Quote from: Targonis on January 19, 2022, 04:14:46 AM

Quote from: Dyl-I-Y on January 17, 2022, 11:16:17 PM

Hey everyone! Hoping to troubleshoot a MATIC mining issue on my end. My graphic card( AMD Radeon RX 6600 XT ) is on the latest driver(22.1.1 released 1/7/2021 …10 days before this post). I seem to be running into an error and cannot move past 0Mh hashrate, along with the error no CUDA driver found, when using version “PhoenixMiner_5.9d_Windows”. I went downgrading version by version and had some success with version “PhoenixMiner_5.5c_Windows” with a 28.6Mh(people seem to be getting higher on newer versions) but thought I would at least illuminate the issue in hope to use a newer version 5.9. Not sure if this is impacting anyone else with an AMD Radeon RX 6600 XT for a graphics card, or because of the newer driver, but any help is appreciated!
-------------->MINER LOG START<--------------

Phoenix Miner 5.9d Windows/msvc - Release build
-----------------------------------------------

No CUDA driver found
Unknown OpenCL driver version! Hashrate and stale shares may suffer
OpenCL platform: OpenCL 2.1 AMD-APP (3354.13)
Available GPUs for mining:
GPU1: MSI AMD Radeon RX 6600 XT (pcie 9), OpenCL 2.0, 8 GB VRAM, 32 CUs
Eth: the pool list contains 1 pool (1 from command-line)
Eth: primary pool: etchash.unmineable.com:3333
Starting GPU mining
GPU1: AMD driver 22.1.1
Eth: Connecting to ethash pool etchash.unmineable.com:3333 (proto: EthProxy)
GPU1: 48C 0% 3W
GPUs power: 3.0 W
Eth: Connected to ethash pool etchash.unmineable.com:3333 (165.227.119.242)
Listening for CDM remote manager at port 60080 in read-only mode
Eth: New job #622207c1 from etchash.unmineable.com:3333; diff: 4000MH
GPU1: Starting up... (0)
GPU1: Generating etchash light cache for epoch #239
Light cache generated in 1.4 s (34.0 MB/s)
GPU1: Free VRAM: 7.922 GB; used: 0.063 GB
GPU1: Allocating DAG (2.88 GB); good for epoch up to #241
GPU1: Generating DAG for epoch #239
GPU1: DAG 24%
Eth speed: 0.000 MH/s, shares: 0/0/0, time: 0:00
GPU1: DAG 48%
GPU1: DAG 72%
GPU1: DAG 96%
GPU1: DAG generated in 6.4 s (457.8 MB/s)
GPU1: Using Ethash OCL kernels (gfx1032; -clkernel 1 -clnew 0)
GPU1: no -gt value specified, switching to auto-tune
GPU1: starting auto-tune process
GPU1: clSetKernelArg (-48)
Fatal error detected. Restarting.
Eth speed: 0.000 MH/s, shares: 0/0/0, time: 0:00

Phoenix Miner 5.9d Windows/msvc - Release build
-----------------------------------------------

Waiting 15 s for previous instance to close

-------------->MINER LOG END<--------------

PhoenixMiner doesn't support anything newer than 21.11.3. I don't know if the drivers are so significantly different to cause this problem every three months, or if PM is just coded to look for specific driver versions and won't even TRY to use the AMD path with newer drivers than what it specifically looks for. It's annoying, that's for sure.

An easy way to TEST would be to:
1) Create a Symbolic link/Junction link to the latest (Stable AMD supported PM drivers) but RENAME the folder to previous drive yet in the system ENV_VARS %path% ../../../%Path_to_Symlink% <== OR PATH_POINT_TO >= build 21.12.x (ABOVE) stable 21.11.3 drivers.
2) I suspect WMI maybe queried by PM and this is how driver version is derived.
3) Ask the dev @PhoenixMiner what the checks are AND if its hardcoded in the code to search for driver version or is it because PM fails because of an incorrect dependency contained in drivers grt > v21.11.3 which PM relies on.. OR missing *.dll it could be a million possible things.
4) It could be that AMD drivers compiled past version 21.11.3have been totally refactored and rebased and what PM relies upon is no longer a resource that is in the codebase any longer, thus fail.
5) It could also be a part of PM's dev workflow to test against a set of Drivers to keep development under control by stipulating a range of supported drivers.
..Buts thats what I would do to try and isolate reasons for fail on latest driver above PM recommended.