We got a workstation pc with i9 13900k and ubuntu 22.10 is installed.
When all cores at their max, system becomes unresponsive. Rejects all ssh connections and keyboard and mouse does not work.
Tried to troubleshoot it but wasnt able to as we cannot get any information while system is at full load.
What would be the reasons? How we should approach the problem?
processes at hand: 31 pybullet simulations.
I'd be inclined to monitor the system while ramping up the workload. So rather than hit 31 pybullet simulations, try building up to that. Before you start, I'd install dstat and leave it running in a terminal. Start the workload and watch the output of dstat.
sudo apt install dstat
Run in a terminal:
dstat --time --load -cdngy --top-cpu --top-mem --top-io
It'll look a bit like this:
----system---- ---load-avg--- --total-cpu-usage-- -dsk/total- -net/total- ---paging-- ---system-- -most-expensive- --most-expensive- ----most-expensive----
time | 1m 5m 15m |usr sys idl wai stl| read writ| recv send| in out | int csw | cpu process | memory process | i/o process
05-05 11:43:10|1.99 2.02 2.17| 10 2 87 1 0|2106k 2208k| 0 0 |7742B 24k|8606 13k|msedge 2.8|msedge 920M|msedge 11M 170k
05-05 11:43:11|1.99 2.02 2.17| 11 2 87 0 0| 0 896k|1174B 1117B| 0 0 |8305 10k|msedge 6.3|msedge 918M|msedge 29M 63B
05-05 11:43:12|1.83 1.99 2.16| 6 2 91 1 0| 0 1168k|2098B 2197B| 0 0 |6219 8362 |msedge 4.4|msedge 918M|msedge 20M 54B
05-05 11:43:13|1.83 1.99 2.16| 7 3 90 0 0| 0 0 |1524B 2136B| 0 0 |8403 12k|msedge 2.1|msedge 917M|msedge 3874k 5092B
05-05 11:43:14|1.83 1.99 2.16| 5 2 92 1 0| 0 0 |2200B 554B| 0 0 |6227 8621 |msedge 2.2|msedge 918M|btm 189k 6355B
05-05 11:43:15|1.83 1.99 2.16| 9 2 89 0 0| 0 8192B| 248k 5862B| 0 0 |7933 11k|msedge 4.8|msedge 919M|msedge 19M 260k
05-05 11:43:16|1.83 1.99 2.16| 9 2 88 1 0| 0 1488k|1194B 997B| 0 0 |6372 8858 |msedge 6.3|msedge 918M|msedge 28M 204B
05-05 11:43:17|2.00 2.02 2.17| 9 2 89 0 0| 0 3368k|2385B 447B| 0 0 |5557 7323 |msedge 6.2|msedge 918M|msedge 29M 49B
05-05 11:43:18|2.00 2.02 2.17| 9 2 88 1 0| 0 80k| 999B 351B| 0 0 |8337 12k|msedge 6.2|msedge 918M|msedge 27M 7521
You can optionally write it to a file in the background, in case the GUI locks up:
dstat --time --load -cdngy --top-cpu --top-mem --top-io --output report.csv 1 5 &
Start building up a workload, and note the time when you do it. You'll see the dstat output has timestamps so you can correlate what's happening. You'll likely see the cpu usage rise, and load average.
