i would like to gather as much data as possible about the performance of this game in regards to all the various systems out there. this game really stresses out systems and brings even the mighty ones to their knees so i think it is the best game for benchmarking your cpu and gpu. the game is DX9 and is multi-threaded as well. if we can keep the test procedures the same it would give us clinically lab-tested results and we can see just how good the systems are between eachtother.
the rules basically are; state your system setup including your native resolution. do not run the tests beyond a wuxga resolution and run all 3 tests the same. your 3 tests should utilize low, med, and high settings. no AF or AA allowed. basically you will want to start off by running with low settings and everything turned off, then the defualt medium settings with bloom, then the default high settings with bloom. you must state if you are overclocked. you must run with sound enabled in the game.
i run supreme commander on my older sager laptop. p4 3.6 with geforce 6800 and i have to run it at all low settings or else the frame rates are unplayable.
the game has a built-in benchmark test. this test does not take very long and basically auto-runs a map and plays by itself for a few minutes. to run the test go into the properties of the game's shortcut icon on your desktop and in the target line at the very end add a space bar and then type in /map PerfTest
then simply launch the game. the higher the scores the better. to run the game normally just delete what you added. you may want to run the game normally to set the game's graphical settings first, then run the perftest after.
the game creates a log file for each test you run and the results are in there. log files are located in C:\Program Files\THQ\Gas Powered Games\Supreme Commander\bin. what we are looking for are the sim, render, composite scores and the average FPS (which can all be found in those log files that are created). i will start off with my results.
Sager 9860-S Laptop computer
3.6Ghz P4 HT
6800 GDDR3 256
2x 80 gig sata drives on raid 0
XP PRO
WSXGA 1680x1050
1.5 Gigs of RAM
system is not overclocked
-sound on in all tests
-game played in window mode at about a 1400x900 resolution as i need all the fps i can get. (i really should run the tests are wsxga though.)
1st test] low settings with everything to off / low
SupComMark (sim) : 7591
SupComMark (render) : 6408
SupComMark (composite) : 13999
AVG FPS = 25.354
2nd test] medium settings with bloom render on, no aa
SupComMark (sim) : 7451
SupComMark (render) : 2061
SupComMark (composite) : 9512
AVG FPS = 12.496
3rd test] high settings with bloom render on, no aa
SupComMark (sim) : 7517
SupComMark (render) : 978
SupComMark (composite) : 8495
AVG FPS = 11.196
edit; i also want to add that the sim score is for the cpu, the render score is for gpu performance and the composite score is the total of the two. in my case, my gpu is more of the bottleneck here as the render score gets absolutely pummeled at medium and high settings. so you can easily tell that on my system you will have to run the game at low settings to have a playable framerate. average fps is a great indication also.
i am very curious to see just what the dual cores and sli rigs can really do as the general census for this game is that it is cpu-intensive and dual cores are a requirement, etc. i was surprised to find out that my prescott p4 3.6 HT held up reasonably well for being a single core. i hope we can get into a situation where one rig tests with one gpu and another rig with similar setup tests with sli to see how good that sli is!
-
Will run it tonight on my MBP (8600GT 128MB GDDR3).
-
See below post
-
My desktop: Athlon 64 X2 4200+, 4GB RAM, 640MB 8800GTS, WinXP Pro
Settings: 1680x1050, Maximum details, 16xQ AA
SupComMark (sim) : 8513
SupComMark (render) : 5837
SupComMark (composite) : 14350
AVG FPS: 26.091
I'm fairly certain it's my processor holding me back. Always is. The minimum framerate was 3.89, too. -
how do u turn everything down all the way? that game runs like below 25fps all the time for me even if i turn everything down. i dun get it.
-
Wu Jen, you need to have the newest patch for the game as it greatly improves performance. also, the newest patch implements a new benchmark formulation where you will be getting results like mine (higher scores = better).
also, can you guys not use AA or AF as it skews the numbers? i think it is best to leave them out so that way we can compare raw benchmark numbers with older systems, as older systems cannot ideally run with aa or af for this game. also, do not enable v-sync for either test.
if you insist on running with aa or af or any other combination you can do so after you completed the default low and high settings.
imhungry29, i don't quite understand your question. but you basically want to load the game, go into the video options and put the fidelity pre-sets to low, turn off bloom and disable v-sync. the exit the game and add the "/map perftest" (without the quotes) in your desktop shortcut to the game and luanch the game. it will run it's own test by itself. then take out the command line you added and load the game up so you can change the vid options to high fidelity presets with bloom on, (v-sync still off) to prepare for the other benchmark.
thanks! -
there is no option for me to turn bloom off. i might have a wierd version or something.
-
you have the retail game, right? options > video > bloom render is the last option on that menu, you might need to scroll down to see it.
-
Geez..those numbers I see above aren't an accurate representation, I hope, or else I won't be playing this anytime soon
-
Here are my results:
Lowest Settings possible
SupComMark (sim) : 8476
SupComMark (render) : 7173
SupComMark (composite) : 15649
minfps 8.73 maxfps 61.77 avgfps 32.09
MidHigh settings - ShadowFidelity on meduim and fidelity on low(my normal settings)
SupComMark (sim) : 8195
SupComMark (render) : 1393
SupComMark (composite) : 9588
Minfps 3.26 Maxfps 21.99 Avgfsp 10.37
Highest settings
SupComMark (sim) : 8318
SupComMark (render) : -733
SupComMark (composite) : 7585
minfps 2.75 maxfps 16.29 avgfps 8.1
Highest settings OCed
SupComMark (sim) : 8313
SupComMark (render) : 190
SupComMark (composite) : 8503
min max avg: 4.20, 16.76, 8.63
lowest settings OCed
SupComMark (sim) : 8486
SupComMark (render) : 5713
SupComMark (composite) : 14199
min max avg: 7.62, 29.76, 19.09
Now just clearify, i didnt OC alot. only an extra 50-75mhz on cpu and mem. Im afraid of OCing to its fullest because im afraid of frying it. now to be honest I have a (not by much) outdated notebook compared to the ones out there now. But I dont know what all those numbers mean. I mean I kno i have a pretty ok notebook but what do those #'s mean, is my notebook decent, crappy, good enough to last a lil longer? Especially since i play alot of supreme commander and those big maps get in the way of the frame rate. (Its playable, just not way into the game when alot of units are on the map)
PS. Someone mentions a patch, I think I have the latest patch but idk if there's a different one to reduce the frame/lag issues. I usually log into GPG so it can auto update but not sure if its the latest one. -
-
malikite, if you log into GPG you have the latest patch as it updates automatically.
it appears that my geforce 6800 is outperforming your geforce 7600 by a little at medium and high graphical settings, so that is interesting. although it is a moot point as i cannot play at medium or high so overal, you have the better rig when it comes to low settings. i am surprised to see my single core p4 3.6 ht is holding it's own quite well to your dual core.
what your numbers basically mean are that you need to play the game on low settings in order for it to be playable. i would say your notebook is at the low end scale in terms of gaming at this moment largely due to your graphics card. your gpu may be 128 bit or only has gddr2 ram i am guessing. -
i am sure there is headway to improve the game's code. but i do not think the poor performance is largely due to coding. this game has real time trajectory and ballistic calculations instead of the "roll the dice" method used c&c3 where everything is predetermined on what gets hit, does damage, etc. you can see this in the built in benchmark where sometimes the battles do not always end the same way.
so i dont think it is a poorly written game. we are seeing a ton of physics calculations for thousands of units at one time. this is why i think it's pretty much the best game to use as a benchmark for today's gaming rigs instead of using half life 2 and doom 3 engines. -
I would be interested in benchmarking this one. i think i will go and buy this tomorrow
-
well my rig isn't exactly a low end, tho it could be considered a fairly low powered machine, i'll vouch for that. But it still holds its own, maybe not in supreme commander. tho i'm shocked at my results, but not surprised at the expectations of notebook. there are newer video cards out there that have surpassed mine.
Btw. i ran them all on 1440x900 display setting and on the lowest setting i ran it at 1024x768. maybe that affected the trail. but regardless its still disappointing on a 256 card -
Wow. This game seems to have insane requirements to play at a decent clip. Are powerful desktops even able to play this game at max settings above 60 fps?
-
seems like the amount of logic and physics in the game make it tough to run smooth on even high class modern pcs. But i bet the cpu is critical in having a substantial machine and an exceptional one.
-
According to a magazine review- the desktop alienware alx system burned through supreme commander at the highest settings with all the goodies maxed. (of course its specs are two 8800 ultras and a quad core / liquid cooled) as far as notebooks... mine is as high end as it gets, and i am nearly positive that when i test this...I wont even be able to max everything and have very smooth play
-
Just to inform you guys of my desktop:
Dell XPS 400, PentiumD @ 3.0ghz, 2GB RAM @ 533mhz, 7900gs.
I can comfortably play it on MAX, no AA, at 1280x1024.
So, I don't know about the efficiency of the coding really, but as the others have said; Even though the model detail isnt the best on some of the units on full zoom, the max unit counter per player is 1000. All units get fully softened realtime shadows, bloom takes up GPU power, and with all the physics, the game seems to be more CPU bottlenecked than GPU.
Not to mention, even on OLD games, if there were equivalent maps to the 80km x 80km monstrosities and you maxed out the player count, nearly anything would grind to a halt.
8 players x Average 600-800units on a big map= 4800-6400 units to track AI, shadows, and physics for. If put into that perspective, I'd be surprised if a computer doesn't slow down with that... -
Edit: Nvm found it here. -
if you launch GPGnet (the shortcut on your desktop that you use to go online with the game) it should automatically apply updates for the game i believe.
-
Okay, updated and ran it at 1920x1200 resolution. Spec's for my laptop are in my signature. Ran it in Windows XP SP2. Here are the results.
1st test low settings with everything to off / low
SupComMark (sim) : 9076
SupComMark (render) : 8017
SupComMark (composite) : 17093
AVG FPS = 41.287
2nd test medium settings with bloom render on, no aa
SupComMark (sim) : 9064
SupComMark (render) : 6533
SupComMark (composite) : 15597
AVG FPS = 25.829
3rd test high settings with bloom render on, no aa
SupComMark (sim) : 8989
SupComMark (render) : 5998
SupComMark (composite) : 14987
AVG FPS = 21.305
Updated Video Drivers!
4th test high settings with bloom render on, no aa
SupComMark (sim) : 9017
SupComMark (render) : 6181
SupComMark (composite) : 15198
AVG FPS = 23.418
So, it's playable at high on WUXGA. But I'll probably play it on 1600x1200 to keep the FPS above 30. With all the goodies turned on
When the 8800's are released it will probably be very playable on high at 1920 x 1200.Attached Files:
Last edited by a moderator: May 8, 2015 -
-
-
I actually have one more concern about the game. IDK if its my notebook or something with the game but it seems to heat up unmercifully when playing the game. Luckly i thot ahead of time and I had a cooling pad to prolong the life of the notebook and to keep my knees cool. But its strange how any other application doesnt heat it up as much as this game does. ANY info?
-
I had a problem like that with my desktop & the Warhammer video games. In CS:S I'd run at like 75C, and during Warhammer it would be around 89C for some reason. It depends on the coding and how much it takes advantage of the hardware. It also depends on the consistency with which it puts a load on your hardware. If its constant, then the heat never gets a chance to fully dissipate, so it tops off somewhere. I know with SupCom,on a dual core setup, the second core goes to AI functions. And with a lot of units and/or bots, then it could get hairy. Not to mention all the bloom, shadowing, etc that would consistenly load a GPU.
Basically, it means that its making your laptop crap itself, so it gets hot. -
I ran this test on my MBP. It runs perfectly until the end when it is supposed to display the score. Instead, Windows states that the program has crashed and closes Supreme Commander. I've tried it several times with several different graphical settings (including all low).
I'm running Vista Ultimate, 8600m GT with the 162.18 patch. Anyone have a fix or an idea? Is it Vista?
I used sysinternals process explorer to check and see if I was hitting the 2gb memory barrier, and I'm not. -
wu jen, is it possible for you to turn off one of the graphics card? if so, i would be curious to see your scores without sli.
i don't think the game is compatible on a mac. -
So no, it's not going to work in osx, but it WILL work on a Mac running Bootcamp. -
Would anyone like to run a benchmark, or make an estimate on what an X1400 would get?
Thanks.
the supreme commander benchmark thread!
Discussion in 'Gaming (Software and Graphics Cards)' started by ARGH, Jul 25, 2007.