100% CPU usage while loading datasets

[thecccut]

Hello @gab i have another request for a fix :).
When i start Japonicus and it loads the datasets it uses 100% of my CPU wich dosent make any sense and is a big problem for me because i just got a 32 core cpu and when it runs at 100% for a while it almost crashed my computer and could have damaged my cpu and motherboard ;(. Do you think its fixable?
I cant see any reason for japonicus to use 100% of the 32 cores when its just loading the datasets.
Hope you have a great new year ;D!

Regards Me lol

[Gab0]

Hey
I have never paid attention to this... but its true and strange XD
The thing is that at this point japonicus does not trigger any multicore code, so this should be a thing from gekko... and indeed we can confirm by running this code: (I guess you wouldn't wanna try)

import requests
requests.post("http://localhost:3000/api/scansets")

Japonicus does make multiple requests to the scansets api though, so I may improve this.

Anyway I think the usage of all cores would not explode the cpu... did you get proper cooling system?
This should be a beast running genetic algorthims, share results ;d

Happy new year too man, lets hope the apocalypse is postponed!

[thecccut]

Haha me to lol ;) will share results when i have some nice ones. the problem is when i have multiple datasets selected i was trying with 20 datasets it would run 13mins on 100% and 4 minutes in i was getting glitching on the screen. I got a cooler wich was recommended for my cpu but i think im gonna save up so i can buy parts for a custom water cooling system. Anyways lol i would love if you could improve it :).

[thecccut]

AHHH i think it could be because my History db is 10gb i tried with a lighter one and had no problems :D

[Gab0]

@thecccut Oh yeah, makes sense... gekko reads all the db files on this request... I'll change those methods to make only 1 request if possible (if its not already this way).

Btw is this cpu the Threadripper 2990WX? lool thats why I'm buying AMD stocks XD.
It should be important to get the watercooler before running on 32 threads though..

ok, if you have more thoughts just send in.

[thecccut]

Thank you for trying to fix the bug and that explanation makes sense since i get 50+ node windows in my task manager when loading my datasets with my big db file XD.

Yes it is indeed the 2990wx its a fucking beast especially if you also get some b-die ram and the x399 msi creation motherboard which dynamically overclocks it depending on the load its fucking amazing i would 100% recommend them but they are quite pricey :/.

And sure would be happy to report on more thoughts if i get some lul :).