More RAM seems to improve FPS smoothness in BFBC2

I do not know extroardinarily much about ram usage and such, but from my understanding, most games will use up a certain amount of your ram, but If you have more ram, I think for some reason it can use up more of it? also the ram lag is when it is pulling up textures and meshes for a bullet or an explosion, so having faster frequency ram I beleive can solve some of this. I run BFBC2 on maxed out every single thing except i think it was AA and multisampling I left at 8x, and I get something like 40-100 fps, sometimes in some ultra super fire fight where i am running through explosives and bullets flying everywhere, It can even go down into the 20s of fps. this is just my opinion, and experience. hope this helped.

 

That actually is repeatable. You could actually do several two hour sessions and *use math* to build a confidence interval about the range of frame rate over that time period. In fact, you could do 30 minute sessions. That would be repeatable also. You'll just need more of them to generate the same confidence interval.

Also, the fact that you didn't purchase the 8 GB of memory for the purpose of gaming doesn't mean that it wasn't placebo or another factor.

---

The problem with this is that we have one person who had one experience where A changed and B changed and got it in his head that A affects B. We also have several people who believe A might affect B, or heard it somewhere. None of that is scientific.

Having gone through a CS major and having a complete understanding of the mechanics of the computer, I *know* that A does not affect B. Additionally, the only scientific evidence anyone was able to provide was by me - and it showed very explicitly that A does not affect B. But, that doesn't seem to be enough to end it for some reason.

---

I feel like a physicist having a debate with aristotle about the speed at which objects fall to the earth.

 

Without explicit proof, I changed the thread title to say, "seems to."

 

Gonna be a long wait as there won't be any ..

By the way , the recommended system requirements regarding Rams is 2GB.

 

qft.

10char

 

I haven't seen my system use more 3 GB of ram so far while gaming. So I doubt having more than 4 GB would make any difference.

ROFL, even with 8GB, DA:O still gets sluggish with it's terrible memory leak problem. No amount of ram will save you from BioWare's ineptitude.

 

i feel more RAM could help in mmo's where textures or maps could be dumped into memory, but really, i don't see the 8gb improving fps by a significant amount by just the memory alone.

i have not noticed any gaming improvements since i installed 4 more gb to my motherboard.

 

I feel it's ridiculous that you're arguing facts and don't have the hardware with someone that has the hardware. Additionally things happen all the time that we think we know that we really don't. So again, reserve further comment until you have the hardware or evidence to back up your claims as well. The specs on a hard drive says it consumes 1W at idle, so you can keep spouting that all day, when I am running that hard drive and find it runs only 0.7W at idle. Are you going to continue to tell me I'm wrong?

I have an Engineering degree, MCSA certified, and a certified pilot, and have 15 years experience as an engineer, I don't need to be educated on scientific methods or lectured to about it. I've run enough experiments and tests to know I know a lot less than I thought I did, not to mention our collective understanding of lots of things in general. Not to mention I've got over 25 years dealing with computer hardware and software and gaming.

Running two 30 minute sessions (or however long) would have to have the same number of players, on the same map, running identical scenarios for it to be considered repeatable. With your suggestion, it'd be like driving a ferrari from A to B in traffic and see how fast you can get there, and then do it again with different traffic patterns. You can't deduce anything from that.

 

Yeah, that's cool. But it would be neat to see if there are any advantages that could be averaged. Like say, average of 4 runs @ 30 mins w/8GB versus 4 runs @ 30mins w/4GB. It's possible the average minimum frame rates would differ to a meaningful degree.

Actually, I was hoping someone could do a few single player tests. I'm guessing since the OP didn't specify SP or MP, either would do.

Though I might have to change the thread title again to add at the end "in BFBC2." Seems to be the only game which unanimously seems to benefit from 8GB according to anecdotal evidence supplied so far.

 

Hehe kinda this. 30FPS is barely playable, and even 28FPS isn't.

 

Yeah, I haven't noticed any change in anything else I've played. I can run some static sequences in BFBC2 and see if there's a difference, but with MP there's a lot more to it than that with all the net code.

Problem with BFBC2 in multiplayer in those two 30 minutes scenarios you could play 2 or 3 completely different maps in that timeframe, unless you find a server that only plays one map which is rare.

Here's an example. I just ran the same map four times for five minutes each with my current config. Different servers, same number of players (30-32), same map, five minutes starting from the beginning. Here's the spread.



You can't even run a reliable trend line through that. Granted more time will be a little better, but not by much.

 

Very true. Though, I'm thinking that perhaps with only 4GB, some of those lows will be spiking even lower. But it will be hard to test in MP since you would have to wait until the conditions were the same (same map, same players, etc.).

 

Right. But I also don't know if its a higher minimum framerate or what it is. I don't know if FPS will tell the whole story. Either way, since it's not easily repeatable and I'm not being paid for it, I will just let it rest, and let people make of it what they want. I don't have the time or patience for it at the moment. I was just trying to support what the OP stated by offering a second subjective vote, which obviously turned into a sparring session of wits.

 

All good and understandable. Though the OP was asking if someone could verify that the claim (found offsite) is indeed true via a test that could demonstrate higher FPS minimums before he drops the coin on extra RAM.

Hopefully he'll come back and specify if he wants to know in single or multiplayer. Single would be easy enough to tell.

Hey I wonder if a video could show the difference?

 

Doubtful. Almost all smoothness and stuttering issues are minimum framerate drops, so what would be most informative is just plotting out what the minimum framerate is, as well as the amount of time it spent below, say, 30FPS (or whatever threshold). Even if you can't get a totally repeatable run in BFBC2, you should be able to run through similar scenarios that cause framerate drops, and if it drops less or less often with 8GB you would have some pretty strong evidence.

 

GAH! It's darn near impossible. Not sure if either one of you have played the game (no offense). Explosions, destructible environment, which vehicles are in the vicinity, etc, all affect framerate. One game you could have a run where you get little of that simultaneously and next run a lot of it. If I could do it repeatably I would, but again, it would be a moot point because it would mean nothing unless you did it a dozen times (I'm not gonna six-sigma this thing) for an extended period, which personally I'm not about to do.

So to make things simple, the answer then is no, more than 4GB will not affect performance in pretty much all cases. I think that answers the OP's question. We have no anecodtal evidence for or against it, but everyone here seems to think no, so I guess we'll run with majority vote.

 

Nawh, don't do that. Hold to your guns. There are other games that are said to have this benefit with 8GB that would be easier to test. Crysis, for one.

Eventually, someone will have the time to run some tests and post the results. You'll be able to say, "see, I told ya." Then posters will argue about the FSB, CPU and/or GPU matched with the RAM.

 

If you so so, then it's easy. Just load up the Crysis benchmark and run through it with 2GB / 4GB / 6GB / 8GB ram, and see the results.

 

Yeah, Crysis should be cake. If I get a chance later I'll do that. Just set it up and let it run. Although personally I'm most interested in BFBC2.

 

you can deduce something from that. you can deduce how fast the ferrari moves through traffic between points A and B. if the traffic changes, it doesn't matter. with enough data, i can make a prediction about the amount of time it will take the ferrari to go from point A to point B.

Anyway, I am the only person here who provided a benchmark. i don't have the hardware on me to run tests but I will in a few weeks when I get back home. And I *shouldn't* need the hardware to request the experiment or hunt for the results.

 

You don't need the hardware to request an experiment. But to blatantly call someone out on something that they haven't done themselves is basically calling someone a liar.

 

I'm *definitely* not calling you a liar. I do think that there is a third variable. I am interested to figure out what it is.

Maybe the game received a patch at about the time you swapped the memory? BC2 has received pretty consistent patches, often affecting performance, and as far as I can gather you haven't been swapping back and forth between 4 GB and 8 GB, it was 4GB -> 8GB, correct?

 

You're right, I was a little harsh. But I'm not crazy (well kinda). No placebo effect because there was no preconceived notion that there would be any improvement.

Feel free to bench because I don't plan to. I never meant to take it this far, I only backed the original post by stating I noticed an improvement, albeit perceived improvement. I wouldn't call it "silky smooth" as the OP stated, only a noticeable improvement from what I was used to. Take it or leave it as you will. I'm done here, leave all messages with my secretary.

 

Uh, there might be a way to get a multiplayer benchmark, such as it is. That's if I'm not completely wrong how the benchmark/recording works. And I know it's not the game in question, but who knows. It might be relevant.

Anyway, I know some Valve games allow recording of gameplay using some in-game recorder. It's different from FRAPs in that it doesn't record to a 2D video file like .avi, but rather records what's happening and replicates it when the user wants to run the demo again.

So, if an experimenter wants to, hit up, say, CS:S or Left 4 Dead 2 (or whatever), do something crazy for thirty minutes, and then end the demo.

Open the demos with the variable RAM, and see what comes up. I guess run a FRAPs benchmark to see minimum-average-max.


Of course, this all depends on my understanding of Valve demo-recording being right.

 

I don't think there's a way to record a gameplay demo (a la Starcraft 2). Plus I think you might get skewed results because live game is different from a playback, even if it is in real-time using the 3D engine. Either way it would have to be designed within the game itself.

 

Oh well. Guess there's no decisive way to tell.

On a related note, I know I will see an improvement when I upgrade my 1215n to 4GB RAM.

 

lol look guys let's not hate. Htwingnut you give me hope that I can tap into more potential of this notebooks power without doing a drastic upgrade with either my cpu or gpu, if it's even possible. I'm already decided that I will purchase more RAM in the future, and I'm crossing my fingers that it will indeed enhance the experience of BFBC2.

Anyway I still couldn't find the benchmark by a reputable source that did the study that indeed shows more RAM holding the framerate of this game steady by increasing the minimum fps. I actually thought you found it, Htwingnut, since your graph was pretty similar in looks.

I just wanted to mention that I'm pleased to have been able to smooth out BFBC2 on my notebook, by installing Nividia 265.98 drivers. Its an official Toshiba driver released by laptopvideo2go, since it's not on the official Toshiba site yet, and my X505 is incompatible with the 260.99 drivers on Nvidia site. So now instead of 20-30 fps, I'm averaging 30-40 fps! And I didn't even spend a dime on extra RAM. Although I do hope someone does do a benchmark to see if 8gb would improve game performance in BFBC2, before I spend more on RAM.

 

You understand that the placebo effect doesn't require you to have the specific conscious notion that you described. It could absolutely occur anyway. And, at this point, I'm starting to think the third variable was a game patch.

@OP I don't just mean to trash your hopes or anything, but the fact is you aren't going to get better performance out of adding more ram. I'm actually just trying to stop you, and possibly others who read this thread, from investing more money into memory when it isn't going to help. That said, for your case, you just need to experiment with your graphics settings to get performance up. You don't need to invest any money- just time.

 

I understand this masterchef, ever since I owned Tandy computers with a 286, 386, and 486 CPUs. Even before the GPU was invented, I always had an issue with low framerates and way back in 1991/1992, I asked the same question. Would extra RAM solve the "choppy" framerates? I bought some RAM, installed it and no cigar. It never helped. BUT, times have changed and technology too, especially software and with new programming techniques, maybe just maybe those articles and studies that I read about Battlefield Bad Company 2, being able to utilize and benefit from more RAM are true. Heck, you can even find articles of Crysis running better with 8gb of RAM. So maybe if this was true in the past, could you be wrong now with modern game software? I mean most older games didn't utilize a quad core to full potential, but BFBC2 does...Maybe it also utilizes 8gb of RAM?

 

If that's the case, the programmers need to be fired for bad coding.

 

I meant to say, "Maybe it could access and utilize more RAM, up to 8gb, if it was available?"...

 

a program should utilize as less RAM as possible , utilizing more RAM doesn't mean it is better.

 

I'm not saying that no game ever will use 8 GB of ram. But- slap a 6 core processor on bad company 2 and it won't help over a quad core. Put in 8 GB of ram and it isn't going to do anything over 4 GB. Now- get a faster processor, or faster memory, or a faster graphics card, and those things can help. It doesn't use the extra cores, and the OS doesn't need it. Same with the memory. The extra resources are just gonna sit there idle. The end.

Go get the memory. I have a feeling you that you will *definitely* notice a change (though no change occurred). Get another hard drive while you are at it - the game will be able to use that just as much as the memory.

/done. I can't help you any more with this. You want to believe it *so* *bad*. So just go believe it. Nothing else really matters I suppose. Please feel free to link to these articles.

1+1 = 2. But, maybe this time it's 3?

Nope. It's 2.

But I heard somewhere it might be 3?

No, I'm definitely positive it's 2.

It was definitely 2 before, but now times have changed and it might be 3.

No. It's still 2.

But I went through the 1+1 calculation and didn't check the result, but it seemed like it might have turned out to be 3, even though I know it doesn't make any sense.

You're right. It doesn't make any sense - it's 2.

But I thought I heard somewhere that maybe it was 3 this time. I heard some article somewhere out in the ether about it.

It really doesn't matter. Unless you can do the calculation and show me that it's 3, or point to the article where they did the experiment correctly and came up with 3, we really need to accept that the result is 2. I even have a nice experiment link showing 1+1=2 over and over and over again earlier in the thread. That's the only experimental source anyone has shown in this thread.

I know you want it to be 3 because 3>2 and it would be really nice and convenient and a simple way to run BC2 at 32xaa on a mobile gtx 460 or whatever, but the truth is, you just need to turn down the aa and accept that.

 

@All - I thought the executable for BFBC2 was 32 bit? Would this not indicate some type of limit to RAM usage?

Thanks,

StevenX

 

Ok, here's three sessions approx 20 minutes each. Two with one RAM config, one with the other. They are all Atacama Desert for a full 500 point conquest map mode from start to finish. I basically just cut off the data for all of them at the shortest time to keep the graphs equal.

If you can tell me it's possible to make any sense relating to appeared smoothness, then have at it. I won't say which config is which to eliminate any "placebo" effect it might have.

FPS 1: Min 23, Max 103, Avg 51.719
FPS 2: Min 21, Max 94, Avg 48.628
FPS 3: Min 19, Max 113, Avg 53.099

 

No, placebo effect is having a preconceived idea about what result you're going to have. If you're none the wiser about a change in hardware configuration, there is no placebo effect. It's all considered the "same" to the test subject. In my case, I installed the RAM the day before and completely forgot about it (yes my life can cause me to forget a lot, I have the memory of a goldfish, about 3 seconds). I didn't play BFBC2 to see the effects of what the memory had, I just played it.

@OP - Yes, sorry about all the bickering. I'd say for the most part you will NOT see any improvement from adding more RAM. I was just commenting that I noticed a differene in BFBC2 after the fact which spurred a debate amongst the skeptics and geeks.

 

This was pointed out earlier. It is a 32 bit .exe

ehem.

This means that, (and this is 100% infallible) BFBC2 can only use ~4GB itself. It can only address 32 x10^9 memory locations, the word length used to access those locations (32 bit), combined with hexadecimal addressing, means that there is not enough permutations of hexadecimal to address more than that. It' like numbering an index of people:

You have 10 digits to work with (0 through 9), and can only have a single digit number for each index (this is your word length of 1)...this means you can only index 10 people (0 through 9). The same works with memory, you run out of "names" for each memory location after 4GB (with 32 bit word length)...

IF BFBC2 IS RUNNING FASTER ON 8GB, it is because the people playing it have a ton of crap running in the background slowing things down with 4GB, and now that they have 8GB, the game finally has access to the full 4GB it didn't before, or there is a memory leak in BC2... End of story. Done. Maths FTW.

I don't mean disrespect to HTwingnut, or the OP, but its math, the conclusion is inescapable.

 

Thanks HT and Classic (but I hate math!) J/K

So the bottom line is that really not much difference between 4 and 8...

Could we take this from another angle... Is their anything within Windows 7 that detects 8gb ram and decides to put more of the OS page file into mem rather than HD? Like the FSUTIL BEHAVIOR command (which I know most people use for NTFS drive config).

Fsutil behavior

Again just hitting this from all angles here.

BW, StevenX

 

Just installed 4gb more in my desktop, since my bro got a newer pc and gave me his old hardware.

Luckily he was running same ram that i do (XMS2 Corsair, DDR2 800mhz) and based on a few tests i made, theres not a single improvement in fps or performance, nor in general "smoothness".

Every benchmark i did scores the same althought i had to lower a bit the OC otherwise it was unstable (i was running @ 4,25ghz, now 4,05. I doubt that 200mhz difference really makes worth a notice).

So my 2cent: more ram doesnt improve gaming performance. Maybe it does regarding general OS management and such, but not in games.
Definitely its a placebo effect, but math isnt an opinion, so if my benchmarks scores the same points with 8gb instead than 4 then its a rock solid proof.

 

My guess is the red is the 4GB config.

If I'm right, I should win something, no?

C'mon. Gimme-gimme.

 

lol ? this is so funny...Why can't notebookreview.com just do a benchmark on this subject? If they did then case closed. I mean all the other guys are just assuming without even having tried it, and the ones who said they actually tried could be 12 year old obsessive liars. IDK, but the guys that actually write the reports and articles on this site would have more credibility to me anyways, and if they falsified a report then well it's all over the internet for other sites to ridicule. You see what I'm sayin?

 

The whole problem is that you can benchmark for video card performance using some of the in-game sequences that are scripted. But that doesn't really tell the story for multiplayer when there's so much dynamically going on, not to mention execution of net code in addition to everything else.

 

Well. I'm out. Tried to help.

 

No, thanks for your insight masterchef341. You are 100% on track with all your comments.

Just for the record, 2.0 was right, FPS 2 is the run with 4GB. Not that a few runs means much, but the 4GB FPS range was a lot less than with 8GB for some reason, but both 8GB runs were about 4FPS faster than 4GB. It would probably need at least a half dozen runs or more with each running the same map on same server with same number of players to make it reasonably respectable.

And for the record I *AM* a 12 year old obsessive liar...

 

What i dont understand is two people with lots of rep and experience can disagree. I am fine with that. Then someone that's relatively new to the forums chimes in, states someone could be a 12 year old and that notebook review forums will get it right? I am just confused on this one.

HT - Based on what you found, did the 8gb ram improve fps or no?

Best Wishes and Thanks for the effort.

StevenX

 

Well, based on avg framerate, yes. But just too little data to be reliable. I'm not about to go play a dozen games on the same map and same server any time soon though.

 

@ Htwingnut, nah I wasn't referring to you when I mentioned 12 yr old obsessive liar, I was just generalizing and besides it was a joke. Oh also I think benchmarking through single player would be ideal since we can reference at what points there's slow down and loss of frames with 4gb and then compare the same points and framerate count with 8gb, but it's up to you if you want to do it, since I know it does take time on your part.

@StevenXowens792 I don't think this forum is some kind of "elite club" for senior members and besides I'm not new here. I joined a mere 6 months after you, and I was reading this site and forums even before then. Also rep don't mean nothing on here. I won't mention names, but there's some members with 50+ rep that give bogus info, even laughable, and there's some who are world class hackers that have well below 10 rep.

Anyway if anyone would want to do a formal benchmark with graphs, video, or anything to prove the results, I'd appreciate it.

 

oh I just found a credible source regarding the issue does BFBC2 uses more than 4gb of RAM? Well coming straight from the horses mouth, Anders Gyllenberg, the developer of the PC version of BFBC2, when he was interviewed by pcgameshardware.com

PCGH: Do you have an explanation why games fully utilizing 64 bit architecture are so rare? If Crysis 2 is 64 bit ready, what benefit players can expect? If no, what where the reasons to make the game 32 Bit only?

Anders Gyllenberg: The main reason to go 64-bit is if the application needs to access more than 2GB of memory at the same time. Bad Company needs a bit more than 1GB of memory for the game itself, and therefore would not benefit from 64-bit.

Well at least I get to save my money and even better I get to play BFBC2 at maxed graphics settings with everything set to high, AA 32x, AF 16x, and HBAO on, at 30-50 fps because of the newly released 265.89 Toshiba/Nvidia drivers I just installed . That's a huge jump from the previous 20-30fps! and my first graphics driver update. Imagine the next driver release, and this one wasn't even official, I got it from laptopvideo to go, since the official Toshiba site didn't post it yet. This Nvidia GTX 460m is a BEAST! Happy New Year!

 

oh wait, I spoke to soon. Here's another argument for extra RAM and BFBC2.

XxUserxX makes very good arguments for RAM holding your framerates more steady and smooth.

GameSpot Forums - Battlefield: Bad Company 2 - Does 64-bit give you more performance over 32-bit?

I just might have to spend money again...

 

You know BC2 is 32bit right ?

Anyway , it's your money .. I say go get those 8GB and may be you will have enough time to prove all of us wrong. Can't wait.