Hi!
Happy to use this miner it's work perfectly and I'm using it with 5 rigs of 12 GPU each (RX580,GTX1070/TI/1080TI ...) since month now, and I'm an early adopter of PhoenixMiner.
Everything is fine until last upgrade (2.7c) where I meet 2 strong problems appearing on 2 of my RIGS (and not the others):
- first a crash of the miner show CUDART errors (never met before)
- then the software is trying to launch a new session BUT keeping the previous one alive, showing the GPU with 0.000 MH/S and many errors. Then server is KO.
First errors appearing:
--> random "CUDART error in CudaProgram.cu:434 : an illegal memory access was encountered (77)" crashing my miner.
Then after the 1st error above when miner is trying to launch itself through a new session BUT the previous one stay present under W10 so GPU are having 0.000 MH/S ...
I'm on W10PRO, latest drivers, etc... everything uptodate.
Of course I tried to lower OC and so on, without success.
What's in common with those 2 RIGS: having RX580 NITRO+ with HYNIX memory (1st rig with 12 of those cards, the others with 6 of them + 6 GTX).
Any idea how to solve please? Thanks Phoenixminer devs and other members!::
Maybe try -mi 8 or -mi 9 take intensity down a bit ? im no expert but it helped me with cudart