More RAM seems to improve FPS smoothness in BFBC2

I have been researching this for awhile now on the internet. I have a Toshiba X505 Q892 with an intel i5 460m and a Nvidia GTX 460m and 4 GB of DDR3 RAM. I love Battlefield Bad Company 2 and I don't want to play anything less than maxed out graphics settings. Although my notebook can run BFBC2 at max settings, it spikes up to 100+ fps at some points and dips to 19 fps at others, although it mostly averages a very playable 20-30 fps.

While doing research on this subject online, I've come across many resources, but I can't seem to find the page or link that I've read without spending hours again looking for them. Anyway, one such resource was a video with one of the games developers that insisted getting more RAM to have a better experience with BFBC2. Also one of the members on this site said that upgrading from 4gb to 8gb made the game "silky smooth" for him, so I had to research this and again I came to a study that showed BFBC2 benefiting from 8gb of RAM by holding the average fps steady and consistent, reducing the constant dipping of the fps to anything lower than what your system can average. You get a higher minimum fps. I also read Crysis framerates are "smoother" by having more RAM.

I would like to know before I spend $150 dollars on more RAM if this is actually the case. Can someone with at least the same specs as mine try to see if BFBC2 does infact benefit from 8gb instead of 4gb of RAM. I know it won't give me higher framerates, but according to what I've researched it will give me a "smoother" experience by reducing the fps dips.

Also please don't just post "more RAM is a waste of money and you won't see any improvement in gaming", because even though I've read this in many forums, there were the studies that I mentioned that proves other wise. So at least test it for real before you respond with that. Thanks.

 

$150? That seems too much for 8GB Kit today. And, sure you'll get a overall better performance... I'll say that a hardcore gamer will easly notice the diffrence, I've also have 8GB 1333MHz modules and must say that I've have enjoyed the incrase and overall performance...

 

>very playable
>20-30 fps

Pardon me while I vomit.

 

In most games it actually is the case, very few games benefit past 4GB. (I have experience and have messed around).

 

First, what's your laptop?

Secondly, I have 8GB RAM in both my desktop and laptop now and BFBC2 most definitely has smoothed out. FPS hasn't really improved much, but the visual smoothness has improved greatly. Most likely minimum FPS has improved, not sure how memory would help that, but it clearly does.

 

That's right htwingnut, and it makes hard to explain why to X2P...
I've often got my 4GB all eaten up by GTA 4, even a small windows pop-up and tells me that I must close some programs etc. I really got suprized when I first saw that... 4GB RAM should be more then enough for most users, in my, or in our case, it didin't.

 

lol yeah 20-30 fps sounds "choppy", but believe me on this laptop without FRAPS showing on the upper corner, it feels like 30-40 fps. And I believe 24 fps is what cartoons run at and is the point where the human eye is immersed with the illusion of fluid movement. Something like that...

 

When you talking about max do you mean by running it at 1080p at dx 11 and 8AA and 16AF? I don't think with your system spec could run it at max setting. I ran mine at that setting and the max fps i ever saw was 50( no 100+ fps)and average 37fps unless there a major fire fight it drop down to like 24. I don't think the ram would improve the fps of the game, I have it and mine still dip down to the low 20 when there a big fire fight.

 

Thanks Tunay and HTwingnut. So it will increase my minimum fps with 8 gb of RAM? and I have the Toshiba x505 q892, it has intel i5 460m and GTX 460m vid card. It says I have 4gb of RAM and I figure it's using 2gb in each RAM slot, so if I wanted to increase, I thought I'd have to buy 2 4gb modules, That's why I'd have to spend $150, or can I just buy 1 4gb module and use one of the original 2gb modules? I've read it's not good to use different types of memory.

 

I turned off vsync to get the 100+ fps at some points, and yes everything set to high, Anti aliasing 32x CSAA, Anisotrophic Filter 16x, HBAO on.

EDIT: and running at native resolution of 1680x945

 

I know bfbc2 is very cpu dependent, I think you would benefit more by upgrading your cpu then your ram.

 

That's just foolish. With those AA and AF values no wonder you're getting 20-30 fps.

Also, no, the difference between 30 fps and 60 fps in a computer game is EXTREMELY blatantly obvious, unless you have some fatal eye problem. So tweak your settings so that the average fps is at least above 60 (or 45 with fraps on). There is barely any difference between 8x and 32x AA in a first-person shooter, and even less for AF; the most important thing is framerates.

 

unfortunatley that's not an option, since I just got this notebook, and it's more complicated to upgrade a cpu, so I'm looking at RAM for an easy fix. BFBC2 is very playable on my system, I just get annoyed watching FRAPS drop below 24 fps once in awhile.

 

I'm not saying it DOES improve lowest framerate, as I haven't measured it, I'm just saying it runs noticeably smoother. I can't explain it.

Ah, I don't want to get into the whole what the eye can discern argument again. But everyone's comfort level is different. Some people say they can't live without 60fps, but in my eyes, 35-40FPS+ is usually more than adequate for me. I've learned to live with 25-30FPS depending on the game. Heck a lot of RTS games are locked at 30FPS.

 

True that. My comfort level is 45 fps. Anything considerably lower (e.g. 35 fps) I will notice right away, but I might not distinguish 55 fps from 45 fps. Still framerates are important even if just in the competitive aspect of first-person shooters, since "twitch time" tends to be more framerate-dependent than just viewing.

 

I still do not have to pay that much (Theese are even @ 1333Mhz!)
Newegg.com - G.SKILL 8GB (2 x 4GB) 204-Pin DDR3 SO-DIMM DDR3 1333 (PC3 10666) Laptop Memory Model F3-10666CL9D-8GBSQ

 

I play at around 20-30 fps on my laptop as well (specs in sig). Yes, there are the times when it feels sluggish, but hey, I still usually end up whuppin' some rear ends so I'm okay with it. I'm hoping to build myself a rocking desktop soon enough, so I don't think I'm going to bother upgrading my RAM on this notebook.

I agree though, everyone has their own comfort level when it comes to fps. imho, just go with what looks/feels good to you, not what fps you're trying to attain.

 

lol, back on my old GMA950, I didn't realize I was getting 3-5 fps in CnC3, until I eventually checked FRAPS o.0... It seemed smooth until the panning views of the green tib fields. But apparently, it was slow every bit of the way!!

Anyhow, maybe I'm just a really bad RTS player

 

no you won't ... more RAM simply means your OS don't have to make use of swap space that much because everything is on the RAM ( game files )

i'd say, you might want to lower your "nothing below maximum" standard
because playing with all max settings with good fps just won't happen in a laptop with limited performance like yours

 

well again there are studies with benchmarks that show BFBC2 and Crysis benefit from more RAM by increasing the minimum fps. Have you tried to see if this is actually the case with YOUR laptops limited performance?

 

29% of ram usage during normal browsing and 47% in Bad Company 2 maxed out ( AA 2x , Anisothropic 4x and everything else on the highest settings )

8Gb is pointless ..

 

while running BFBC2 I have Nvidia monitor showing 61% memory usage, just running windows it uses about 30%. How much RAM do you have? I have 4gb.

 

I'm using my Logitech G13 to monitor it. I have 4Gb aswell

 

Something isnt adding up.
My xps m1730 config is in sig.
I get with BFBC2 a min fps avg of 35, max 64, everytime, even with big explosions.
Game is in dx10 mode, 4xAA, 16xAF, maxed out, 1200p, no vsync, using quadros 265.90 modded drivers.
Whats up with the low performance with newer quad based systems? Its not the first time i see it.
BTW Vietnam runs much better since its the updated frostbyte engine (2.1), my avg is around 40-42 fps while gaming.

 

I was hoping you had 8 to actually confirm the extra ram wouldn't make a difference.

 

Could it be that it's because you're running in SLI? I kinda find it hard to believe this though since I have newer hardware. I do get a higher framerate count going DX10 and even more in DX9 mode. Although it annoys me if I have to lower my game settings, I want to play it in all it's glory. One thing that I've observed though is that my hit detection doesn't register as much with max settings than when I lower my settings I get a better hit detection and more kills. That's why I'm wondering if more RAM would help.

 

If my laptop's limited performance is indeed caused by low RAM , then i would sure to see an increase of performance when i upgraded to 8GB , but its not.

i think thats obvious that while you're lagging , every single input commands you send to your computer will be delayed .
You need to sacrifice either a bit of quality for performance...or upgrade your graphics card to the ones with even higher performance

if windows doesn't prompt you that you're low on virtual memory , means your memory usage never exceeds 4GB

 

so you did upgrade to 8gb and don't see any difference of smoothness in either BFBC2 or Crysis?

 

hahaha, I had the same feeling, to me anything less than 40 is not a good experience, I mean seriously, Id rather have crappy settings than 20-30 fps.

 

May I just facepalm this whole thread?

Really, people.



No, 8 GB will not help battlefield bad company 2. it doesn't USE 8 GB of ram. Unless you are doing other things also on your computer which in total sum up to something close to 4 GB, having *MORE* memory WILL NOT help. Having faster memory might help your overall processing performance, but having more memory will not. For those of you who say "it just seems to help, but I am not sure why and the frame rates aren't different" - I suggest placebo.

And, in the OP's case, running 32x aa on a gtx 460, overall processing performance isn't a factor, anyway. You are only able to render 20-30 frames each second, so your processor is just sitting idle most of the time, anyway. Having more memory isn't part of the render path for your GPU to do 32xaa after the system is completely done setting up the scene for rendering and is just sitting idle waiting on the GPU to push its result.

You should give up the idea of running the game "maxed out" if you want reasonable performance and instead try to find visually balanced settings. the mobile gtx 460 is relatively powerful, but wasn't designed to run games at 32xaa. the end.

Do more research and experiment with different settings. You might find that experimentation with a variety of settings gives you more insight and is of more ultimate benefit to you than experimenting with different hardware setups, especially given the fact that your idea of graphical settings requirements, hardware requirements, and frame rate requirements are all in completely different ball parks.

-----

More reasonable approach:

Normally, instead of focusing on maxing out every setting and option in a graphics program or game, we strive to generate the best graphical output. That metric includes the frame rate. So now, it is a balancing act of settings to achieve a desired frame rate. Step 1 is to achieve that balance as best as is possible given your hardware. This requires experimentation and actually looking at the game to see if you can discern differences in quality during gameplay, and how those differences affect frame rate. This will NOT result with you selecting 32xaa, ever. If you are unable to obtain a suitable balance, you may choose to invest in additional hardware. Generally, this means a new graphics card, cpu, or memory (usually in that order). I guarantee you that adding more memory will not make BC2 faster, and that it will not improve the outlook of trying to run it with 32xaa.

ok. the end.

 

In my own experience I also never felt the need to get more ram (went from 2 to 3b withouth any noticable difference in gaming experience). So while everything masterchef says makes perfect sense, this thread is intriguing since it would be a big surprise to me if some games actually do benefit from more ram.

Anyway, are there any people who actually tested BFBC2 with 4 vs 8 gb's of ram? Not that it would matter much to me personnaly, unlike most of the people here i'm not a fps freak (I played AC2 on my 8600mgt at 12-15 fps and thought it was fluent lol). I'm just curious about this...

 

you made me find the stupid benchmark.

Benchmark Results: 3D Games : Do You Really Need More Than 6 GB Of RAM?

anything over 3 GB doesn't matter for all tested games and crysis is one of them. not a SINGLE game got ANY advantage from more than 3 GB of memory.

there is your benchmark. this result is also obtainable via thought experiment, given that you understand how random access memory works (versus, say, a rotary hard disk). Bad company 2 uses about 1.5-2GB of memory. Windows 7 might use about 1 - 1.5 GB.

You need to keep your total memory usage under 4 GB to avoid having to use the hard disk (slow) as additional memory. It doesn't matter as long as you have some memory free, because the memory is random access. You can access all elements with equal speed. This is unlike a hard drive.

The hard drive situation can be like trying to find a parking space in a crowded city. The more inhabitants and cars you have on the street already, the longer it will take to find an open spot. You can combat this by keeping the city organized, to a degree, and filling up parking spaces in order. With memory, it doesn't matter. It's totally unlike that city. If there is a space open, it is always available to you right there, and it doesn't matter how many spaces are already taken. It is as if every space in the city is in a giant ring surrounding you, all equally distant. As long as one is open, you can drive right in.

You need to have enough memory to meet the demands of the applications you are running. You do not need any more. It will not change your performance.

Exceptions - in some circumstances the application may be able to change it's behavior based on the amount of memory and other resources available. Also, changing in game settings may affect the memory required. Battlefield Bad company 2 may or may not have such logic, but it doesn't matter because we know experimentally that the *most* that it takes is 2 GB (regardless of hardware). If you could demonstrate that it took, say, 4 GB under certain conditions, or more, then additional memory might help performance in that case. However, again, we already know it doesn't take more than approx 2 GB at the most.

8 GB of memory will not give you a more consistent frame rate or a smoother frame rate or a higher frame rate, or change your performance in any way, for that matter, unless you force the matter by loading up a ton of other applications and do the test that way. Your performance in that case still wouldn't be better than the baseline with 8 GB, but you would be able to reduce performance below the normal level with 4 GB.

----

Obviously, eventually, games will be targeted for 6 and eventually 8 GB machines. Then, it will help. You will probably also need a better graphics card and processor because those will also be targeted higher, and you still won't be able to run 32xaa.

 

Thats why I'm only running 4GB. Notice how most games install to the "program files(x86)" directory? lol (not that this actually means that something installed in the x86 folder cant be 64bit, its just that this far, most games are x86, and therefore incapable of having more than ~3.5GB of accessible memory allocated to them by the OS.

If you want proof that your RAM isn't your bottleneck, turn down a setting that pretty much only uses up RAM: textures. Reducing textures should more or less simulate what having more RAM would be like. Very unlikely that this is your bottleneck, and that higher texture resolution is slowing you down.

 

except i think those textures are loaded onto the graphics card directly, not system memory.

and, yes, the fact that BF2 is using a 32 bit executable means that it can't use up your 8 GB of memory even on a 64 bit operating system. The application itself is limited to less than 4 as noted above by constraint. doesn't change the fact that it only uses 2 GB of memory anyway.

 

Don't most games have built-in FPS counters? And just because it is going fast enough for the images to blur into fluid motion doesn't mean that that is as far as you need to go. On something interactive, you need to bring the response time as low as you can go, and 20-30fps is certainly not it. Our eyes can distinguish far greater than 20-30Hz.

 

masterchef341 - I would have to agree with you in theory. But I played BFBC2 at 4GB then updated to 8GB and there is a noticeable difference. I can't explain it either and don't know how/why it would work. I can't comment on Crysis because I haven't played it since it was released.

I have pretty barebones apps loaded in the background, and my memory useage is around 1.5GB, although that doesn't mean much with Windows 7 because it caches lots of stuff in RAM when available, and purges when the need arises for higher priority stuff.

However, you do have to be sensible with your game settings depending on your hardware. My settings were untouched from 4GB to 8GB, a mixture of medium and high, with little difference in performance from medium. Biggest thing is if you run at native resolution, get rid of AA, not really really necessary.

So in this case, I'd say reserve any comment until you've tried it yourself. I really don't have the inclination to do testing, as there's no set benchmark for BFBC2, and trying to reproduce a similar experience with 4GB and 8GB will be darn near impossible to do.

edit: Let me just state that it made a noticeable difference on my laptop but nothing really on my desktop. But my frames were quite high to begin with on my desktop.

 

can you do a benchmark?

also, how is your memory set up? if it was 1x4GB and then 2x4GB, then that would improve performance, but not because of the additional memory. you need to get memory in pairs. also, if it was 2x2GB of slower clocked memory replaced with 2x4GB of higher clocked memory... that would also do it.

another thought: running poorly coded 3rd party software with memory leaks could end up eating up all your memory over time. in this case, 8 GB would appear to have better performance than 4 GB, for a period of time, before performance dropped with both setups.

----

anyway, you'll need to explain the difference in the setups and the difference in game.

 

Newer hardware doesnt mean "better" performance mate, infact a fully loaded xps m1730 can kick pretty much good and newer machines in the butt based on fps count when sli kicks in and works as it should.
What im saying here is that a SLI setup really helps, plus a dual core running at 3,8ghz its the added value.

Consider how much BC2 uses the cpu, around the only thing i read is "go for quads" when you can play it just fine.

Maybe because its higher clocked. I have no doubts that with a quad it will skyrocket with fpses, but thats what im getting with my configuration. No 8gb or 6gb, just 4gb DDR2 ram, and still my performance is on par with my dekstop system.

The problem, afaik, lies within the GTX460m. The same problem is afflicting fellow desktop users, those with GTX460-470.
Nvidia already confirmed this: basically the Fermi architecture doesnt have a decent driver yet and a newer one is expected to come in first week of january with - hopefully - a full fix for gpu usage and performance.
Considering the state of the card drivers right now, i think you should get your performance back with the newer driver set.

Also, dont forget that nvidia 260.89-260.99 have some serious issues with GPU usage/performance and punkbuster on everycard, in particular with GTX4xx series (based on fermi chipset, both desktop and mobile). This is a well known fact and everywhere around the interwebs users are experiencing the same issues, low fps and general unsmoothness even on high end systems.

Your only hope for now, and partial fix, is to use modified 265.90 quadros that include BC2 and Punkbuster performance fix, raising the gpu usage up to 85% ingame.
This method has been approved from ManuelG, nvidia community manager and part of the driver team as well.
So, basically, what ive written is all already affirmed and tested by Nvidia itself.

Throw in the mix the fact that i had the same worst performance on both laptop and desktop running 260.89-260.99 drivers, while performance was fine with 258.96 whql, althought not stellar.
Upgrading from 260.99 to 263.00 then quadros 265.90 nearly fixed everything.

 

Imho its a waste of time. If he's running a GTX460m then probably its the driver as i said in my precedent post. The problem is "officially" aknowledged by nvidia itself so its not a rumor or something like that - a quick search around the web can confirm this.

The only driver set that brings back BC2 performance is the quadro 265.90 that contains all the latest fixes, BC2 included+NFSHP+Punkbuster fix+COD:BO.

He needs a modded inf, im sure that his performance will be much better with his system.

 

i recently just bumped up to 8gb of RAM, but i don't soley run just games on my computer while gaming. i often have a lot of other things going on like large VS projects and photoshop/illustrator projects that i don't feel like closing out.

however, coming from 4 gb of memory, i have yet to notice any differences between the two when just gaming.

if your minimum FPS is when you're outside in the game world and the problem is loading many textures and objects, perhaps an SSD will benefit you more than more memory.

 

From 2x2GB to 2x4GB both same timings at 533MHz. Nothing else changed, same PC same config. One day had 4GB playing BFBC2, next day 8GB playing BFBC2. I have no 3rd party apps loaded except MSE and Steam and drivers, all the same. Like I said, I would have been on your side of the argument if I only had 4GB, but since I upgraded to 8GB there's some difference. This is on my Sager NP8662, C2Q 9200 @ 2.4GHz, GTX 260m, all stock clocks.

You can argue as much as you want, but being on the side of the fence with the hardware, I'll believe myself before I believe you. No offense.

 

Why don't you try OC your flu and memory before buying new ram. Ram might help but usually only the memory for the graphics helps with visual stuff but in some games like LFD2 the CPU matters alot so try overclocking and it might avoid the 8gb ram

Oops didn't realize there was 5 pages this has already been explained probally ZDNET

 

I don't mean to change your belief, I just am curious what actually happened.

Did you check frame rates or did it just feel more smooth? I think it was placebo. I'm seriously gonna have to do a benchmark now. How frustrating.

 

How do you measure smoothness exactly? The only way would be to check that the difference between minimum and maximum framerates is smaller, achieving nearly stable fps, but you don't seem to mention this.

 

there was a graph of a benchmark done with BFBC2 using less RAM and more RAM, I forget if it was 2GB vs 4GB, or 4gb vs 8gb. It was a link in another forum, I think the Battlefield forums, but again I went through so much sites and reading it's hard to find it now. Anyway the graph showed a comparison with the fps "spikes and dips" with less memory compared to the more memory and the less memory showed drastic dipping to low fps while more memory showed the dipping to low fps was reduced and held more steady at the average fps. I was trying to look for it all night, since MasterChef put up that other benchmark, which was not the one I was reffering to, anyway I can't right now since I have to go to work I'll try when I get home though.

I think the misunderstanding that Masterchef and the others have is that they think I'm saying adding more RAM will increase top end fps performance. No this is not the case, from my understanding more RAM should increase the minimum fps, in certain games i.e. BFBC2, Crysis, thereby reducing the drops and giving you a more constant average fps, and making the game more "smooth" so as not to break the illusion of being immersed in a virtual world.

 

Benchmarks will solve the matter and add weight to the idea that more RAM = better overall performance. So take the new RAM out and put the old RAM setup back in to run benchmark 1. For benchmark 2, put the RAM back in.

Benchmark 1: 4GB - FPS Min, max, avg.
Benchmark 2: 8GB - FPS min, max, avg.

Make sure all game settings are the same for each test.

Post screen shots of game settings and benchmark results.

Without benches, it will have to be filed under "placebo effect."

 

Nope. There is no misunderstanding. I understand exactly the performance change you are suggesting, and I also know that it doesn't work the way you have described. I'm waiting for the bench. My desktop is out of commission for a while. Someone run the test as 2.0 described please.

 

To be honest, I didn't even think about the 8GB when I went to play. As a matter of fact I forgot I put it in there. I went to play BFBC2 and I was like, wow this thing runs so smooth. It didn't even dawn on me until I thought about it, and remembered I plugged 8GB into my machine the day before. No placebo.

I agree, but as I mentioned earlier, there are no repeatable benchmarks for this game, especially multiplayer. Otherwise I would have done it. Some review sites use the game's single player intro sequence, but I don't think that shows what we want. What we want is in-game multiplayer performance numbers.

I actually recommend the naysayers to buy 8GB RAM and try it themselves.

 

Just play the game for 2 hours each time, and log the min/max/avg. It'll be accurate enough.

 

I can't play for two hours. 30 minutes max usually... darn life gets in the way you know?

Even so, that's still not a repeatable benchmark.