Topic: PhoenixMiner 6.2c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux) - page 132. (Read 784965 times)

here is my daggen.sh for HiveOS

#!/bin/bash

a=$(cat /run/hive/last_stat.json | jq '.params .miner_stats .uptime')

echo "$a"

if [[ "$a" -gt "120" ]];then
        echo "greater than 120 seconds, restarting"
        sudo reboot
fi

You need to install JQ first: sudo apt-get install jq (json parser)

Basically, if the miner has run for more than 2 minutes and sees a new dag, reboot the machine.

There is obviously a chance for a new dag to come before 2 minutes and you wont get a reboot, but this will solve the 99.9% situation as the dag is only generated once every 6000 minutes (100 hours).

Same with all of my 1070 cards. As I understand it, it will become slower each time the DAG increases. I am stable with 26.25 - 26.9 for all 8 of my cards on HiveOS. +100 core, +1200 mem, power limit set to 132 (because these are plugged into a 1200w PSU.)

I'm receiving this error in 5.5c and 5.5b
https://iili.io/fqJ62p.jpg
I'm using HiveOS latest version
RX580 8g

I am trying to restart the miner every 6 hours.
I used -timeout 10 (10 mins)
It shutdown the miner but does not restart it.
Other parameters are default.

Am I doing something wrong?

So on the first dag generation it runs daggen.sh? We need it to run on a new dag coming in, not all dags.

That's 4Gb, whattomine doesn't have 570 4gb and 8gb option. 4gb is no longer relevant for eth. If you have 570 with 8gb use the 580 (which defaults to 8gb)

Hello all,

What are the possible values for “-eres”?

Documentation says default is 2 but what other values would work when miner crashes during dag switch?

RX 570 seems to have 13 to 15 ish hashrate in Whattomine. Is this real or the web page become buggy?

-cvddc doesn't work on 5.5C. already ran it as administrator

there is support for the 500 DAG, links above

+1
And right now with everything that's going on, that doesn't seem like such a bad thing.

Always download from the first page!

PhoenixMiner 5.5d real?

   No, you haven't mined ETC, you still mined ETH - this was possible before epoch 390 because ETC still used the unmodified Ethash algorithm, and as long as you mined on the proper pool address (domain and port number), you will mine whatever the pool wants you to mine. However since the ETH epoch is already 390, when you have -coin etc, you are forcing PhoenixMiner to use the ETChash algorithm, which is incompatible with Ethash and leads to 100% incorrect shares. The lesson is to use the -coin parameter only if you mine ETC or other coin instead of Ethereum.



   The overclock on Nvidia cards with all relatively recent Windows drivers is only possible if PhoenixMiner.exe is started "As Administrator". The settings seem OK but are probably not applied on the first card - you should see some warning messages on the screen (in yellow). You may add the parameter -hstats 2 which will print the actual clocks on the cards and you will see which one are not set properly. Finally, you may try to add -straps 1 for the Nvidia cards, which should give you a little more hashrate.



    First, you need to run PhoenixMiner "As Administrator" for any of the overclocking options to work on Nvidia cards (older Nvidia drivers didn't require this but 30x0 GPUs obviously doesn't work with older drivers).  Second, you should use something like this: -powim -30 -mclock +700 -mcdag 1 for 30x0 cards (these are just example values, you need to find the best ones for your own cards). Note the following things:
 

• Do not try to lower the core clock of the card with -cclock as it often doesn't lower the voltage and the card uses much more power than necessary. Instead use -powlim to limit the power that the card can use. Start with -powlim -30 (this means that the card must use 30% less power than the normal limit) and then try lower values to see how low you can go with the power consumed by the GPU without affecting the hashrate
• Do not give absolute memory clock values in -mclock. Instead use relative value like -mclock +700 which means 700 MHz more than the normal memory clock.
• When the -mcdag 1 parameter is added, PhoenixMiner will keep the GPU memory on the normal clocks during the DAG generation and will apply the memory overclock about 40-50 seconds after the DAG is ready. So you need to wait about a minute to see the final hashrate when testing the memory overclock settings.

   If you want, you may send us the logs via PM (the first two minutes or so are enough). We can't diagnose the problem without any details.




   We will add this to the TODO list for a future release



      Check if you have -coin etc on your command line or config.txt and if so, remove it. Otherwise, post a full log from the start of the miner to the first rejected share (you may send us as PM) in order to help you.





   daggen.sh is provided by the user and you can place whatever you want there (including command to restart the rig, reset the overclock, etc.).

PhoenixMiner 5.5c is officially released. You can find the download link and the checksums in the first post of this thread.

It is the same release as the one that was posted previously as release candidate, so if you are already running it, there is no need to upgrade.

What we need is daggen.sh to just reset the whole computer. Most of us are using HiveOS and such for overclocking. I guess I can use it like you've set. However just having -NewDagScript that runs NewDag.sh would be ideal. Let us handle it from there - let it be reset or just miner reset in HiveOS.

Where is this daggen.sh btw? should it have been in the compressed file?

thanks one of the problem is that is hard in my country to find those fans and the shipping cost can be a lot.
since the other time it said something about driver error, i thought it might be another thing.

so basically it can be a damaged fan?
thanks
[/quote]

Check paste and thermal pads on it.  As long as you keep it below 80c it will run a very long time.  If its underwarranty maybe you can get it covered in time...but gpus are short everywhere. 

You can always remove the stock plastic cover and fans and slap a larger 12v fan (will have to power off psu or mobo) and zip tie it to the card.  Ive got some older cards that ive done this too.  Fans speed doesnt matter as long as you keep temps down.  Since my gpus shut off when they exceed the 80C for nvidia and 85C for AMD, I dont worry about them.
[/quote]

thanks one of the problem is that is hard in my country to find those fans and the shipping cost can be a lot.
since the other time it said something about driver error, i thought it might be another thing.

so basically it can be a damaged fan?
thanks

You can always remove the stock plastic cover and fans and slap a larger 12v fan (will have to power off psu or mobo) and zip tie it to the card.  Ive got some older cards that ive done this too.  Fans speed doesnt matter as long as you keep temps down.  Since my gpus shut off when they exceed the 80C for nvidia and 85C for AMD, I dont worry about them.

hi guys

one of the msi ocv1 gtx 1060 6gb was getting hot, up to 92ºC.
i says 0% fan. i think one of the fans was in need of lub. also, this model when you open it to lub the fans, you need to remove the heatsink so it would need a repaste.
so i thought: well, since ill change the thermal paste for a new one, plus lub the fans, the problem will be gone.

but no, it is still hot, and also says :
gpu 2, unable to get fan speed. unknown error (999)

it can mine, with the expected hash rate etc, but thats too hot (90ish ºC)
so my guess is the paste and fans are working correctly, but maybe the fan speed is wrong and since it cant get the fan speed, it cant control it so it gets hot?
is there any solution to this problem? or someone else got this too?
thanks