Any way to restart the miner after GPU unresponsiveness?
2018-02-21 14:20:28|gpu_id 4 54 1 unspecified launch failure
2018-02-21 14:20:28|gpu 4 unresponsive - check overclocking
2018-02-21 14:20:28|cudaMemcpy 1 failed
Use this Miner Autorun (Watchdog)
https://bitcointalk.org/index.php?topic=2071108.0Thank you for that recommendation. That looks fine, but requires to use windows. I am running a linux rig... :-/ Any other idea?
You can use this little python script that I made to handle the miner:
#!/usr/bin/env python3
import subprocess
import sys
def restart_miner(proc):
print('Restarting miner...')
proc.kill()
def process_line(proc, line):
if 'cudaMemcpy 1 failed' in line:
restart_miner(proc)
else:
print(line, end='')
if __name__ == '__main__':
while True:
proc = subprocess.Popen(sys.argv[1:], stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
for line in proc.stdout:
process_line(proc, line.decode())
Copy/paste the code in a file, make it executable and run it with "./filename your_launch_command". With a file named "start.py" and for the default dstm's launch script the command would be "./start.py ./zm --cfg-file zm.cfg".
The script will restart the miner when it sees the "cudaMemcpy 1 failed" error message.
BTW, it should work on Windows too if you have python3 installed. Remove the first line and replace "sys.argv[1:]" by "['zm.exe', '--cfg-file', 'zm.cfg']" (or whatever your launch command is).