What is the optimal configure.json setting for mining Random XL using xmrig on Loki?

Question

I'm mining on a threadripper 1950x and I followed this guide: https://medium.com/luxor/loki-randomxl-fork-af9b141e492a

I got to the point of mining, but now I want to optimize it further for mining Loki Random XL. Currently it only shows mining with 2 threads at 800h/s. I'm not sure what to change to make use of more threads in my cpu. What are the optimal settings?

  {
    "api": {
        "id": null,
        "worker-id": null
    },
    "autosave": true,
    "background": false,
    "colors": true,
    "cpu": {
        "enabled": true,
        "huge-pages": true,
        "hw-aes": null,
        "priority": null,
        "asm": true,
        "cn": [
            -1,
            -1
        ],
        "cn-heavy": [
            -1
        ],
        "cn-lite": [
            -1,
            -1,
            -1
        ],
        "cn-pico": [
            -1,
            -1,
            -1,
            -1
        ],
        "cn/gpu": [
            -1,
            -1,
            -1,
            -1
        ],
        "rx": [
            -1,
            -1
        ],
        "rx/wow": [
            -1,
            -1,
            -1
        ],
        "cn/0": false,
        "cn-lite/0": false
    },
    "donate-level": 1,
    "donate-over-proxy": 1,
    "http": {
        "enabled": false,
        "host": "127.0.0.1",
        "port": 0,
        "access-token": null,
        "restricted": true
    },
    "log-file": null,
    "pools": [
        {
            "algo": "rx/loki",
            "url": "loki-us.luxor.tech:9999",
            "user": "my_address_here",
            "pass": "x",
            "rig-id": null,
            "nicehash": false,
            "keepalive": false,
            "enabled": true,
            "tls": false,
            "tls-fingerprint": null,
            "daemon": false
        }
    ],
    "print-time": 60,
    "retries": 5,
    "retry-pause": 5,
    "syslog": false,
    "user-agent": null,
    "watch": true
}

Answer 1

Your fallback rx object only specifies 2 threads with no affinity. Add the following into the cpu object to yield 16 threads with no affinity:

"rx/loki": [
       -1,-1,-1,-1,-1,-1,-1,-1,
       -1,-1,-1,-1,-1,-1,-1,-1
    ],

Replace the -1's with CPU IDs to affine the threads to specific CPUs.

Documentation here.