Slow FPS

[madbammer]

Hey there i just bought this game and i really want to play it, i have a core i7 740, Geforce GTX 260M, that should run this game but im having 01-13 FPS, now with my temper i dont want to play but ive messed around and turned things off, reinstalled, googled it  changed the dust rate etc NOTHING WORKS! i do want to play this game really bad, considering Ive always found space fascinating and what a better concept for a game then this!? hahaha i will be checkin for updates and what not and hope those can help solve my issues if not i wait till this game is premo and play it then......sigh

[LimitlessAeon]

Check to see what your "dust particle" level is in the options tab (it's at the very bottom of the list). You shouldn't have it above 1.0x unless you have a really powerful CPU. See if it helps!

[Dan Dixon]

I posted a new update to Steam that should alert you if you've set the dust multiplier too high or too low.

Let us know if this solved your problem or if you still need help.

[Carnage2K4]

The FPS seems to be a strange issue, I just ran a sim with 4-5 plants with many rings in a binery system (all fell apart btw ) but even with dust off I was getting 7 FPS...
the game/sim seems to not be able to utilise much processing power, I had 1 core running about 50% and the other 3 basically idle (30% overall used) and I was still getting 7 FPS.

core i5-750 OC'd to about 3Ghz

sad as my comp should be able to destroy this sim.

[Dan Dixon]

There's lots of opportunities for multi-core optimizations that I'll be working on in the near future.

No dust, 5 objects, and only 7 FPS?

Trails with lots of segments can also slow a simulation way down. Try turning them off and let us know if that helps.

[Carnage2K4]

well my test to see how stuff is running is on open "our solar system" and place a few rings at 5000 count and see how many it takes to run the FPS down.

basic sim with no trails or dust = 60 fps 13% cpu (over 4 cores)
1 5000 count ring = 35 fps 28% (over 4 cores)
note that it never uses a single core over about 50% but the 28% is the total amount being used at any 1 time.
2 5000 count rings = 19 fps 28% (over 4 cores)
3 5000 count rings = 13 fps 28% (over 4 cores)

I just noticed that changing the affinity to a single core makes 1 core hit 100% yet total still sits at 28%.
so I guess it uses the same processing power for 4 cores that it would normally use for 1, and that has no benefit in the end that I can see.

[Dan Dixon]

Multi-processing is barley implemented in the current version and only applies to objects, not yet dust. I'll be releasing a free update that will enable multi-core support by default and make it work with dust in a future update.

[Joshimitsu91]

One way to boost FPS in complex systems is to select the option to auto-size your trails, so the sim automatically lowers the trail length if the FPS is starting to drop. Helped my performance on the Solar System All Moons simulation.

Josh

[ocbaker]

I also have an i7, I had around about 2,500 objects in my Simulation, and it was really slow (Trails auto, Dust off, calculations low) BUT my FPS was way less than 1.

I checked Task Manager to see if it was using my CPU, but it stuck on 13% an never changed. I tried upping the priority but no change.

Have I missed something, or have you purposefully set some kind of restriction on the game?

ALSO, you should be able to set that dust warning OFF, and I have to change to Accurate Calculations everytime I start up!

[Dan Dixon]

There's no intentional, artificial CPU restriction in Universe Sandbox. I'm not sure why the CPU percentage is so slow.

Give me some time to get multi-core support working correctly. Hopefully this will help a bit.

And yes... I'll be offering a way to turn off the dust warning in a future update.

[rei]

Is this related at all to this hiccupping I'm getting?

Every second or so the whole program hiccups for a split second, and then goes back up to an acceptable framerate.

I've tried setting the process priority high as well as closing other apps.

[Darvince]

That's probably an Antivirus continuously running checks on Universe Sandbox.

[rei]

Hmm nope, tried turning off Security Essentials, didn't make a difference.

I thought it might be the Gen 1 GCs since perfmon shows about roughly the same frequency of GCs as hiccups, but it's only showing a total of 1% CPU cost for all GCs -- so that can't be it.

[killroy]

hm thats strange i have 2.2 ghz and 512 ram and it works perfectly

[ChemicalBR0]

check to see what services you have running in memory.
there's stuff that loads at startup that simply doesn't need to be running all the time.
most notably the windows indexing service which continuously indexes file locations for quicker searching

some useful links
http://www.dq.winsila.com/tips-tricks/disable-windows-indexing-service-to-improve-performance.html
http://www.jasonn.com/turning_off_unnecessary_services_on_windows_xp

I did this a long time ago as i was getting stuttering in most of my games.
I don't see it at all nowadays.

when I run the solar system sim at default (1920x1080)... with trails and dust my average framerate is around 40-45fps
with just the dust removed, around 80fps
with dust and trails removed it jumps to around 390-450 fps. (thats a crazy performance hit just to show some trails and particles)

(phenom2 quad core. 4gb DDR3 ram. MSI GTX460 1Gb) by no means is it a super powerful computer but my framerates are acceptable. (for what it's doing it should be a LOT more but ho hum)

last night i added a model of the babylon 5 space station to a sim...
there were 2 stars 3 planets and the station (the station has 38000 polygons) so its quite high detail.
(1 of the gas giant planets had a 32 megapixel texture map)

the sim ran super smooth at over 200fps.
turned on trails with a length of 50 and it died a slow death dropping down to around 20fps
I think the code for trails and dust needs to be looked at.

[Puddin]

Just throwing out ideas here but one thing that may be to blame is that different CPU models crunch mathematical problems differently.  I read somewhere that this program is a n-body simulation which if my memory serves right is solved out by differential equations.

If I can find a benchmark program that  will work for this (other than matlab)I will repost.  Just something to think about though

[Dan Dixon]

I think the code for trails and dust needs to be looked at.

Yep. Drawing trails makes the simulation run very slowly. I'm working on a long term plan to fix these issues.

Hmm nope, tried turning off Security Essentials, didn't make a difference.

This is what I use.