Evidence of major performance increase in new version(?) 56k

Caydr · Post by **Caydr** » 21 May 2007, 02:35

Righto, here I am with some illegitimate, unverified, and most likely inaccurate news! As usual.

Let's begin with screenshots of the current version of Spring, taken with Fraps running for FPS:

Now moving on, here's the current SVN build in action:

And, for what it's worth, SupCom in action:

Conclusion: the new version of Spring brings with it a performance increase of between 45% and 26%, depending on the settings you use. Also, if you can currently run Spring with all settings maxed, you can also probably run SupCom at maxed settings as well... at least, until things begin to heat up...

jackalope · Post by **jackalope** » 21 May 2007, 02:41

lol supcom is crap

Caydr · Post by **Caydr** » 21 May 2007, 02:46

jackalope wrote:lol supcom is crap

lol u r fag

SwiftSpear · Post by **SwiftSpear** » 21 May 2007, 05:48

Your computer has always been kinda weird with handling spring though caydr... It always kind of made me WTF that it didn't see any performance different between running many high poly models and many low poly models.

LOrDo · Post by **LOrDo** » 21 May 2007, 08:00

Hes not using antialiasing, that's problably why polycount didnt make a diffrence for him. That was my crappy conclusion anyways.

Ishach · Post by **Ishach** » 21 May 2007, 08:04

svn screenshots are a little more zoomed in, less stuff on the screen --> higher fps

Post by **jcnossen** » 21 May 2007, 09:56

Not enough for such an increase
AFAIK there hasnt been any update of the rendering system

Neddie · Post by **Neddie** » 21 May 2007, 10:25

jcnossen wrote:Not enough for such an increase
AFAIK there hasnt been any update of the rendering system

JC is on the button, we need more evidence.

Tobi · Post by **Tobi** » 21 May 2007, 11:56

Caydr, SVN build from buildbot or self compiled? If self compiled, which compiler?

1v0ry_k1ng · Post by **1v0ry_k1ng** » 21 May 2007, 14:31

supcom IS crap. have you tried playing it competively online cadyr? everyone plays aeon, and everyone spams aurua tanks. only when complete noobs are playing does the game reach T4

drolito · Post by **drolito** » 21 May 2007, 14:36

Supcom sucks and use too much CPU memory ...

I cant play it without lag but i have a good PC which allow me to host for 16 ppl in Spring ...

moreover, u can't saw real unit and play in supcom ... obliged to be in symbolic view (with square, round or triangle) for fight ...

It sucks a lot !!!

and has ivory just say : spam T1 is the only good strategy ... and units are less varied so its always the same poor battles !!!

I want to be refunded for the purchase of Supcom !!!

Zenka · Post by **Zenka** » 21 May 2007, 14:49

well indeed supcom runs nice, untill you've a lot of units and your CPU melds away (that the case at my pc anyway).
And spring doesn't slow down very much with a lot of units. And with lower settings you won't need a pwnage pc anyway.

however, with a lot of units the lag becomes to annoying. the game speed is lowered to keep the slowest system connected happy. thus, there might be a single player being a bottleneck. sometimes the host if forced to ask the player to leave to regain a reasonable gamespeed. people with a Pentium 3 600 mhz and a dial connection should not be allowed in great 10 player games.

drolito · Post by **drolito** » 21 May 2007, 14:54

In spring i always host for 16 ppl and i saw few lags ... its sure u have lag in a Green fields when u have 200 fighters patrolling arround each base ...

But a "normal" game like altored or deltasiege dont make lag ^^

today its deltasiege for 16 ppl with me if u want ^^ it will be very funny !!!

maybe silent will make a deltasiegeRevX dry for more fun ??? i hope so :p

AF · Post by AF » 21 May 2007, 16:29

Visual studio 2003 builds regularly give 30-40% more performance than the mingw32 builds used in the installer releases.

Boirunner · Post by **Boirunner** » 21 May 2007, 16:47

AF wrote:Visual studio 2003 builds regularly give 30-40% more performance than the mingw32 builds used in the installer releases.

If this is true somebody should host a binary made with VS. As long as the hoster owns VS there should me no legal issues, right?

Caydr · Post by **Caydr** » 21 May 2007, 16:54

The version I used was an automatic build from that... place with all the automatic builds...?

As for zoom making such a big difference, zooming out *very* slowly, my FPS ranges from 341 to 143 with the most recent build. 143 is at a point *just* before I hit max zoom. edit: I actually posted "143" not remembering that my first image (74b3) gave me 142... it's just coincidence.

As for my computer "just being weird", this is a brand-new computer completely unlike the one I had before my disappearance.

tooleh · Post by **tooleh** » 21 May 2007, 16:58

AF wrote:Visual studio 2003 builds regularly give 30-40% more performance than the mingw32 builds used in the installer releases.

Why don't they release vs2003 builds then?

Could you host some binaries for us?

Tobi · Post by **Tobi** » 21 May 2007, 16:59

They desync.

Caydr · Post by **Caydr** » 21 May 2007, 17:03

Tobi wrote:They desync.

They desync period, of they desync with linux builds?

Boirunner · Post by **Boirunner** » 21 May 2007, 17:05

So Linux compatibility cost 40% performance

Aren't there any other cross-platform compilers that would produce better binaries?

Spring RTS Engine

Evidence of major performance increase in new version(?) 56k

Evidence of major performance increase in new version(?) 56k