Borderlands Performance Review

[LAiN_OfTheWired]

Let me start off with the fact that this is not a 'clean' test. I had Facebook open in Firefox, uTorrent downloading OpenSuse 11.1, and GPU-Z 0.3.7 open in addition to FRAPS to measure my framerate.

Also, my monitor's native res. is 1440x900, so a bit less than your average enthusiast. Still, looking at the results, how long the test ran, and taking into account the settings I used for everything, you should be able to gauge your performance from this somewhat well.

In-game settings are all set to high (16x AF), bloom+depth of field+flare outs+dynamic shadows+bullet decals are all on, 16xQAA is forced in the nVidia control panel (globally and confirmed for Borderlands profile).

I also set everything to stock clocks for this, so my Phenom II X3 720 is at 2.8GhZ, and my GTX 275 is at 648/1440/1152, RAM is at DDR3-1600 7-7-8-24. I currently have 190.38 drivers installed, have not patched Borderlands at all, and have not modified any of the game's config files to help performance.

The reason I did this? The complaints that some systems are running it REALLY poorly. If your system is running it poorly on your $1200 worth of hardware, you've definitely got a problem.

The drops on the graph DO play into the framerate results you see there, but those are actually when I'm going from one area to another. The transitions screens were showing ridiculous results when I had the FRAPS overlay on my screen, when switching areas it would always drop to some single-digit number.

As you can see, my system doesn't even sweat trying to run it and my system is definitely not some monster with a Radeon 5870/GTX 295 and an overclocked Core i7 with 6GB of DDR3-2000.

EDIT: Lifted framerate cap to 120, still bounce off the limiter sometimes. I'll do OC results in a few. You'll hafta ignore the end of the chart there, thats where I was exiting the game.

 

[pandora's box]

how did u run the benchmark?

btw those dips in frame rate would drive me crazy.

 

[Intel_Hydralisk]

how did u run the benchmark?

btw those dips in frame rate would drive me crazy.

He said he used FRAPs. FRAPs has built in FPS logger = benchmark.

 

[overclocking101]

i play this game with my hd4870 512mb and 8800gt for physx and i have no dips whatsoever its 100% running at 88fps with other apps running everything maxed at 1440X900 odd maybe its my physx card that gives me so much better performance??

 

[LAiN_OfTheWired]

i play this game with my hd4870 512mb and 8800gt for physx and i have no dips whatsoever its 100% running at 88fps with other apps running everything maxed at 1440X900 odd maybe its my physx card that gives me so much better performance??

The frame rate is capped, I mentioned above that I have not modded any of the configs. I suppose I will give that a go for further comparison results.

EDIT:

Your sig seems to say a bit otherwise about your setup... not that I don't believe you, but you should aim for consistency I'd say the results are pretty much the same, anything over 62 and you won't see hardly any difference, Unreal Engine is capped at 62 by default for a reason. Your 5FPS average could be attributed to a couple of things. 1) The slightly higher clockspeed of the E8400 2) the fact that you have 4x the RAM that I have 3) You're obviously on a 64-bit OS so some things are running down the CPU pipeline faster, whereas I'm on Vista 32-bit. I'd have to see the chart parsed out from a FRAPs benchmarking run to believe you have "no dips." My benchmarking run for the 120FPS cap was taken while I was completing those two first missions for Crazy Earl, LOTS of gunfights with the bandits, explosions, those crab things, etc etc

Also, I'm totally looking forward to tax rebate time, I want a bigger monitor! I thought I was the only person alive left on a 1440x900 lol

 

[sirmonkey1985]

i play this game with my hd4870 512mb and 8800gt for physx and i have no dips whatsoever its 100% running at 88fps with other apps running everything maxed at 1440X900 odd maybe its my physx card that gives me so much better performance??

yes it is your extra physX card that is helping.. physX is a biotch when using single cards.. doesnt matter what resolution or what card it is, physX still effects the performance..

but you are also running an overkill card at the resolution you play at..

op'er: and as much as you want to say that its our fault the game runs like crap.. try playing the game at 1920x1200 or higher with your card and you will be in the same boat as us.. this game runs like crap because its yet again another crap ass console port.. most cames are scale much better the higher you raise the resolution.. in borderlands thats not the case.. each step higher you go in resolution takes out anywhere from 25-30% of the frame rates.. 1920x1600 i can run at around 55-60fps.. 1920x1200 im lucky if i can hold 30fps.. hell i get a better frame rate in crysis with that maxed out except AA and AF..

 

[BababooeyHTJ]

yes it is your extra physX card that is helping.. physX is a biotch when using single cards.. doesnt matter what resolution or what card it is, physX still effects the performance..

but you are also running an overkill card at the resolution you play at..

op'er: and as much as you want to say that its our fault the game runs like crap.. try playing the game at 1920x1200 or higher with your card and you will be in the same boat as us.. this game runs like crap because its yet again another crap ass console port.. most cames are scale much better the higher you raise the resolution.. in borderlands thats not the case.. each step higher you go in resolution takes out anywhere from 25-30% of the frame rates.. 1920x1600 i can run at around 55-60fps.. 1920x1200 im lucky if i can hold 30fps.. hell i get a better frame rate in crysis with that maxed out except AA and AF..

This game doesn't use Physix so the dedicated card is just idling.

I run this game at 1920x1080 with 4xAA on a GTX280 and Q9650 with Vsync since the screen tearing is horrendous otherwise and have yet to notice a dip in framerates.

 

[LAiN_OfTheWired]

OK, so... A few things to address. I've spiffied up my testing to a specific run and tried to create utter chaos throughout pretty much the whole thing. The test time is exactly 5 minutes, and I accomplish 3 strikes within that limit. I'm in a vehicle a good amount of the time equipped with the rocket launcher (which I use quite heavily to really put the stress on). I have done two tests, one without a modded config (except I've raised the FPS cap to 180) and one with a config that should allow for more eyecandy.

I'm not saying its the end-user's fault the game is running like crap (if it is for you), I'm saying that there may be some things you need to change, though. Obviously, this game was built on newer hardware, so you can't really expect 8000+9000 series Geforce cards to max this game out at higher resolutions. The same really goes for anything below a Radeon 4850.

Also, keep in mind I have 16xQAA, transparency supersampling, 16xAF, high texture quality forced in the Nvidia control panel in addition to having cranked all of the games settings to highest. The developers even mentioned that there isn't a significant image quality improvement with certain settings even though there is a large performance impact with them. Google should help you find more specific information.

Here are the results:

FPS Uncapped, Unmodified Config, Stock Clocks

FPS Uncapped, Unmodified Config, OCed

FPS Uncapped, Modified Config, Stock Clocks

FPS Uncapped, Modified Config, OCed

OCed Settings

Now, here's the funny part. I modified the config to be more intensive. Kinda odd how I got consistently better results with it, huh?

Changed made:

Turned hardware physx support on:

bDisablePhysXHardwareSupport=False

Turned FPS cap up to 180:

MaxSmoothedFrameRate=180

Turned motion blur on:

MotionBlur=True

Turned high-quality bloom on (turned off by default, even when bloom is on):

UseHighQualityBloom=True

Allowed DX10:

AllowD3D10=True

Now I'm unsure if a couple of these are even doing anything, like allowing DX10 and motion blur. What I can tell you is things are a bit more shiny and lights glare a bit more. This seems pretty typical of DX9 vs DX10 in the unreal engine (gears of war PC was the same way). It shouldn't be ignored, though, that I'm getting a better frame rate with these changes made than the default config. If you're getting a crappy framerate, try these (especially if you have an Nvidia card or dedicated physx card).

This testing was still done with the older 190.xx drivers and original version of the game. I don't expect the patch to really affect the performance, but drivers might so hit up those 195.xx betas if you're having trouble and using an Nvidia card.

ALSO: IF YOU'RE WILLING TO DONATE A MONITOR CAPABLE OF 1900X1200 LET ME KNOW

 

[piratepress]

There's a built-in benchmark tool that was fixed with the 1.0.1 patch. Using it, my minimum was 1 fps, avg was 65 fps, and my max was 275 fps.

 

[shansoft]

He said he used FRAPs. FRAPs has built in FPS logger = benchmark.

meh, this game does cripple nVidia cards in some specific parts, but its rare to happen.

the fps drop to 10 like in certain area.

IIRC, there is only 3 places in the whole game that will do that, so not a big deal.

 

[LAiN_OfTheWired]

There's a built-in benchmark tool that was fixed with the 1.0.1 patch. Using it, my minimum was 1 fps, avg was 65 fps, and my max was 275 fps.

Is there a config hack for the time demo? I purchased physical media for the game and it is not to be found anywhere in my install.

@BababooeyHTJ - The game definitely does use physx. I don't know where you got the idea that it didn't. GPU-accelerated physx is off by default, and is one of the changes I made in my config. There are physx DLLs in the install folders in addition to the config switch. Enabling GPU-accelerated physx definitely did make a performance impact.

 

[shansoft]

Is there a config hack for the time demo? I purchased physical media for the game and it is not to be found anywhere in my install.

@BababooeyHTJ - The game definitely does use physx. I don't know where you got the idea that it didn't. GPU-accelerated physx is off by default, and is one of the changes I made in my config. There are physx DLLs in the install folders in addition to the config switch. Enabling GPU-accelerated physx definitely did make a performance impact.

uh, its UE3, all UE3 have integrated physX in there IIRC.

and no, there is definitely no impact when I turn that on or off, at least I don't see a difference throughout the whole game.

 

[LAiN_OfTheWired]

You might want to run your own individual tests and parse out the graphs if you want to see. The physx in the game are really minimal, but a config with extra eyecandy (and we all know DX10 also almost always incurs a performance impact) consistently performs better in both minimum framerates and average framerates. I would have to say that letting the GPU handle what little the CPU was being forced to do helped. I have nothing else to explain the 5.5% performance improvement in average FPS and 9% improvement in minimum FPS, especially since the game DOES look a bit different (another thing I mentioned, shiny things were a bit more shiny and lights had a bit more of a glare to them amongst other VERY LITTLE details). The evidence really is there to support my theory.

 

[piratepress]

Is there a config hack for the time demo? I purchased physical media for the game and it is not to be found anywhere in my install.

Mine is on physical media also. I ran the demo from a command prompt and then looked in the log file to see what the min/max/avg fps were. Google it and I'm sure you can find the directions for running it.

 

[shansoft]

You might want to run your own individual tests and parse out the graphs if you want to see. The physx in the game are really minimal, but a config with extra eyecandy (and we all know DX10 also almost always incurs a performance impact) consistently performs better in both minimum framerates and average framerates. I would have to say that letting the GPU handle what little the CPU was being forced to do helped. I have nothing else to explain the 5.5% performance improvement in average FPS and 9% improvement in minimum FPS, especially since the game DOES look a bit different (another thing I mentioned, shiny things were a bit more shiny and lights had a bit more of a glare to them amongst other VERY LITTLE details). The evidence really is there to support my theory.

how does physX make things more shiny?

and this game don't even run in DX10 mode, even if you set it to DX10 in config but it still run DX9.

 

[robble]

Using the rig in my sig I hit low 30fps,ave ~55, max 180 fps using the built in time demo. 1680x1050 all settings at max in the config using dx10 mode.

and this game don't even run in DX10 mode, even if you set it to DX10 in config but it still run DX9.

do you have a source for saying that?

 

[WabeWalker]

This title is selling for 33$ on Steam for another 15 hours - worth it?

 

[pandora's box]

Using the rig in my sig I hit low 30fps,ave ~55, max 180 fps using the built in time demo. 1680x1050 all settings at max in the config using dx10 mode.



do you have a source for saying that?

run a program that detects what directx version is being used. msi afterburner is capable of this i believe. setting it to directx 10 = directx 9, the setting is broken

btw:

1920x1050 max in game settings 16xaf, no aa:

min= 36.72 avg= 73.24 max= 254.49

 

[Q-BZ]

This title is selling for 33$ on Steam for another 15 hours - worth it?

Hell yes! One of the best games I've played this year.

 

[WabeWalker]

Hell yes! One of the best games I've played this year.

Really!

Okay, it looks like it's going to be Borderlands, Trine, and Dawn of Discovery then - and on Thursday it was Grid and Resident Evil 5.

Normally, I would've said that since all these games are on sale now I'll be buying them now, but will be reserving them for the lean months of January, February, and March - except that argument doesn't hold water since during those 3 months we're going to be getting Mass Effect 2, Star Trek Online, Bioshock 2, and Assassin's Creed 2... a start to the year like no other before... and I mean that literally... no year has ever started off with such a bang for PC gaming.

The thing is, you never know which game is going to knock your socks off until you've actually played it. Some of you guys will know, for example, that Risen just blew me away this year. Sometimes you just have to buy a whole bunch of titles (preferably discounted) to find the one game that really does it for you.

 

[BababooeyHTJ]

Is there a config hack for the time demo? I purchased physical media for the game and it is not to be found anywhere in my install.

@BababooeyHTJ - The game definitely does use physx. I don't know where you got the idea that it didn't. GPU-accelerated physx is off by default, and is one of the changes I made in my config. There are physx DLLs in the install folders in addition to the config switch. Enabling GPU-accelerated physx definitely did make a performance impact.

That could be some leftovers from the port or for modding or something but the game obviously doesn't use hardware physix.

Oh, I lied about my performance before. In some of the more demanding areas I'm seeing dips down to about 40fps.

 

[Q-BZ]

The thing is, you never know which game is going to knock your socks off until you've actually played it. Some of you guys will know, for example, that Risen just blew me away this year. Sometimes you just have to buy a whole bunch of titles (preferably discounted) to find the one game that really does it for you.

Borderlands was that game for me, although I, too, really like Risen a lot.

I figured I'd like Borderlands...I didn't expect to just fall in love with the damned thing like I have. Even playing single player by myself has been a great blast.

For me, the best games that I've actually played so far this year are Batman, Risen, and Borderlands.

I'm still trying to get into Dragon Age. That's an awesome game but I've never been that good at "old school/hardcore RPGs."

It certainly is an excellent game of that genre, though. So yeah, technically, I guess I should mention that as well although I really haven't played it a whole lot because I'm just not in the mood for that kind of RPG right now.


When I play Assassin's Creed 2 I'm sure that will quickly join the list. Might even be my GOTY when I play it.


I think it took me maybe 2 minutes tops for me to fire up Borderlands, see the art style, get used to it, and then really like it. I honestly can't imagine the game any other way now. I love the art.

The gameplay is fantastic, the game itself is addictive as all hell...and as far as the story goes? Sure, it could be better but they kind of pull a Bioshock here: If you don't haul ass and rush through the game, you get a lot of side missions and some of them include picking up recorders from certain characters that flesh out the story a bit more.

There is no real ending to speak of. I can't defend that. I can only hope this game gets maybe a Broken Steel-esque DLC down the road some day. But it really doesn't matter to me because everything else just kicks ass and I, at least, come away with the idea that there's more than meets the eye with the story and things where your imagination can plug in some holes. I definitely come away with the idea that there's a larger mystery and situation being implied.

Is the story Dragon Age caliber or even Risen caliber? Of course not.

But it really doesn't matter. The journey kicks ass, everything else in the game kicks ass. I can't see a FPS fan worth their salt not liking this game A LOT. I love it. I really don't have a single complaint about it.

I think there easily be a Borderlands 2 some day.



http://gbxforums.gearboxsoftware.com/showthread.php?t=79043

^^ This thread is a must read for any owner of the PC version of Borderlands. Now you can download a nifty little app that sets you right up, right out of the box. No .ini tweaking necessary.

Works with the Steam version just fine as well.