waifu2x-converter-cpp ... -j ?

[idanka]

My question would be, what is the optimal value of "-j"? waifu2x-converter-cpp ... -j xy?

i have i5-4570 (4 Core, 4 Thread) and nvidia 1650, 16 GB RAM I use Ubuntu Linux 20.04 What to consider?

conversion is currently running (waifu2x-converter-cpp --scale-ratio 2.667 --noise-level 2 -m noise-scale -i /img -o /img2x --model-dir .waifu2x --processor 0 -z -j 4)

• GPU max73%
• GPU VRAM 442 MB
• CPU all 10-20%
• RAM 4.41GB

Thx!

[idanka]

Lol:) test 31 image (--processor 0) gpu or cuda used, not authoritative

-- -  Done, took: 2m 26.313s total, ETA: 0.000s, file: 4.714s avg: 4.720s
-j 4  Done, took: 2m 26.342s total, ETA: 0.000s, file: 4.738s avg: 4.721s
-j 8  Done, took: 2m 26.492s total, ETA: 0.000s, file: 4.715s avg: 4.726s
-j 16 Done, took: 2m 26.583s total, ETA: 0.000s, file: 4.732s avg: 4.728s
-j 64 Done, took: 2m 26.528s total, ETA: 0.000s, file: 4.729s avg: 4.727s