of the DCDC's less than 5amp output for more than a few mins? If so, thats why its carrying out the hard reset. Paste contents of /var/log/monitordcdc.log if ya could =) thanks If it turns out that one of the DCDC's is putting out current but the other is not, I may just change the comparison to DCDC1 AND DCDC2 less than threshold rather than OR. Because both will be less than 5 if the die is not working at all. |
|
|
|
|
615
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 01:24:28 PM
|
cool, I'll check it out
It looks good, however I did see one small issue. I raised the MHz on the flaky soft reset die to stress it and bfgminer wouldn't load. It got stuck in a hard reset loop when the die failed. I was able to change the advanced setting to get out of the loop, but maybe need to delay the hard reset routine for a few min to let bfgminer load when there is a bad die that's overclocked or not turned off ok I just noticed something ... If I get a bunch of "Got nonce for unknown work in slot xx" errors scrolling then it goes straight to hard reset ... (this may tie in with the bgfminer not loading loop). other than that everything is working fine I'll get another cup of coffee & watch this for awhile, see if I can notice anything else that might help you troubleshoot it ...... ok this time the "Got nonce for unknown work in slot xx" errors didn't trigger an instant hard reset and it reset successfully, must have been coincidence earlier The hard reset most ikely took place because that die was probably not able to be reconfigured via waas -s and that command failed, whenver that command fails it will do a hard reset. That command should never return failed. If it does that means soft resets wont bring it back. The 2nd time it may have just been that the soft reset worked for it =) As far as ur first issue ... again, if u raised the die to a mhz where it wont even acknowledge waas -s commands or it returns failed it will do a hard reset no matter what. Can you post the contents of ur log file where this all took place so I can guauge to see if my guesses are correct? I also had some interesting behaviour on my titan this morning. One die went down, 10 soft resets failed then it did a hard reset on the cube - worked as designed. Then another die went "down" shortly after and it completely failed waas -s command so it carried out the hard reset. Worked as designed as well =) |
|
|
|
|
616
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 03:34:23 AM
|
TXSteve, so the way these things work is every loop of the monitoring script it scans how much current is going through the DCDC's and if under 5 amps is going through it basically incriments the value /var/run/dieXX once /var/run/dieXX reaches over the threshold variable it takes action to reset the die. So...basically if the ASICs dont have work due to a shitty pool configuration like when renting... that /var/run/dieXX will get incrimented, if the pool is shitty enough and it increments over threshold ... no matter what... one of the reset actions will be taken. Im not exactly sure whats a good way to recode this so it can differentiate between possible pool comm issues vs an actual die needing reset. I could impliment maybe 2 thresholds, one being a lower one for the pool issues then another maybe being double of the first one?  and if thats reached then it performs a hard reset. I dont know, what do you think? Or anyone else care to chime in? Im just kinda doing a ton of trial and error here LOL! How this all was originally(the way KNC set it up) was on threshold and if that threshold was hit it just performed a soft reset for an eternity. In the short term, I was thinking, pool issues aside. Now since we have this other stupid mode of failure where the waas command succeeds but it really doesnt bring the die back to life .... I could do another loop which would attempt soft resets up to say 5x sleep after each attempt, run the dcdc function to see if current is over 5A and each incriment its not, of course increase /var/run/dieXX ... then inside that same loop if /var/run/dieXX goes over the threhold then start a hard reset. I think that would solve the issue I ran into just today =) how about this? An "auto on/auto off" button & a "manual reset" button. On problematic rigs we turn auto reset off, & use manual reset as needed. It still beats physically powering down the rig, and restarting it ...just a thought I only have one die with these flaky soft resets, so it may not be a huge overall problem Trying to keep this as automated as possible and least user invasive as possible =), its hard for me to code custom stuff for webgui since I dont have much experience w/ all the crazy shit they have going on involving it. LOL this v.93 seems to be running pretty good, even the flaky soft reset die seems to have stabilized after several hard resets when bfgminer randomly shuts down I attribute that to hard resets Yeahp, bfgminer cant be running in order for a proper dcdc power down / up ... dont know if it has to do w/ bus traffic or what, but seems the dcdc power down / up does something but bfgminer will continually ignore them. If those loops are too big a pain the neck to implement I wouldn't worry about them Well, I rewrote the soft / hard reset code =) Basically, once it detects a die in error via /var/run/dieXX It calls the reset die function, it first attempts a soft reset... if that fails right off the bat via waas -s failing then it performs a hard reset. (waas -s shouldnt fail because of pool comm errors) If waas -s succeeds then it calls bfgminer to perform its internal die reconfiguration update Then script waits 30 seconds and measures the current output of die in question If either of the DCDC's are below current treshold then it incriments error count It will loop through the soft die resets up to 10 times, if it fails 10x then it performs a hard reset. So, it gives the die roughly 5-6 mins to "work" ... meaning have current flowing through it greater than the threshold, via soft resets. If that fails it hard resets. What ya think? I updated github w/ the changes if you wanna test. I can say it seems the soft reset logic is working. I have yet to be able to see a hard reset take place, I have to actually wait till my unit acts up to confirm hard reset functionality LOL! |
|
|
|
|
617
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 13, 2015, 01:01:48 AM
|
TXSteve, so the way these things work is every loop of the monitoring script it scans how much current is going through the DCDC's and if under 5 amps is going through it basically incriments the value /var/run/dieXX once /var/run/dieXX reaches over the threshold variable it takes action to reset the die. So...basically if the ASICs dont have work due to a shitty pool configuration like when renting... that /var/run/dieXX will get incrimented, if the pool is shitty enough and it increments over threshold ... no matter what... one of the reset actions will be taken. Im not exactly sure whats a good way to recode this so it can differentiate between possible pool comm issues vs an actual die needing reset. I could impliment maybe 2 thresholds, one being a lower one for the pool issues then another maybe being double of the first one? and if thats reached then it performs a hard reset. I dont know, what do you think? Or anyone else care to chime in? Im just kinda doing a ton of trial and error here LOL! How this all was originally(the way KNC set it up) was on threshold and if that threshold was hit it just performed a soft reset for an eternity. In the short term, I was thinking, pool issues aside. Now since we have this other stupid mode of failure where the waas command succeeds but it really doesnt bring the die back to life .... I could do another loop which would attempt soft resets up to say 5x sleep after each attempt, run the dcdc function to see if current is over 5A and each incriment its not, of course increase /var/run/dieXX ... then inside that same loop if /var/run/dieXX goes over the threhold then start a hard reset. I think that would solve the issue I ran into just today =) how about this? An "auto on/auto off" button & a "manual reset" button. On problematic rigs we turn auto reset off, & use manual reset as needed. It still beats physically powering down the rig, and restarting it ...just a thought I only have one die with these flaky soft resets, so it may not be a huge overall problem Trying to keep this as automated as possible and least user invasive as possible =), its hard for me to code custom stuff for webgui since I dont have much experience w/ all the crazy shit they have going on involving it. LOL this v.93 seems to be running pretty good, even the flaky soft reset die seems to have stabilized after several hard resets when bfgminer randomly shuts down I attribute that to hard resets Yeahp, bfgminer cant be running in order for a proper dcdc power down / up ... dont know if it has to do w/ bus traffic or what, but seems the dcdc power down / up does something but bfgminer will continually ignore them. |
|
|
|
|
618
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 12, 2015, 09:13:57 PM
|
TXSteve, so the way these things work is every loop of the monitoring script it scans how much current is going through the DCDC's and if under 5 amps is going through it basically incriments the value /var/run/dieXX once /var/run/dieXX reaches over the threshold variable it takes action to reset the die. So...basically if the ASICs dont have work due to a shitty pool configuration like when renting... that /var/run/dieXX will get incrimented, if the pool is shitty enough and it increments over threshold ... no matter what... one of the reset actions will be taken. Im not exactly sure whats a good way to recode this so it can differentiate between possible pool comm issues vs an actual die needing reset. I could impliment maybe 2 thresholds, one being a lower one for the pool issues then another maybe being double of the first one? and if thats reached then it performs a hard reset. I dont know, what do you think? Or anyone else care to chime in? Im just kinda doing a ton of trial and error here LOL! How this all was originally(the way KNC set it up) was on threshold and if that threshold was hit it just performed a soft reset for an eternity. In the short term, I was thinking, pool issues aside. Now since we have this other stupid mode of failure where the waas command succeeds but it really doesnt bring the die back to life .... I could do another loop which would attempt soft resets up to say 5x sleep after each attempt, run the dcdc function to see if current is over 5A and each incriment its not, of course increase /var/run/dieXX ... then inside that same loop if /var/run/dieXX goes over the threhold then start a hard reset. I think that would solve the issue I ran into just today =) how about this? An "auto on/auto off" button & a "manual reset" button. On problematic rigs we turn auto reset off, & use manual reset as needed. It still beats physically powering down the rig, and restarting it ...just a thought I only have one die with these flaky soft resets, so it may not be a huge overall problem Trying to keep this as automated as possible and least user invasive as possible =), its hard for me to code custom stuff for webgui since I dont have much experience w/ all the crazy shit they have going on involving it. LOL |
|
|
|
|
619
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 12, 2015, 08:41:48 PM
|
TXSteve, so the way these things work is every loop of the monitoring script it scans how much current is going through the DCDC's and if under 5 amps is going through it basically incriments the value /var/run/dieXX once /var/run/dieXX reaches over the threshold variable it takes action to reset the die. So...basically if the ASICs dont have work due to a shitty pool configuration like when renting... that /var/run/dieXX will get incrimented, if the pool is shitty enough and it increments over threshold ... no matter what... one of the reset actions will be taken. Im not exactly sure whats a good way to recode this so it can differentiate between possible pool comm issues vs an actual die needing reset. I could impliment maybe 2 thresholds, one being a lower one for the pool issues then another maybe being double of the first one? and if thats reached then it performs a hard reset. I dont know, what do you think? Or anyone else care to chime in? Im just kinda doing a ton of trial and error here LOL! How this all was originally(the way KNC set it up) was on threshold and if that threshold was hit it just performed a soft reset for an eternity. In the short term, I was thinking, pool issues aside. Now since we have this other stupid mode of failure where the waas command succeeds but it really doesnt bring the die back to life .... I could do another loop which would attempt soft resets up to say 5x sleep after each attempt, run the dcdc function to see if current is over 5A and each incriment its not, of course increase /var/run/dieXX ... then inside that same loop if /var/run/dieXX goes over the threhold then start a hard reset. I think that would solve the issue I ran into just today =) |
|
|
|
|
620
|
Alternate cryptocurrencies / Mining (Altcoins) / Re: Swedish ASIC miner company kncminer.com
|
on: August 12, 2015, 07:45:40 PM
|
|
Came across a behviour issue on my Titan, I just now noticed it issues a soft reset via waas and that returns success yet it still fails to RECONFIGURE successfully in bfgminer. So, a hard reset would be inevitable and really theres no way to differentiate at this point between a soft reset full success vs failure =/
May have to reimpliment hard reset no matter what.
EDIT: yeah what a bummer, its attemping multiple soft resets w/ no success yet waas doesnt fail. I dont know if there is a way around that at this point, may have to revert just hard resets, bummer.
EDIT: rethinking it out, I may have another way to detect die status as a fallback. Coding that in will be tricky, will have to wait till later =)
Damn these things for failing so many different ways ROFL!
|
|
|
|
|
Show Posts
