Open Mini-Digger opened 6 days ago
Hello Mini-Digger, Can you provide us with more information about your environment and the detailed error message? The support of WSL may not be optimal since we have not tested it on such devices. If you can run it on one GPU device, you can try running multiple instances with distinct device ids to mine
2024-10-11T13:55:42.757919 INFO Prover start. version(0.2.0) commit hash(ed3b854)
Failed to get count of cuda capable device. Exception:out of memory
thread 'main' panicked at src/main.rs:152:14:
Failed to get gpu number: Failed to get count of cuda capable device.
note: run with RUST_BACKTRACE=1
environment variable to display a backtrace
WIN10,WSL2,Ubuntu22.04
It is not possible to run multiple instances with a single GPU, as each instance detects multiple GPUs, resulting in the above error message. Disable all GPUs in the Device Manager except for the individual GPU to be used before enabling instances of a single GPU. In short, the kernel encountered an error while detecting the initialization of multiple GPUs.
The kernel does not support multiple GPUs well, and running multiple GPUs under WSL may result in errors. Can it be compatible with WSL?