lmao

This is the OFFICIAL PhoenixMiner.

+1 QFT

Wanted to share this because I'm positive more people are bound to have this issue as well...

I have a miner with x2 Nvidia GPUs and x3 AMD GPUs. One is a 1070 8GB and the other is a 1060 6GB.  The 1070 gives me a hashrate of 31 and the 1060 gives me 21. This past weekend Phoenixminer quit working. I was getting the following error message:

main Unknown OpenCL driver version! Hashrate and stale shares may suffer

I thought it was odd since the miner had been working well and I had made NO changes.  I did the following troubleshooting:

-Rebooted miner and ran batch file with admin rights.  No changes
-Recreated the Phoenixminer folder and a new batch file and ran batch file with admin rights. No changes
-Uninstalled Nvidia driver using DDU and re-installed the 461.09 version. Rebooted and tried running miner again. No changes
-Uninstalled Nvidia driver using DDU and re-installed an OLDER driver 451.48. Rebooted and tried running miner again. No changes
-Uninstalled Nvidia driver using DDU and re-installed 461.09 version. Replaced nvml.dll file found in C:\Windows\System32. No changes
-Ran TeamRed Miner. This miner ONLY works with AMD GPUs. The x3 AMD GPUs mined for several hours w/out any issues.
-Ran LOLMiner. This miner failed to run as well. No error messages, it would simply shut off.  I copied and pasted the SAME LOLMiner to my gaming PC, which has a Nvidia GPU. LOLMiner ran w/out any issues.
-Decided to re-install Windows 10 Pro x64 Build 20H2. After 2 hours of configuration and setting up. No changes
-Decided to installed Windows 10 Pro x64 Build 1809. After 2 hours of configuration and setting up, the issue is NOW RESOLVED! 

Conclusion: A Windows update that is found in build 1909 and 20H2 is affecting Nvidia GPUs and mining.

I have a mining rig with 5 GPUs.  Three of them are RX 580 8GB and two of them are Nvidia 1070 8GB / 1060 6GB.  I'm getting the error for the Nvidia ones.

2021.01.31:16:14:08.370: main CUDA version: 11.0, CUDA runtime: 8.0
2021.01.31:16:14:11.974: main Unknown OpenCL driver version! Hashrate and stale shares may suffer

and

2021.01.31:16:14:30.465: GPU3 GPU3: clSetKernelArg (-48)
2021.01.31:16:14:30.468: GPU2 GPU2 initMiner error: Unable to initialize CUDA miner
2021.01.31:16:14:30.562: GPU5 GPU5 initMiner error: Unable to initialize CUDA miner

I tried older drivers - 442.74 - no luck
I tried using the most recent drivers - 461.09 - no luck
I tried replacing the nvml.dll mentioned in a previous post - no luck

This is a relatively new mining rig but it was mining just fine for about one month until two days ago.

This didn't work

Hello.  I'm currently using Phoenix Miner 5.5c and getting this error message with my NVDIA GPUs. I have already copied the nvml.dll to c:\Windows\System32\ and still getting the same error.  Any suggestions?

[/quote]

I'd go to the Nvidia website, grab the latest driver and do a clean install of the driver.

Here's the link:
https://www.nvidia.com/Download/index.aspx?lang=en-us
[/quote]

I did this already and it didn't resolve the issue

Hello.  I'm currently using Phoenix Miner 5.5c and getting this error message with my NVDIA GPUs. I have already copied the nvml.dll to c:\Windows\System32\ and still getting the same error.  Any suggestions?

I'm gonna test it tonight.  I have rig running on an AsRock H81 Pro BTC R2.0. I currently have x6 RX 580 8GB GPUs. But I'm going to disable onboard graphics, run display from a 8GB GPU and swap x2 RX 580 8GB GPUs for x2 RX 580 4GB and see if that works. 

I'll report back.

Interesting.. What about motherboards that have onboard VGA / HDMI? It needs to be disabled and use a GPU with 8GB instead?

Maybe he's trying to flex?  hehe

On another note, I'm seeing people claim they have x5 4GB GPUs and x1 8GB GPU and have no issues mining.  I had a miner with x4 8GB GPUs and x2 4GB GPUs and Claymore 15 would error out saying it could not create the DAG for the 4GB GPUs.

It must be the people mining with 4GB GPUs are not mining coins that are passed Epoch 360.

Which coin are you talking about?

I had a similar issue.  I had a miner doing 190 MH/s and I was mining ETC. When the DAG reached 3.8 GBs, the miner started having errors, glitches, crashing.  I replaced the 4GB cards with 8GB ones and I'm back to doing 190 MH/s. 

Not sure which coin you're mining but it is apparent you're hitting the limits for your 4 GB cards.  Remember GPUs come with different type of RAM, even if same model.  The best card usually have SAMSUNG RAM. ElPidia and Hynix are a mixed bagged. There's a good chance if you're getting 31 MH/s from a 4GB card, it most likely has SAMSUNG RAM.

You're gonna need the following tools:

DDU (display driver uninstaller)
GPU-Z
Crimson Beta 22.19.659.0 (blockchain drivers)

You need a graphics card that has 8GB of RAM or higher. If you have at least one GPU that does not have 8GB of RAM, your miner is gonna have some issues.

I have x6 RX 580 8GBs and I currently get 192 MH/s with Claymore 15 and 182 MH/s with Phoenix5 miner.

Also keep in mind that Windows is stubborn about updating video drivers. You'll have to adjust the OS windows policy to prevent from drivers from being updated.  I had this issue. Windows 10 wants to update from blockchain data drivers to latest which don't mine well at all.