Jittor is using 16 procs here (see logs below). nproc outputted 30. How to use all the 30 procs?
[i 0914 16:57:25.240278 12 compiler.py:956] Jittor(1.3.8.5) src: /home/ubuntu/anaconda3/envs/test_env/lib/python3.9/site-packages/jittor
[i 0914 16:57:25.242855 12 compiler.py:957] g++ at /usr/bin/g++(9.4.0)
[i 0914 16:57:25.242917 12 compiler.py:958] cache_path: /home/ubuntu/.cache/jittor/jt1.3.8/g++9.4.0/py3.9.17/Linux-5.15.0-6xbb/IntelRXeonRPlax9b/default
[i 0914 16:57:25.258817 12 install_cuda.py:93] cuda_driver_version: [12, 0]
[i 0914 16:57:25.261936 12 **init**.py:411] Found /home/ubuntu/.cache/jittor/jtcuda/cuda11.2_cudnn8_linux/bin/nvcc(11.2.152) at /home/ubuntu/.cache/jittor/jtcuda/cuda11.2_cudnn8_linux/bin/nvcc.
[i 0914 16:57:25.306967 12 **init**.py:411] Found gdb(20.04.1) at /usr/bin/gdb.
[i 0914 16:57:25.309797 12 **init**.py:411] Found addr2line(2.34) at /usr/bin/addr2line.
[i 0914 16:57:25.402290 12 compiler.py:1011] cuda key:cu11.2.152_sm_86
[i 0914 16:57:25.679032 12 **init**.py:227] Total mem: 196.58GB, using 16 procs for compiling.
[i 0914 16:57:25.762435 12 jit_compiler.cc:28] Load cc_path: /usr/bin/g++
[i 0914 16:57:25.832214 12 init.cc:62] Found cuda archs: [86,]
[i 0914 16:57:25.909511 12 **init**.py:411] Found mpicc(4.0.3) at /usr/bin/mpicc.
[i 0914 16:57:27.461097 12 cuda_flags.cc:49] CUDA enabled.