Post Reply 
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
[2.0] Negative Performance Changes on Low-end Machines
2017-07-26, 10:50 PM
Post: #1
[2.0] Negative Performance Changes on Low-end Machines
Disclaimer : This is not a proper "bug" report but more of a "general kind of issue" report. (Bug report section is the one that seems the most fitting place for this)

P.S.: Before I start, I'll stop ppls as stupid as the ones on discord, stop saying that "it's due to new shadows" because everything is done at the lowest resolution (resolution and graphics) and the only version of the game which do have shadows is the highest one (btw I can't even play the game with it as it basically insta-crash before even the game's starting loading screen).
It however can be "UV mapping" and "bump mapping" which is the name of the "shadows" that has always been in this game. (aren't actualy shadows)


So, what I found out is that for this version of the game, is that performances on low-end machines were hammered.
I recorded my tests (only FtD ones, as I'm not here to advertise other games), but the record file got missed into oblivion thx to no reasons, I would probably redo it sometimes.

So, the platform I've performed the test on is my 3 year-old low-end laptop, it got a 4 core/4 thread Pentium CPU @2.6 GHz and it has no GPU (it's an APU with similar performances to Intel HD Graphics 3000) with 4 Gb RAM and the game on a SSD.

I will compare performances between 1.968 (best mod compatible version with equal performances as 1.969 (which isn't as mod compatible)) and 2.0 (apparently not the lastest version of it, but the one that pirates share)
Both version are pirates (because I can't buy the game yet, [and considering 2.0's result if not fixed, I mostly likely won't even buy it when I will actuay be able to], and also because it actualy make sure that data is truely the one of the said version)
I'll start of with a comparaison of the FtD version, and give a comparaison with other games I run on this PC (just so you can size the performances of this PC and maybe compare it to yours)

I tested the game running alone with the Task manager (to see real-time performances)(I redid it without running the recorder because that could influ the results)(no sub-apps running either exept lightshot to take the screenshots)

I) Settings :
1.968 =
[Image: zbk9J3ODQ1asTafAUukVQw.png]
Vanilla (no mods) and no external content exept my profile and everything related to it

2.0 =
[Image: VGPUIFRJR2WkCal_pZhqXA.png]
Same (Vanilla (no mods) and no external content exept my profile and everything related to it)

II) Loading time of the Starting load :
1.968 = 62 seconds.

2.0 = 72 seconds.

III) FPS in Main menu (with standard Neter crafts battle in background) :
1.968 = ~16.66... FPS on average with up pikes at 21 FPS and down pikes at 12 FPS. (I did rounded the pikes)

2.0 = ~11.4 FPS on average with up pikes at ~16 FPS and down pikes at ~6 FPS. (I did rounded the pikes too)

IV)FPS on "Vanilla Designer" with the classic "raft.blueprint" loaded in :
1.968 = ~18.8 FPS on average rather stable.

2.0 = ~15.9 FPS on average very stable however, a very annoying kind of slow-motion on the mouse. (can't realy explain it with otherwords exept "my cursor was going super slowly")

V)FPS on "Vanilla Designer" with the mighty SS Godly craft"Thyr.blueprint" (that is known by my PC to be the more CPU/Memory intensive craft) loaded in :
1.968 = ~7.4 FPS on average which is unstable with best result being at under 9 FPS.

2.0 = ~5.9 FPS on average more stable than the previous version but best results under 7 FPS and still that very annoying kind of slow-motion on the mouse.

VI) Building a small ship (59 blocks) while only using 1x1x1 blocks in an empty "vanilla designer" (to show impact of "slow-mo mouse" with mouse based building (do not, it's slow-mo in building mode as in normal mode as well)) :
1.968 = 33 seconds.

2.0 = 37 seconds.

(the shown result is the shortest of 3 tries (in a row) and this is for a very small craft, the bigger the craft the more horrible it becomes)

In general :
1.968 = The CPU got all it's cores at 80~100% all the time exept while loading (80~85%) and exept on main menu (which is continiously being at 100%), it don't use the disk at all, exept when 1st loading and when loading a craft. The memory took is about ~1.2 Gb RAM

2.0 = The CPU got all it's cores at 90~100% all the time, did used 4~17% of the disk continously, even when not doing anything, does use 0.2 Gb less RAM than the previous version on average

Thoughs : Somewhat, 2.0 uses less memory but more CPU intensive and use the disk continously for no reasons (alltho, I'm on SSD so I don't realy give a Lamb) and it does perform worse (at a point where it's basicaly unplayable)

Other Games Comparaison :
Game name (state) - graphical settings - ingame fps
Minecraft (vanilla) - 1280x720 standard settings - ~50 FPS
Civilization IV (vanilla) - 1366x768 high settings - ~35 FPS
Minecraft (110 mods) - 1280x720 minimal settings - ~27 FPS
WoT (basic XVM) - 1366x768 medium settings - ~20 FPS
Civilization V (vanilla) - 1366x768 low settings - ~10 FPS

(for the most famous ones, civ 5 is the "biggest" one I got, and I barely run it without melting the CPU xD I did took vanilla minecraft and modded minecraft to show mod impact on performances (do note, on FtD it's rather irrelevant as only graphical/algorithic calculations (and mods) do seems to impact its performances at least until 2.0 (and as I can't run it properly vanilla, I didn't tested mods yet)))


I tried to list as much data as I could with averages and pikes that I wrote down to try to help you figure out what's the origin.
Tbh I don't know if it's the UV mapping "fix" or if it's the passage from unity 4 to 5 or if it's something else, but do note that if I replace 1.968 (or 1.969)'s StreamingAssets folder and replace it by 2.0's there's no major performance changes, so it's not due to new textures nor new models.

King of the Idrillian Kingdom ; my BP Thread ; my To-Do on Trello.

[Image: h6npcjh]
Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump:


User(s) browsing this thread: 1 Guest(s)