1 serverperformance
mckaygerhard edited this page 2023-08-12 20:02:56 +00:00

testing on empty world: just engine 5.5.0 + mtg 5.5.0 run several times, some time, default options/settings generated just standard area, mean near 125 Mapblocks (555), not moving gamer.

base is minetest game as 100 %

test with just one other mod, to see their impact On low part world, NO gamer in: % of all used CPU time / rising load in % to basic config (100/(100 - tested mod %)) = + load

  • 93.7 - 93.8 % / + 1.487 % load : advtrains ... advtrains via github.com (3cb21a8/10.02.2022) tested ...
  • 84.3 % / + 536 % load : poshud
  • 75.6 % (18.4 + 57.2%) / + 310 % load : technic ... via technic ... plus basic_materials, moreores and pipeworks
  • 69.1 % / + 223 % load : petz via github.com/runsy/petz 4688047 / 10.03.2022 - tested
  • 67.1 % / + 204 % load : pipeworks via ... github.com/mt-mods/pipeworks with basic_materials - tested
  • 64.7 % / + 183 % load : Witt-2
  • 61.9 % : technic + pipeworks with Festus1965 both mods tuning options (no mt.conf)
  • 51.9 % / + ~109 % load : poshud
  • 51.8 % / + 107.5 % load : character_anim
  • 51.1 % / + 104.5 % load : sprint
  • 48.7 % / + 95 % : animalia 13.8 % + with needed Creatura 34.9 %
  • 40% : animalworld (know from PAW server)
  • 30.4 % (32.3 %) / + 43.7 % load : techage via github.com/joe7575/techage 8b0d340 20 days ago (? 22.03.2022)
  • 30.4 % / + 42.9 % load : wielded_light, new tested, 12h
  • 28.9 % / + 38 % load : hbsprint
  • 21.0 % / + 26.6 % load : modlib as a dependency alone (doing nothing)
  • 20.7 % / + 26.1 % load : playeranim
  • 18.5 % / + 22.6 % load : unified_inventory
  • 18.4 % technic beside pipeworks/ + 22.5 % load only extra: but ! as pipeorks is needed for technic = 75.6 %
  • 3 % / + 3.3 % load : mobs_redo plus ... animals, monster, sky water, horse (30 diff - as petz 48?)
  • 0.1 % / + ? % load : airutils (d1e5ec2/03.05.2022), as base to supercub
  • 0.1 % / + ? % load : supercub ... supercup, as depend on airutils
  • 0 % / + 0 % load : terumet (never played myself)
  • 0 % / + 0 % load : mobkit (as needed depend mod) tested here
  • 0 % / + 0 % load : biofuel (dd8a199/21.02.2022/V0.7), as base to supercub
  • 0 % / + 0 % load : what_is_this_owo

new coming in:

  • ???

So as a result over 'theme'

technic:

  • first choice is terumet as 0 impact on empty world, or let say ' no overhead'
  • second is techage (+43.7 % load) as include tubesystem
  • not recommended : technic, with need of pipeworks (together + 310 % load) not recommended as of pipeworks AND technic refuses to rethink their lag impact

mobs :

  • first choice is mobs_redo with belonging mobs with just + 3.3% load,
  • before next animalia (+needed creatura) + 95% load (double lag)
  • not recomended: petz with raising 223% load (tripple lag) ... maybe for fast and strong servers only

cars / airplanes ?

as some base on mobkit as mobs, interesting

  • supercup as even with 2 depending mods, as a sum of only 0.1% CPU time on blank MTG

sprinting

  • hbsprint : 28.9 % + 38 % load, on only mtg base server (= lower impact, better choice based on lag/cpu)
  • sprint : 51.1 % / + 104.5 % load

node info

hud top show what you look at, name ...

  • what_is_this_owo ... 0 % empty server impact
  • witt ... 64.7 % / + 183 % load

So how to calculate your servers expected impact : Take your server, only allow mtg and let it run for 24 hours. check /status maybe 4-6 times short logged it, take a profiler save extract watch your CPU avg usage ... via top, htop or whatever.

So if you had an average of 10% all CPU was working, add/multiply with the % load shown: mean, with your server mtg and maybe pipeworks your avg CPU usage is then about ~30 %, and the measured lag tripled also.

You don't believe ? Just go this kind of server sorting, see the high lag server and look what mods are running. Got it ?

IF you want to have a low lag server, with option for much gamer having fun ? Read and learn how to keep lag low - with choose the right low lag mods

games on a Intel E1, 1400 GHz, 2 Threads, 4 GB RAM reference is same time running htop time used on CPU: (from bad high er impact sorted down to lower impact compare to same time running htop and it's usage)

  • factor 2.75, or 175 % more / between 0.7-1.6 CPU / 3.9 % RAM : mineclone5 (361314e / Dec 10, 2021)
  • factor 1.71, or 71 % more / between 0.7-2.0 % CPU / 3.9 % RAM: minetest_game (5.5.0 cc3e7be/31.01.2022)