Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[SabbyKat]

I just thought I'd slap this info down for !SCIENCE! Purposes, and detail my on-going FPS experiment and the issues there of.

Below is a list of FPS drop 'types', and the results from my personal testing so far! This is VERY fresh in experimentation, I plan to update it as I have more information. I will also begin to attach forts to let others test on their rigs and see if they gain similar results or not to help cross-confirm exact !SCIENCE!.


~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~

Total Items (MASSIVE)

~ in my 10 year old fort, I had gathered 50,000 or so items. Prior to my deletion of ALL items (not furniture, workshop, etc) – I had 139-144 FPS. After deleting ALL items, I jumped to 480-495 fps! This is a truly massive thing to focus on

Total sentient Entities (Invaders and Dwarves) (Major)

~ As is well and likely known, the amount of dwarves has a fairly major impact on FPS. It's hard to put a science  on their specific fps lose due to the constant changing path finding, work flow, degrading clothing, and all the FUN of that. Further, Invaders act like a dwarf on the map in fps lose. It's normal to see a siege of 80 drop my fps by over 100, especially if they have to run through a 'trap hall' to get to me (more on that below).

~ I am too early in testing to suggest any improvements to this that isn't obviously, less invaders, less dwarves! But I can say a fort of 35 dwarves is completely self-sufficient, unless you NEED a 10-20 man military. though, I may question why, especially when I show you 2 of my dwarves solo a 110+ frost giant invasion, in steel armor... did I mention all invaders have +15 skill levels? >.>


Stock Piles (Moderate)

~ using default stock piles, storing all stock piles in a 3x3 embark floor (Cleared all stone/etc, and used it for stockpiles) – I noticed a 35-43 FPS gain by REMOVING the stockpiles. Only about 1/3rd of the stockpiles had items or bins.



'Complex Halls' (Moderate)

~ The good ol' classical 'trap hallway of zig-zagging doom!' - I've noted such zig-zagging and turns, ESPECIALLY when invaders enter causes a MASSIVE fps lose. I mean, 100+ fps from a siege of 80.

~ Further, even a '3 turn' trap hall minimizing the turns and complexity, when I 'locked off access' to it, gave me a 30 FPS boost. Nothing complex – it was a 5 wide hall, from left side to right, up 5 squares, then left to the next map edge for only half a map Z level. Extremely simple, yet showed a huge lose of fps. Even when not part of a barrow (denying Dwarves access), and 'RESTRICTED' traffic, neither improved fps. Only sealing it physically did.



Traffic Designations (Confusing?)

~ I am... puzzled. Traffic designs should improve fps, yet... my experiments showed fps LOSE in almost ALL scenarios I used them! Leaving areas not needed for travel as low, 'empty areas' that are sealed off, as restricted, etc – I saw no gain. Removing all traffic – I gained about 15-20 FPS! No matter what combo I used and waiting 5 minutes to let 'things adjust' – I would get an fps LOSE for traffic controls?

~ If others have insight, it would be very handy as I could not discern a way to cause a fps gain by this?

Visible Areas (underground - Minor)

~ Total 'dug up and accessible' area is a moderate concern for fps. When I replaced the HAPPY FUN DEPTHS, Magma sea, caverns, and my entire fort up to to the surface with stone (all open area=walls), I went from 530 fps to about 690 fps. Less so you should be worrying about this, as it seems even quarantining off  entire dug out Z levels (walling physical access to the area), added maybe 1-2 fps? I need to test more. This may be different based on more complex forts, where I use a centralized stairway to ensure equal 'access and exit' to all floors.

Visible areas (Channeled surface - Game Breaking!)

~ I do not know why, but if you channel out the entire surface from edge to edge, you'll notice a ABSOLUTE INSANE fps lose. I am talking losing 500+ fps from digging out the surface, down 1 layer. Access to the floor below is there, nothing is 'floating above' – a clean, fresh map and I lose 500+ fps doing this. Want a simple really, really deep 3x3 moat around the map? Good luck. At 30+ depth I went under 50 fps!

~ I am completely uncertain what causes this, and would others can tell me if this happens to them or not too, so we can theorize what the heck is going on!

~*~*~*~*~*~*~*~*~*~**~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*

Things I need to test still:

~ Soil/Gem/Mineral 'types' and their impact on fps. Simple soil, etc, if it matters.

~ Amount of plants/wood and their impact both in species, and on embark (1500 trees vs 150)

~ World generation, civilization count, world size, history 'issues', etc etc. Extremely complex thing to test and one I am avoiding for now.

~ Specific biome 'quirks' and their fps. I have had embarks identical on the same world, have drastically different fps based on biome.

~ More to come!


As I said, this is a VERY beta draft of my experiments. I will update it in time, I just thought I'd share specifics to MASTERWORK and not 2-10 year old info on google for vanilla!

[LMeire]

Traffic designations are more for keeping dwarves away from specific obstacles than just willy-nilly-everywhere. They can help if you're having issues with haulers running through crowded pastures, frequently-used minecart tracks, and the bottoms of fast-flowing waterfalls- but only then because when they get interrupted they have to recalculate their paths again.

[Putnam]

This probably belongs in the DF general board instead of here.

[SabbyKat]

This probably belongs in the DF general board instead of here.

Actually, no it does not. I am doing tests in MASTERWORK for fps about MASTERWORK (not to sound rude). Vanilla and Masterwork have entirely different systems, dfhack, all that fun that has given be drastically different results from vanilla vs Masterwork in fps matters.

if Meph disagrees, he can tell me himself and I will remove it. It's truly no issue to me, as I am doing this for my own benefit after hitting various, frustrating fps issues without logical cause.

[Putnam]

All of the causes you described for FPS death are exactly as valid for vanilla DF. The research is very useful and very good and applies to all versions of Dwarf Fortress. I would prefer it be more visible to everyone rather than just people who play Masterwork.