I have launched a cluster with 3 nodes on AWS, and when I issue sky exec, I see that the resource section shows 2x [CPU:0.5]. Does this mean it's only using 0.5 CPUs? How can I specify the amount of CPU and memory to use for each job/exec? I tried using --cpus parameter directly in the command but it was ignored. Zongheng mentioned setting OMP_NUM_THREADS, but I'm not sure how to do that. Can you provide guidance on how to specify CPU and memory usage for each job/exec in a cluster?