How I got the most speed out of spring-How I tripled my FPS

[Jazcash]

I'm always going to keep my graphics settings at the lowest quality simply because the fact I'm a performance > graphics person. I like to whore FPS as much as I can. I have the capability to use Lups but if it means me getting an extra 10+ FPS, no thanks.

Someone ring me when you've fixed the GML spam.

[ginekolog]

OMG


I just tripled my FPS with Behes advice. With normal single core spring i get 7 FSP with 1500 corvamp on merath4 patroling on 1 spot.

With MT build of spring i get 21 FPS with all figs on screen and 50 off screen. YEEEHA :


Thanx again Behe for this awsome idea.

Btw e8400 and nv9600GT here.

[Neddie]

I'm always going to keep my graphics settings at the lowest quality simply because the fact I'm a performance > graphics person. I like to whore FPS as much as I can. I have the capability to use Lups but if it means me getting an extra 10+ FPS, no thanks.

Someone ring me when you've fixed the GML spam.

Why would you whore FPS over, say, eighty? Around fifty/sixty most people can't see further improvement, and while I can see some difference through seventy it is highly unlikely that you can see any improvement beyond that.

[Jazcash]

I'm always going to keep my graphics settings at the lowest quality simply because the fact I'm a performance > graphics person. I like to whore FPS as much as I can. I have the capability to use Lups but if it means me getting an extra 10+ FPS, no thanks.

Someone ring me when you've fixed the GML spam.

Why would you whore FPS over, say, eighty? Around fifty/sixty most people can't see further improvement, and while I can see some difference through seventy it is highly unlikely that you can see any improvement beyond that.

Because I play SpeedMetal every now and then. And I'm not talking a few frames difference, I used to get about 40 FPS with the default graphic settings but after setting everything I can find to the lowest it can get without just displaying blobs of colours I'm now getting around 100-120 FPS.

[imbaczek]

Why would you whore FPS over, say, eighty? Around fifty/sixty most people can't see further improvement, and while I can see some difference through seventy it is highly unlikely that you can see any improvement beyond that.

Input. The higher your fps, the more responsive your game becomes, even if your display only supports 60 refreshes/second, as most lcds do.

[Gota]

-

[Frostregen]

Input. The higher your fps, the more responsive your game becomes, even if your display only supports 60 refreshes/second, as most lcds do.

Sure?
I did not verify it in the code,
but most (rts-)engines allow gamestate-changing input at every game(sim)-frame only, which is always fixed. (around 30 fps usually)


On Topic:
going to try MT now...

[Auswaschbar]

Input. The higher your fps, the more responsive your game becomes, even if your display only supports 60 refreshes/second, as most lcds do.

Sure?
I did not verify it in the code,
but most (rts-)engines allow gamestate-changing input at every game(sim)-frame only, which is always fixed. (around 30 fps usually)


On Topic:
going to try MT now...

Input is not gamestate-changing.

[Frostregen]

1. input
2. simframe reads current input-state, applies it to gamestate
3. user sees response

Ok, I just thought about gameplay relevant input (unit orders/movement) which by definition lead to gamestate changes.
I forgot about scrolling around the map, selecting units etc.

[Tobi]

The input will have to roundtrip to server anyway, before being processed by simulation, so FPS doesn't matter at all for this. (since network lag is 2 orders of magnitude higher usually)

It's only the "unsynced" input that feels more responsive with higher FPS. (Scrolling map etc., like said before.)

[imbaczek]

non-hardware mouse cursor and most importantly mouse clicks. spring used to have a bug on right-click attack where it'd issue a attack-ground command instead of attack-unit if you clicked on the unit bet let the button go when the unit moved away. with low fps, this can happen unintentionally.

[Neddie]

Virtually nobody has an average APM over sixty at present, however. It isn't reasonable.

[aegis]

my clicks per minute is usually over 60, but my frames per minute is normally over 3600

[SirMaverick]

I get around 10 fps more than normal.

The real gain is when the Sim uses nearly 100% CPU. In that case in ST my fps drops to 1. Then you basically can't do much. Selecting and giving orders takes seconds (for 1 command). Not playable.
In MT under extreme load (speed drops to 0.5) the fps stays mostly between 10-15. I can almost give commands as usual, its more responsive.

[Pendrokar]

HOLY! And I thought Supreme Commander ran better because Spring doesn't support multiple threads.

Then again I just tried 4 vs 4(single player) in Supreme Commander:FA and although it ran flawlessly the game speed became slower and slower(With about 1000 units) till I would need to play multiple hours for a one hour game.

With 1500 Hawks flying I had game speed of 0.85.

Why do people worry about FPS on an processor intensive RTS?

BTW does this experimental exe support more than two cores? If so I regret buying only a two core processor!

[Argh]

BTW does this experimental exe support more than two cores?

Not at this time, no.

[Otherside]

ill try this when i get back from portugal and i upgrade my comp a bit should get some sweeet fps i already got 100 + :p

quadcore 3.6 ftw

[Beherith]

If it hasnt been pointed out yet, lolui doesnt work with this either, at least some components spam gml errors. (ally resbar for one)

Also, one of the unintended but very pleasant side effects is that I can zip through replays really fast now :)

[aegis]

Also, one of the unintended but very pleasant side effects is that I can zip through replays really fast now :)

/skip [seconds]

[Beherith]

Also, one of the unintended but very pleasant side effects is that I can zip through replays really fast now :)

/skip [seconds]

Yes, but its great to be able to watch at nearly skip speeds :)