Invalid shares как исправить

A share is an outcome of the mining process. This result is sent from a miner to a pool and contributes to finding a new block on the blockchain (confirming transactions in the process).

You can think of a share as a mathematical problem solved with a random number. 

Miner tries many various random numbers until the solution is valid. The more random numbers a miner can try out in the given time, the greater the probability of solving the mathematical problems accurately.

Miners are solving mathematical problems and sending the solutions back to the pool. If they solve the problem correctly, they get (rewarded) a new block reward.

But why do you get invalid shares when mining on HiveOS? Is one invalid share something you should worry about? Could this be due to your flight sheet?

Maybe you tried to overclock and “underclock,” to lower the memory clock, etc., but you just can’t fix it.

Ideally, the normal amount of invalid shares per 24 hours per card should be 0. If there is an invalid share reported constantly, you need to turn the clocks down a little or give it more juice – more power or voltage or both.

But let’s just dive into details and solve this nasty invalid share problem!

What Are Invalid Shares on HiveOS and Why Do They Happen?

Getting invalid shares means that the calculation it submitted was wrong. Let’s say it’s 2+2, and your GPU said 5.

Invalid shares might be caused by either a problem with the miner program (bad settings) or your hardware (the most common ones are unstable overclock and incompatible driver versions).

Some pools pay like 5% per invalid share; others dont pay at all.

If you start to get invalid shares from only one card, and the rest is good (before you notice it and reboot), you may be wondering why only one card gives invalid shares? After all, all cards are the same, and they all have the same OC settings.

Well, not all GPUs are the same, even the same brands. People call it the silicon lottery.

So, in this case, it’s always the same GPU even though you have the same settings for all. This one, in particular, is “slower” than the rest.

That’s why it’s recommended to change clock settings to slow it down, and you will see the hashrate also decrease. So if it’s “slower,” it will likely commit fewer errors.

Sometimes you’ll see invalid shares during the day when it gets really hot. No, it’s not always purely coincidence, and temperatures could affect invalid shares.

The offset core could possibly help with temps. The thing with invalid shares is that the memory settings are usually too high or running too hot.

Keep an eye on it and, if needed (if you start seeing lots more shares), lower the memory clock on that card by five, restart the HiveOS rig, and continue this process until it resolves the issue.

Also, when you get invalid shares, try rebooting (without underclocking), and it might resolve the issue fast.

If you are sure your software is up-to-date, all the drivers are updated, and you did not overuse the GPU overclocking, contact the HiveOS support team with the following information:

• Hardware information (CPU, GPU)
• Logs (if you are using our soft, then both HiveOS logs and currency logs for the period that invalid shares started occurring; if you are using some alternative software, try to provide currency logs for the same period). Please send the logs in .txt format, separated (and named differently) for each worker.

What is a solid accepted/rejected share ratio?

It would be best to aim for 100% accepted shares, and 0% rejected shares. 

In the long run, this is virtually impossible since there will always be some latency between the miner and the HiveOS stratum server.

When it comes to types of rejects like an invalid, stale, or duplicate shares, only stale shares should be fine (up to 1%-2%, depending on the algorithm). 

Any other rejected share is a consequence of a misconfigured system and has the potential to be fully removed.

1. Check your internet connection. You can write “net-test” in the HiveOS console and check if you have a high ping. This may be producing a lot of invalid or stale shares.

If you’re on WiFi, switch to Ethernet cable if possible. Or upgrade the WiFi router.

1. Look for the typical OC settings of your cards on the internet. If necessary, start from the default settings and start overclocking slowly until it’s unstable.
2. Lower your memory by 50 until you stop getting invalid shares. This might take a few hours. 

Rebooting doesn’t solve the problem. It just clears the log. When you get invalid shares, turn down the memory setting until it stops throwing them out. Then tweak from there.

1. Try an absolute core clock of 1085, not offset.

How do you configure the absolute core clock if your UI only seems to provide configuration for offset?

Well, when you enter s value over 500, the text over the field switches to absolute.

1. Start mining with no overclock on the RTX cards and then overclock.

When you initially begin mining, there are some DAG issues, and most of these can be solved by running stock to load DAG.

1. Set your overclocking to 0 on each position and restart your rig after 20-30 seconds.

You can set your overclocking again and set the delay to 40. That helped me many times.

A lot of mining software has issues with DAG change with 3060 Ti and 3070 cards. 

My friend had this issue with Phoenix, Gminer (before Gminer actually fixed it, so it’s fine to use now), and occasionally with NBminer. 

The DAG would be corrupted from high OC settings with every epoch change.

Changing my mining software to T-Rex (and now Gminer) fixed this for me. 

The issue could also be related to a power supply:

Invalid Shares Caused by a Power Cut

If you’re quite new to mining, you could experience a problem with invalid shares after a power cut in your house. 

You could be running smoothly with 0 invalid shares for quite some time, but after the power cut, you could come to the nightmare of hundreds of invalid shares on your dashboard when you turn on the system.

You might try multiple overclocking settings, but none will eliminate the issue of the invalid shares popping up.

This also happened to me. I was getting a lot of rejects, tried a lot of OC settings, and finally, I noticed that it was because the ping to the pool was very high. 

I changed pools, and now I have zero rejects or invalids. Things are looking very stable, and the system is cooler than it was before.

Try these OC settings to find out how to make them all run at the same pace:

• Update drivers.
• Update HiveOS.
• Try other OC settings.

Invalids/Rejects are generally caused by overclocks. Try lowering your mem clock by ten until the invalids stop.

• Try other miners.

You’ll get better results from T-Rex or Gminer over NBMiner.

NBMiner uses corporate-level product obfuscation to hide its source code, preventing people from reverse-engineering its code.

• Put the OC preferences to -500 2200 120 and 60 seconds delay. Then increase the memory by 30 (2230-2260-2290-etc) until you get your expected value, or the best one before getting errors.

One time I switched off the Power Limit (it went to 237W), and everything got smooth.

I then began decreasing the limit. 180W got 100% efficiency, and 150W got one invalid share over an hour. Setting to ~155W seems ideal for my rig.

Hope this works for you, and if nothing helps, I’ll try to give more information I’ve gathered in the future.

Recommended reading

Nicehash vs. HiveOS – The Data-Driven Analysis (2022)

HiveOS or Windows for Mining? HiveOS, Here’s Why (2022)