* BEWARE*GTX 1070 mining effects on the card

Hrithik_JayanthHrithik_Jayanth Member Posts: 12
edited June 2017 in Mining
miners who are mining with gtx 1070 ,stop if u dont wanna end up breaking ur card, i will tell u guys what just happened , i came to my room ,turned on the monitor only to find the screen flooded with purple hue, the screen was tinted. i had my gpu mine eth with claymores miner with a few adjustments like increasing memory clock by 650 mhz and reducing the power target by 20 percent or so thus obtaining 31mh/s,keeping the card at a stable 64 degree centigrade,at fan speed being 50 percent, i ve had the gpu run non stop for about 40odd hours . fear struck me hard,i immediately closed the minining prog,and got the gpu to default settings and restarted , now its normal phew . mining can take a toll on your cards ,be wary of my incident
Post edited by Hrithik_Jayanth on

Comments

  • Hrithik_JayanthHrithik_Jayanth Member Posts: 12
    just thought of letting the community know :)
  • EMWEEEMWEE Member Posts: 60
    Your card is fine. Its probably because u decreased the power to much.
  • BmoneyBmoney Member Posts: 12
    I have 10 GTX 1070 and they've been mining 24/7 for 2 weeks now with no issues. Some of them have power target set to 60 and with memory clocked at 700.
  • Hrithik_JayanthHrithik_Jayanth Member Posts: 12
    @EMWEE, I had set it at 60 percent, is that too low?
  • Hrithik_JayanthHrithik_Jayanth Member Posts: 12



    That is the setting that I had used
  • deepminer1deepminer1 Member Posts: 17
    edited June 2017
    I have two rigs one with 6 EVGA FTW and another with 5 ASUS STRIX and one evga SC. Both of the rigs have the same issue if I push the core clock above 50 MHz. and memory above 575 MHZ ( for asus) and 600 MHz for EVGA FTW. I see artifacts .. but everything becomes okay when i restart the machine. power is set at 80%. I think there is something called "fatigue" .. i researched but dint get anything on it.

    i have also noticed that if you have only two GPUs .. you can push it more than one with 6 GPUs. i also want to try out overclocking individual GPUs because i know each chip is different. But i was looking for some program which will do it automatically for me.
  • Hrithik_JayanthHrithik_Jayanth Member Posts: 12
    So what is the solution, should I decrease the memory clock, or should I increase power limit?
    And if it's fatiguing enough to cause such phenomena, is it a good idea to continue mining with it?
    Won't this reduce life span, I really don't wanna lose my card..
  • EMWEEEMWEE Member Posts: 60
    I think u should do a combination. U need to find a memory clock / power limit that works for your card. Not every card is the same. If you get artifacts your pushing your card to much.
  • deepminer1deepminer1 Member Posts: 17
    Bmoney said:

    I have 10 GTX 1070 and they've been mining 24/7 for 2 weeks now with no issues. Some of them have power target set to 60 and with memory clocked at 700.

    How much did it earn you in 2 weeks? which pool ?
  • deepminer1deepminer1 Member Posts: 17
    i have noticed if i increase the power by around 80% its stable. even better with 85%. I have never heard 60% though .. 75 is what i saw and use in sometimes. mostly things are stable at 80%. I m not so worried about the fatigue as I just feel its okay . its all digital in the that world. it will or will not work. if its working its all fine.
  • BmoneyBmoney Member Posts: 12
    edited June 2017
    @deepminer1 I really cant't say for the 1070's alone.. but at the rate I was going at which was 980Mh, I was getting .45 ETH a day on ethermine

    For my 1070's.. I have:
    7 ASUS ROG STRIX, power limit 70% and memory at 700mhz (30.9mh)
    1 Gigabyte 1070 G1 gaming, power limit 60% memory at 750mhz (31mh)
    2 MSI 1070 GAMING X, power limit 60% memory at 700mhz for one (31mh) and 600mhz for the other (30.3mh)

  • deepminer1deepminer1 Member Posts: 17
    Bmoney said:

    @deepminer1 I really cant't say for the 1070's alone.. but at the rate I was going at which was 980Mh, I was getting .45 ETH a day on ethermine

    For my 1070's.. I have:
    7 ASUS ROG STRIX, power limit 70% and memory at 700mhz (30.9mh)
    1 Gigabyte 1070 G1 gaming, power limit 60% memory at 750mhz (31mh)
    2 MSI 1070 GAMING X, power limit 60% memory at 700mhz for one (31mh) and 600mhz for the other (30.3mh)

    What drivers buddy ? Also did you not touch the core frequency ? Also, do temperatures matter? I have everything below 61C. Did you use hit & trial to get to these frequencies?
  • Hrithik_JayanthHrithik_Jayanth Member Posts: 12
    http://www.legitreviews.com/geforce-gtx-1070-ethereum-mining-small-tweaks-great-hashrate-low-power_195451


    That is the thread that I had followed, @Bmoney, even bmoney set it at 60 percent power limit with the same model (g1 gaming)
    Can you tell me your exact memory clock?
    Not the offset, i had it near 9290mhz as you can see in the pic that I posted earlier.

    And how will I know if the new setting that I choose won't start straining the card again, because for the previous setting that I had used, everything was looking good for the first 40 hours, there were no signs of system instability initially, even the temps were good.
  • BmoneyBmoney Member Posts: 12
    edited June 2017
    @deepminer1
    -Nvidia drivers I use: 382.05
    -I left core frequency alone because it doesn't really change the hash rate much
    -These 1070 cards can handle heat.. so you probably won't see it throttle until high 80C. I like to keep my temps at 55C and below for all my cards. Some of my rigs have a small external fan blowing at it to let air circulate. I find it drops some of my cards by 6-9C.
    -I generated those frequencies myself by trial and error. Any rig running more than 3 days stable is a pass in my book.

    @Hrithik_Jayanth
    -My exact memory speed in GPUZ is 2377mhz for my 1070 G1 gaming. I might be wrong but mulitple by 4 will give you the exact memory clock which is 9508mhz
    -Find a good frequency that is stable and let it run to see how long it goes. Also all cards are not the same.. you just have to get lucky with silicon lottery
  • deepminer1deepminer1 Member Posts: 17
    @Bmoney Dude/Dudette ..
    I overclocked my cards individually yesterday and was able to achieve 75% - 700MHz... both my main rigs are mining at ~185 MH/s ( 6 GTX 1070). My total hashing rate is about 430 MH/s for 14 GPUs. Temp are stable all ranging from 48 to 61 deg C. However, my happiness was not there for long. And systems crashed after 4-5 hours. Now i have one running at 181 MHs and one at 178 MHs.
  • lonkodonkolonkodonko Member Posts: 7
    Bmoney said:

    @deepminer1

    -My exact memory speed in GPUZ is 2377mhz for my 1070 G1 gaming. I might be wrong but mulitple by 4 will give you the exact memory clock which is 9508mhz
    -Find a good frequency that is stable and let it run to see how long it goes. Also all cards are not the same.. you just have to get lucky with silicon lottery


    Do you have Samsung or Micron memory on your Gaming G1 cards? I have one Micron and can only oc memory to 8600 stable.
  • deepminer1deepminer1 Member Posts: 17
    I checked the memory brand its mircon for the EVGA and asus. I was confused when I saw two different frequencies on MSI after burner and the GPU-Z. Which one is true?
    I saw a message saying " GPU#1 got incorrect share, if you see this warning often make sure you did not overclock it too much" .. after that I changed the afterburner GPU #1 frequency to +520 MHz. Rest of the GPUs are at 600MHz ( afterburner). I have noticed one thing even from another rig of mine, if i keep the frequency of my GPU #1 a bit lower than rest of the GPUs then its much stabler. I think it has to do something with me accessing the screen via team viewer to monitor. I wish there was a faster method to know it my machine goes down ( I come to know when I see the worker down on the ethermine or get an email which is like after 1-2 hour). I m also getting one HDMI monitor emulator per rig so that the computer thinks there is a monitor and is stable.
  • BmoneyBmoney Member Posts: 12
    @lonkodonko
    My g1 has Samsung memory.

    @deepminer1
    178-181mh for 6 cards is pretty good I'd say. It's better to have it 24/7 stable than get the few extra hash. Losing money if it goes down :)
  • deepminer1deepminer1 Member Posts: 17
    yeah ... my computer got hacked yesterday when I tried installing a mining monitor from this forum. Lost 1.5 weeks of my earning and will live in fear for the future. Gmail hacked, coinbase hacked, litecoin wallet hacked. So frustrated. This was the link.

    http://forum.ethereum.org/discussion/14051/awesome-miner-v2-6#latest
Sign In or Register to comment.