BFBC2: considering upgrading

[DualOwn]

Exactly!



This wasn't directed at me, but I want to apply it to my situation as I think you're implying that Porter's Q6600 is a bottleneck. If my low GPU % was due to my "slow" Phenom II, can you explain why the 460 performed 22% worse than the 4850 with all things being equal?

because the gpu usage on 4850 was 95-100% and that cpu is a perfect power match for 4850.

and dont forget, you cant compare two different gpu's like that.

the more state of the art the gpu is the more state of the art cpu is needed.

you cant pair up old cpu tech with new gpu tech and expect everything to click.

 

[tlbig10]

because the gpu usage on 4850 was 95-100% and that cpu is a perfect power match for 4850.

and dont forget, you cant compare two different gpu's like that.

the more state of the art the gpu is the more state of the art cpu is needed.

you cant pair up old cpu tech with new gpu design and expect everything to click in.

Don't make me use my Carl Winslow gif again. There are so many things wrong with this post, I'm going to have to break it down piece by piece.

you cant compare two different gpu's like that

Really? You really want to stand by that statement? What's the point of every single tech site/forum if we can't compare GPU's on equal ground? If one card, which is universally accepted as being 2 or 3X faster than another, somehow perform much much WORSE than an old video card, not make you start asking a lot of questions? Isn't comparing the sole purpose of upgrading? You buy something new, you expect it to be better. If it's worse, something is WTF.

the more state of the art the gpu is the more state of the art cpu is needed.

Ok.... yes that is a true statement to some extent, but we're not talking about Pentium 2's vs. Sandy Bridge. Do you think I stuck a 460 in a P4 2.28 Ghz PC and wonder what the problem is?

you cant pair up old cpu tech with new gpu design and expect everything to click in

I'm starting to think you work for Apple. Isn't this the beauty of building your own computer - the ability to upgrade components as needed rather than the whole thing? You have a fair amount of certainty that replacing an older PCI-e video card with a newer, more powerful PCI-e card, will result in better performance and very few, if any, compatibility issues (assuming we're all on board with Windows 7).

This thread has gone way off track. I talk half the blame, but by golly we need to correct all this misinformation!

 

[DualOwn]

460 is not to blame here pal, bc2 is still a console port, its just that AMD drivers are better for BC2 than Nvidia drivers.

devs coded BC2 engine based on ATI/AMD gpu, xenos from xbox360.

but the latest drivers fixed BC2 for me, im with gtx480

 

[tlbig10]

 

[DualOwn]

lol, i dont know what to tell you.

 

[Jodiuh]

To those wondering why tlbig10's 460 experience blew more chunks than most...

1. Phenom II
2. 3 cores

Tlbig10's taco shop has been working sloppy lately because they didn't have enough staff to handle the load. So tlbig10 put up a "help wanted" sign. While looking through the potential candidates, he noticed one in particular that stood out. Mr. Fermi looked great on paper and his salary requirements were low. Mr. Fermi got the call a few minutes later, accepted the offer, and went to the bar to celebrate. He woke up the next day and couldn't find his keys. This really bugged him because he didn't care for cab drivers. He got to work and despite his best efforts, couldn't keep up. He even managed to drag the other employees down with him. This displeased tlbig10 greatly as he'd never had to deal with a bad company before.

 

[DualOwn]

lol

 

[tlbig10]

To those wondering why tlbig10's 460 experience blew more chunks than most...

1. Phenom II
2. 3 cores

Tlbig10's taco shop has been working sloppy lately because they didn't have enough staff to handle the load. So tlbig10 put up a "help wanted" sign. While looking through the potential candidates, he noticed one in particular that stood out. Mr. Fermi looked great on paper and his salary requirements were low. Mr. Fermi got the call a few minutes later, accepted the offer, and went to the bar to celebrate. He woke up the next day and couldn't find his keys. This really bugged him because he didn't care for cab drivers. He got to work and despite his best efforts, couldn't keep up. He even managed to drag the other employees down with him. This displeased tlbig10 greatly as he'd never had to deal with a bad company before.

never had to deal with a bad company before

 

[samuelmorris]

To those wondering why tlbig10's 460 experience blew more chunks than most...

1. Phenom II
2. 3 cores

Tlbig10's taco shop has been working sloppy lately because they didn't have enough staff to handle the load. So tlbig10 put up a "help wanted" sign. While looking through the potential candidates, he noticed one in particular that stood out. Mr. Fermi looked great on paper and his salary requirements were low. Mr. Fermi got the call a few minutes later, accepted the offer, and went to the bar to celebrate. He woke up the next day and couldn't find his keys. This really bugged him because he didn't care for cab drivers. He got to work and despite his best efforts, couldn't keep up. He even managed to drag the other employees down with him. This displeased tlbig10 greatly as he'd never had to deal with a bad company before.

I lol'ed hard.
Brilliant, saving that

 

[Porter_]

FYI, i just installed the new 266.35 beta drivers and i get no change in performance. still 40% gpu usage while playing BC2

 

[DualOwn]

FYI, i just installed the new 266.35 beta drivers and i get no change in performance. still 40% gpu usage while playing BC2

hang in there, 2500k is almost out.

 

[Porter_]

amen

 

[nobody_here]

It'd be great if you could replicate my test just to confirm. 1920x1200, medium detail, 1xAA, 16xAF, HBAO off and Vsync off. Join a populated 32/32 person Cold War map, run fraps for 5 minutes, and take a screen shot of your GPU usage when you're done.

If you'd like, you can send me the fraps spreadsheet of the FPS and I can add that to my line graph.

ok, not apples to apples but maybe helpful, remember i am at 1440x900 resolution, but it works out well

1440x900
Highest Detail
16xCSAA
16xAF
HBAO ON
VSync OFF

i dont have fraps, but i took a screenshot of task manager and MSI Afterburner so you could see what was happening on mine

as you can see not a huge system memory hog, and it only needs around 600Mb of video memory....four cores fairly busy.....60% on three, last one 90%....GPU is as i remembered, fairly taxed out....but then again these settings run butter smooth, flawless on my system...i imagine i might could push better quality AA..?? i dunno, it's pretty nice looking already



Uploaded with ImageShack.us

 

[Jodiuh]

That's some garbage usage right there. Butter smooth my butt. At those settings I bet you're hitting a minimum under 40, maybe even 30. Go get fraps, its free.

Edit: It should look like this.

If you don't mind, would you grab fraps and run a benchmark with 4xaa, 16xas, high detail, and hbao off? Also, snap a usage shot if you can. Make sure to set nvidia driver settings to high quality and lod bias to clamp.

 

[nobody_here]

That's some garbage usage right there. Butter smooth my butt. At those settings I bet you're hitting a minimum under 40, maybe even 30. Go get fraps, its free.

Edit: It should look like this.

If you don't mind, would you grab fraps and run a benchmark with 4xaa, 16xas, high detail, and hbao off? Also, snap a usage shot if you can. Make sure to set nvidia driver settings to high quality and lod bias to clamp.

Um...yeah...its smooth as silk...gpu utilization does not equal quality of gameplay.......that's not a framerate graph..that's gpu usage

So I can get fraps and run it for a few minutes again....what do you mean by benchmark it? Is there a prepackaged benchmark in BFBC2?

 

[Jodiuh]

It does for those of us that can see it.

Given your recommendation of a 2 yr old midrange over today's midrange hotness, I think maybe you might be a bit confused by some of this. Its simple really. A gpu performing less than 98/99% will vary in framerate more than normal. With his settings that means dancing around the 30/40 FPS line. That is not smooth.

When I first read about this issue after experiencing myself...well, I was bummed. By I did not attempt to defend my Q9550. To have done so would have created confusion. That helps no one. I can only hope those reading after can separate the chaff from the wheat.

 

[nobody_here]

ok then, i figured it out, the benchmark feature in BFBC2...got it..here you go, 1440x900 all highest settings on details, 4xAA, 16xAF, HBAO off, set driver to my preference "Quality" and LOD was already set to Clamp, made sure though

Min,Max,Avg:

Frames, Time (ms), Min, Max, Avg
4908, 60000, 52, 183, 81.800



Uploaded with ImageShack.us

i was a few seconds late nabbing the screen shot but the only thing you would see different is the same thing you saw above in the task manager window with three cores at 60-75% and one at around 90%

there was some weird...server glitch..??? i dunno...lag maybe....near the end of the run....it definitely wasn't a drop in performance...as in a decline which i have seen before on an older system...it was all of a sudden and then it was gone, thats why i theorize it was a server side thing

 

[Jodiuh]

There is no benchmark feature in bc2.

Can you see difference in the two usage graphs? One is straight, the others wavy.

 

[nobody_here]

no they are both relatively the same in regards to GPU usage except the second one has that small glitch near the end of the run like i mentioned.....but otherwise the average usage seems to be the same....bobbling between 75 % and 90 % it looks like...averaging maybe 85%

at any rate, that showing up as it is or not....the game is running quite well on this end, couldn't really imagine much more..... im very happy...hopefully he figures out whats ailing him

 

[Jodiuh]

Both of the graphs using an AMD processor that you have posted are wavy. The graph using an i5 760 that I posted is a straight line.

 

[nobody_here]

the cpu being pegged out isn't a good thing though, means the CPU is having to work it's ass off to feed the GPU.....which means the GPU could likely be bottlenecked by the CPU in your i5 example whereas in my example this is not true, the CPU is working moderately but the GPU is working around 85% capacity...it's not waiting on anything...it needs more to do and it's not because the CPU cant feed it more to do, there's just not enough work for them to do....looks like time to upgrade my monitor to a larger one which is in the works in the coming months

in my case the card is capable of a bit more than i am putting it through in regards to BFBC2, i could be running more detail or even a high resolution to get the goody out of it, 1680x1050 is probably just about right....might go 1920x1080 but may have to run 2xAA or less AF, etc.....still, rock solid piece of hardware i tell ya this 460 is, for $200? extremely happy considering i havent even touched the sliders in Afterburner yet.....this is on a totally stock system

i can however peg out this card in this system. i can make it happen for instance on Crysis....i will have only one CPU core doing 99% of the work (because Cryengine2 sucks like that) but the GPU is pegged to the max while the CPU has room to stretch. thats at 1440x900, ALL very high details, 4xAA, motion blur on, etc....

so it's not a system limitation at the core, it's a matter of BFBC2 not using my hardware to it's fullest potential, although i must say, i am much happier with it's SMP vs. Cryengine

 

[Jodiuh]

I give up.

 

[nobody_here]

why...?? i didn't know you were trying anything in particular...

 

[ObiDamnKenobi]

ok, I think we can say slightly off topic now.. Isn't there a way to record a FPS graph, so you can settle this?

Ordered a 6950 to replace my gtx 260 for BC2 yesterday. But last night I tried it at 1440x900, and it didn't look that bad and stayed ~50 fps, decently playable in MP. Wondering if I was a little to itchy to upgrade. Ahh buyers remorse already!

 

[Jodiuh]

Are you saying there wasn't much difference? What CPU? I wondered about the new Radeons in this game, especially the 6950. It looks speedy with decent power requirements.

For a graph tutorial, just YouTube excel % distribution.

New nvidia driver last night, 266.33 beta iirc? Mentions a bump in bc2 performance.

 

[Porter_]

New nvidia driver last night, 266.33 beta iirc? Mentions a bump in bc2 performance.

i saw none

 

[SHiZNiLTi]

ok, I think we can say slightly off topic now.. Isn't there a way to record a FPS graph, so you can settle this?

Ordered a 6950 to replace my gtx 260 for BC2 yesterday. But last night I tried it at 1440x900, and it didn't look that bad and stayed ~50 fps, decently playable in MP. Wondering if I was a little to itchy to upgrade. Ahh buyers remorse already!

You made a good decision getting that card, it's a huge upgrade over the gtx260. Your also running a i7-920 @ 3.4 GHz and 24" (1920 x 1200) monitor. The 6950 does very good at higher resolutions like that and your not bottle-necking the card in any way with such a good cpu.

My upgrade scenario is different though, I'm running a Core2Quad Q9650 @ 3.8gh and my monitor resolution is 1600x1200.

The ati5870 @ 950/1250 actually gave me better fps then ati6950 in my rig. It was within about 10fps, but I bet the results would have changed if I tested at a higher resolution like your running.

Also the 5870 is only $200 right now!, I went ahead and picked one up on newegg.

I still have a gtx570(905/2250) in my rig now that I need to return to Fry's Electronics, I've been testing this gtx570 for about 20 days now. My PC room is consists of 2 ft table side by side, I sit right next to my roommate that has the same exact rig as me except that he is running his C2Q @ 3.6ghz and running a ati5870, he is def getting better fps in BC2, very easy to compare side by side since I can see both monitors at the same time.

 

[SHiZNiLTi]

We are very aware of that, thank you. The issue still exists on high end overclocked C2Qs where GTX 460s will only see about half GPU usage, yet when simply changing platforms (regardless of clock speed!) GTX 460 usage jumps to 100%.


pokevitek: What CPU and setup are you running? My C2Q is fairly balanced accross cores between 60 and 80%.

Can anyone confirm if this happens with GTX 570s and C2Q?

I'll check out my CPU/GPU usage in BC2 for you tonight I have a gtx 570 in my rig right now and a Q9650 @ 3.8ghz.

I also tested out (2) gtx 460 768mb cards in my C2Q rig.

With SLI disable only using 1 card running BC2 @ 1600x1200 highest settings DX11 16xAA(might have been 16xQAA or 32x, not sure) and 16xAF I was at pretty high GPU usage in a full server around 75-90%. When I enabled SLI I only got a 15fps increase, but each card was only running at 50-60% GPU usage.

The single 5870 still out performed (2) gtx 460's in SLI... I was hoping for better results from the gtx460's in my rig, but was let down... it's weird because all the benchmarks done online show (2) gtx460's SLI'd performing awesome and doing better then a 5870, but in my case my Q9650 really bottlenecked them.

Test shown using a Intel Core i7 920(ES) @ 4.0Ghz...

 

[nobody_here]

ok, I think we can say slightly off topic now.. Isn't there a way to record a FPS graph, so you can settle this?

Ordered a 6950 to replace my gtx 260 for BC2 yesterday. But last night I tried it at 1440x900, and it didn't look that bad and stayed ~50 fps, decently playable in MP. Wondering if I was a little to itchy to upgrade. Ahh buyers remorse already!

i did...and ps im not having a problem, i was simply trying to provide data for the OP to compare to when talking about 460GTX performance and a quad core CPU, both of which i have

http://hardforum.com/showpost.php?p=1036657061&postcount=177

the Fraps dump of framerate min/max/average is in that post

Min,Max,Avg:

Frames, Time (ms), Min, Max, Avg
4908, 60000, 52, 183, 81.800

52 minimum, 183 highest, 81.800 average

 

[nobody_here]

The single 5870 still out performed (2) gtx 460's in SLI... I was hoping for bett results from the gtx460's in my rig, but was let down... it's weird because all the benchmarks done online show (2) gtx460's SLI'd performing awesome and doing better then a 5870, but in my case my Q9650 really bottlenecked them.

please be sure you are comparing 460 768Mb cards to 460 768Mb cards and not 768's to 1Gb's because there's a difference other than 256Mb of ram, sometimes more than others, depends on the game

 

[ObiDamnKenobi]

You made a good decision getting that card, it's a huge upgrade over the gtx260. Your also running a i7-920 @ 3.4 GHz and 24" (1920 x 1200) monitor. The 6950 does very good at higher resolutions like that and your not bottle-necking the card in any way with such a good cpu.

My upgrade scenario is different though, I'm running a Core2Quad Q9650 @ 3.8gh and my monitor resolution is 1600x1200.

The ati5870 @ 950/1250 actually gave me better fps then ati6950 in my rig. It was within about 10fps, but I bet the results would have changed if I tested at a higher resolution like your running.

Also the 5870 is only $200 right now!, I went ahead and picked one up on newegg.

I still have a gtx570(905/1125) in my rig now that I need to return to Fry's Electronics, I've been testing this gtx570 for about 20 days now. My PC room is consists of 2 ft table side by side, I sit right next to my roommate that has the same exact rig as me except that he is running his C2Q @ 3.6ghz and running a ati5870, he is def getting better fps in BC2, very easy to compare side by side since I can see both monitors at the same time.

thanks for easing my remorse a bit The 6950 is sitting in the box next to me at work now, can't wait to get home and stuff it in my rig!! It will definitely be nice to go above "netbook" resolution and not get 8 fps when I turn around..

The results you got do seem strange, especially that the 570 is slower than a 5870 The 6950 and 5870 were pretty close in the benchmarks I read so that's not as surprising. Only explanation I can think of is what you said; that the CPU is holding it back somehow. But I would think the C2Q at 3.8 GHz would still perform pretty well?

 

[SHiZNiLTi]

thanks for easing my remorse a bit The 6950 is sitting in the box next to me at work now, can't wait to get home and stuff it in my rig!! It will definitely be nice to go above "netbook" resolution and not get 8 fps when I turn around..

The results you got do seem strange, especially that the 570 is slower than a 5870 The 6950 and 5870 were pretty close in the benchmarks I read so that's not as surprising. Only explanation I can think of is what you said; that the CPU is holding it back somehow. But I would think the C2Q at 3.8 GHz would still perform pretty well?

When you get home from work don't remove your gtx260 right away, gather some results 1st... crank your resolution up and run highest settings, HBAO off with at least 4xAA and 16xAF go to an empty and full server in Valparaiso and take a SS while sitting in the attacking teams blackhawk while in 3rd person. I think I still have the SS from that exact same spot when I tested out the 6950 in my C2Q setup. Will be interesting to see how they compare with your i7 920 CPU.

Ya I know for sure it's my C2Q that was bottle-necking these higher end cards I tested like the gtx570 and $500 ati5970, the 5970 was a hella bottle-necked, I got horrible results with it and was slower then the 5870... I even used my roommates PC to verify results, he has pretty much the same rig as me, same 850 Corsair PS, but uses a Q9450 @ 3.6ghz on a XFX 750i mobo and it yielded the same poor results.

Another interesting BC2 graphics detail I found out is that when you run the game in DX11 it forces shadows to be on high, shadows cause the biggest FPS drop.
DX11 shadows set to low is exactly equal in to how DX10 shadows on high looks like.
So if you need more fps you might want to force DX10 with shadows on medium or low, you will get 15-20+fps boost!

 

[Jodiuh]

Get ur ass signed up for the LAN! Tlbig10 and I expect you to be there!

 

[SHiZNiLTi]

Get ur ass signed up for the LAN! Tlbig10 and I expect you to be there!

lol, sorry have been procrastinating sign up, I'll get myself and my roommate signed up now... I'm not looking forward to loading up our massive 90lb Trinitron CRT's and PC's into the back of my lowered lexus, better not see any speed bumps on the way up, gonna bottom out big time lol

I still have 11 days till I have to return my gtx570 to fry's and I should have that $200 5870 delivered from newegg to my house on Friday just before the LAN so we can do some testing if you want with these cards. I'd like to see what the gtx570 can do in your rig, your running a i5 right?

 

[tlbig10]

I still have a gtx570(905/2250) in my rig now that I need to return to Fry's Electronics, I've been testing this gtx570 for about 20 days now.

If that 5870 gets here before the lan, bring both with you. We could put one of them in my system so the CPU can feel the squeeze!

 

[Az Syndicate]

These guys will jump you and steal you shiz. Do Not LAN.

Unless you bring a sheep. Then it is all good.

 

[Jodiuh]

I'm not looking forward to loading up our massive 90lb Trinitron CRT's and PC's into the back of my lowered lexus..

Strap em to the hood for some ballin' headlights.

I'd like to see what the gtx570 can do in your rig...

Probably not a whole lot more than this. 570 can't compete w/ closeouts.

...your running a i5 right?

Yes, unfortunately it's a dirty 760 and needs around 1.33V for 4.

Unless you bring a sheep. Then it is all good.

C'mon, it was only the one time and no one else saw the pictures.

 

[SHiZNiLTi]

These guys will jump you and steal you shiz. Do Not LAN.

Unless you bring a sheep. Then it is all good.

LOL! Let them try to jump me, I'm 6'8" 300lbs, free piggy back rides is all they will get

Probably not a whole lot more than this. 570 can't compete w/ closeouts.

Dang, you gonna get that card at Fry's Jod? That's a pretty damn good deal, but the 5870 still performed better then the 470 I tested out. I assume they fixed the fan noise issue with this version that you linked above. The PNY I tested out was freaking LOUD!!!

 

[Italiano86]

I don't know how alot of you guys are testing your frames, but anyone running a fermi card with anything less than an i series chip while claiming to have high minimum frames makes no sense.

My rig runs the game at pretty much medium settings with aniso and aa set at 1 and I get down to the 20's on panama canal when theres smoke or a lot of action going on. My 24" runs at 1920x1200.

Maps with less foilage run great, but panama will show you your system sucks if you run fermi with a 2+ year older chip.

I am probably going to snag a sandy bridge in the next week or two to finally be able to play this game properly.

 

[Jodiuh]

i saw none

266.35 beta

Dang, you gonna get that card at Fry's Jod? That's a pretty damn good deal, but the 5870 still performed better then the 470 I tested out. I assume they fixed the fan noise issue with this version that you linked above. The PNY I tested out was freaking LOUD!!!

Just put it in now. It's below the noise floor of a Scythe SFLEX E 1200 RPM @ idle. But the reference 470 wasn't too loud @ idle either IIRC. To be honest, I could hear the ref 470 while playing BC2 provided I took my headphones off. But in no way would I call it loud. Now furmark unlock power draw's a whole nother story...sound like aeroplane.

...medium settings with aniso and aa set at 1...20's...1920x1200...

W/ and i quad, you'd be looking at 60. FWIW, high settings 4xAA on i5 760/460 oced gets 60 min. I'm hoping for some HBAO or maybe a little TSAA w/ the 470.

...sandy bridge...play...properly...

If you live near a Microcenter, I hear 2500K will go for $150 day one. Paired up w/ an Asus P8P67 Pro would make Italiano happy.