Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  

Author Topic: Fort mode performance testing: Sandy Bridge CPU  (Read 1322 times)

merino

  • Escaped Lunatic
    • View Profile
Fort mode performance testing: Sandy Bridge CPU
« on: January 27, 2011, 08:24:35 pm »

I was overdue for an upgrade so I jumped on Intel's new Sandy Bridge processors when they came out. An RMA, a few additional orders, and a few weeks later, and I'm finally ready to do some testing. Setup details in the spoilers.

System setup:

Spoiler (click to show/hide)

Dwarf Fortress setup:

Spoiler (click to show/hide)

Unfortunately I haven't built any forts lately, so I grabbed a few succession game saves to test with a more realistic load. Namely, the latest saves (as linked to on DFFD from the OP of each thread) of Skyscrapes and Ardentdikes. Between the two tests, I reverted back to the downloaded copy of the save file.

Ardentdikes:
Dwarfpop: 45
Animalpop: 30
Fort status: Looks like a 3x3 map, lots of work going on, only 1-2 idlers at once, water flowing as well as lots of lava flowing, steam, and items being burned/temperature transfer, plus rain
Actions taken: none, there was already plenty going on!
Test 1 FPS range: 39-47 (mostly ~43)
Test 2 (weather/temp off) FPS range: 40-100 (almost always 70-80) (given that it was already raining, turning weather off won't have really helped AFAIK)

Skyscrapes:
Dwarfpop: 211
Animalpop: 35
Fort status: Looks like a 2x2 map but with lots of height. Not incredibly deep, but with caverns and the magma sea revealed. A surface river with a dam, so it doesn't appear there's any liquid flows in effect, though I'm not 100% sure of that. 40-60 idlers most of the time.
Actions taken: none again.
Test 1 FPS range: 11-12
Test 2 (weather/temp off) FPS range: 13-15

Ardentdikes was a lot more all over the place after dumping temperature, not too surprising considering the flowing lava. Skyscrapes seems to be more limited by population and size. I didn't get a chance to test on my previous setup so it's hard to react to these without having a basis for comparison. It seems to be doing fairly well, it not as great as I'd liked, but Skyscrapes in particular may just be a much more intensive fort that I'm used to running.
Logged

o_O[WTFace]

  • Bay Watcher
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #1 on: January 27, 2011, 09:26:16 pm »

Try running DF cleanup, march the entire population through a dorfwash and then test the FPS.  Alot of people think thats the real FPS killer, but no ones really sure. 
Logged
...likes Dwarf Fortresses for their terrifying features...

Voyager

  • Bay Watcher
  • Maybe it's supposed to be on fire.
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #2 on: January 27, 2011, 11:29:38 pm »

What is a Dorfwash?
Logged

Girlinhat

  • Bay Watcher
  • [PREFSTRING:large ears]
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #3 on: January 27, 2011, 11:34:36 pm »

Anything that puts water on dorfs.  I'm working on a shower chamber, with floor grates and a vat of water that falls on the dwarves, washing the blood off them like some large delousing chamber.  Others use a small trench in front of the dining hall (or other area) with 2/7 water to cleanse their dirty masses.

krenshala

  • Bay Watcher
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #4 on: January 28, 2011, 12:06:14 am »

Anything that puts water on dorfs.  I'm working on a shower chamber, with floor grates and a vat of water that falls on the dwarves, washing the blood off them like some large delousing chamber.  Others use a small trench in front of the dining hall (or other area) with 2/7 water to cleanse their dirty masses.
If you use the trench, don't forget to "change" the water every now and then, or it will eventually start making the dorfs walking through it dirty instead of clean. ;)
Logged
Quote from: Haspen
Quote from: phoenixuk
Zepave Dawnhogs the Butterfly of Vales the Marsh Titan ... was taken out by a single novice axedwarf and his pet war kitten. Long Live Domas Etasastesh Adilloram, slayer of the snow butterfly!
Doesn't quite have the ring of heroics to it...
Mother: "...and after the evil snow butterfly was defeated, Domas and his kitten lived happily ever after!"
Kids: "Yaaaay!"

Di

  • Bay Watcher
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #5 on: January 28, 2011, 06:31:31 am »

or it will eventually start making the dorfs walking through it dirty instead of clean. ;)
It won't, they always get water covering only. Though they get exposed to syndrome if its carrier appears inside trench. They can clean them by themselves but that depends on how intensive traffic is. Trenches near my hospital get a cleaning in in next month after contamination, those at the entrance were cleaned once or twice in whole lifetime.

to merino:
Could you see if dfclean helps fps? Ardentdikes have lakes full of FB secretions in one of caves. Do it, for !!science!!
Logged
Quote from: Creamcorn
Dwarf Fortress: Where you meet the limit of your imagination, moral compass, sanity and CPU processor.
http://www.bay12forums.com/smf/index.php?topic=103080.0 Fix sober vampires!
http://www.bay12forums.com/smf/index.php?topic=91442.0 Dwarven Cognitive Science

FaultyLogic

  • Bay Watcher
  • [EDIBLE_RAW]
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #6 on: January 28, 2011, 09:34:13 am »

Isn't that a really low FPS for that kind of computer power? I get ~40 fps with 90 dwarves, +river, +magmasea on my three-year old laptop.
Logged
"It's unpossible!"

helf

  • Bay Watcher
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #7 on: January 28, 2011, 10:36:59 am »

What OS are you running on it, are you using the on die GPU, what are the ram timings?

That does seem kinda weak. My e8400 core2 duo @ 3.74ghz (darn chipset wont go higher, I still havent changed cpu voltages!) with cas7 ddr3, winxp has higher average frame rates, iirc. I'll have to check with those saves, but I'm fairly positive I average a good bit higher than that.
Logged
YOUR GAMES GLITCH: Hey, I got out of the map boundry!
OUR GAMES GLITCH: Hey, a horrid monstrosity just migrated to my fortress! Let's recruit it!

o_O[WTFace]

  • Bay Watcher
    • View Profile
Re: Fort mode performance testing: Sandy Bridge CPU
« Reply #8 on: January 28, 2011, 10:49:45 am »

If you use the trench, don't forget to "change" the water every now and then, or it will eventually start making the dorfs walking through it dirty instead of clean. ;)

Thats what dfcleanup is for.  Run it and it removes all contaminants on the map.  A dorfwash removes all contaminants on dorfs and puts them on the map.  Together they remove the liquid Fun that tends to coat everything. 
Logged
...likes Dwarf Fortresses for their terrifying features...