ati 7970 and disappointing battlfield 3 multiplayer performance

I

icanseeu1985

n00b
Joined
Jan 20, 2012
Messages
49
First of all, I get the same minimum fps as I did with the 5870, on the same spots/areas on the expansion maps (Such as gulf of omen or shariq peninsula). The only difference is, I can use 1920 by 1200 resolution now. On the 5870, I was using 1440 something resolution. I got a 1080p lcd (1920 by 1200 to be exact). Second of all, if the average fps is high, you can feel the slow down effect @ 38-45 fps (approx). Third of all, I reformatted and did a clean install of windows 7 (after installing the 7970). Forth of all, I tried the 1/9 drivers, and also the ones that came out today (no difference really). Fifth of all, the default voltage of this sapphire 7970, seems to be 1.7 something (using gpuz, while battlefield 3 is in window mode).





2600k sandy @ default (overclocking to 4.5 ghz did not affect the fps in battlefield 3)
ati 7970 @ default (currently at 1050/1575. Maxing the 3 sliders, give me a 6fps boast to the minimum fps. Right now, I am getting a 4-5 fps boast to the minimum fps, because of this oc)
8 gigs muskin ram
asus p8p67 pro rev 3.0 b3
windows 7 64 bit, just reformatted, nothing running in the background.

























NOW FOR MORE DESCRIPTIVE STUFF ON SETTINGS & FALSE ADVERTISEMENT:


It's funny how you look at these 2 graphs:
http://www.tomshardware.com/reviews/...cn,3104-7.html
This was before I bought the 7970. WOW, LOOK AT THAT FPS. 5760X1080, all on ultra, no aa = 32 min fps. SO that must mean, that at 1920 by 1200 res, I would not go below 49 fps (at the most extreme multiplayer situations in battlefield 3). Not the case at all. Why are all sites picking the most least intensive part of the game, then sayings how great the 7970 performs in battlefield 3. All of this crap, seems like a scam to me. It is like ATI wanted the review sites; to only benchmark certain parts of the game (PAID OFF BY ATI). I honestly feel cheated.


WHY ARE THERE NO REVIEWS OF THE MULTIPLAYER PORTION OF BATTLEFIELD 3 (with 64 players, gulf of omen/shariq peninsula/other graphically intensive maps)?



Judging by literally ALL THE ati 7970 reviews/graphs out on the internet; I should not be getting a minimum of 29 fps, on some of the expansion multi-player maps, @ 1920 by 1200, 4xaa, all ultra, hbao on, motion on, & default fov.
Judging by literally ALL THE ati 7970 reviews/graphs out on the internet; I should not be getting a minimum of 39 fps, on some of the expansion multi-player maps, @ 1920 by 1200, no aa, all high, hbao on, motion on, & default fov. By the way, if the average fps is allot higher (which is the case most of the time); you definitely feel the slow down at low 40s-39 fps.



If a person would read all the reviews out there, before buying this card, this person is literally being misinformed. Some of the stuff is being hidden. Not all the information can be seen. Kepler will destroy this card, watch it have a way better minimum fps. I am not a fan of nvidia or amd. I am a fan of whatever card does what review sites portray them to do.
 
Did I mention, that people are telling me (on the guru forums, WITH PICS); that the gtx 580 is giving them a higher min fps, on the same spots that I tested. They are also using the same graphical settings that I used. For example: shariq peninsuala, on the roof, where the helicopter spawns, & looking at pretty much the entire map. The graphical settings used were; 1920 by 1200, 4x aa, all on ultra, hba on, motion on, & default fov. The minimum fps was 45 on the gtx 580. I get 29 min fps, on that same spot, while using the same graphical settings.
 
Using ultra settings on a single 7970 with 4x MSAA. I can tell you that a single GTX 580 overclocked (I had MSI lightnings oc'ed to 950 mhz/1200 memory) cannot run those settings at 60 fps at 1920x1080 on 64 player maps. In fact, with ultra settings at 1080p I would often dip into the 20s a LOT in multiplayer.

Fact of the matter is tons of websites have reviews minimum/average bf3 framerates on the 580 and 7970, and the 7970 is faster but you can't just throw ultra settings at every single game you have with a single gpu. You can't do that with crysis 2, metro 2033, or bf3 mp.

Those settings really need sli or crossfire. Lower your settings, jesus christ. Or return your damn card.
 
xoleras said:
Using ultra settings on a single 7970 with 4x MSAA. I can tell you that not even a single GTX 580 overclocked (I had MSI lightnings oc'ed to 950 mhz/1200 memory) and it could not run those settings at 60 fps at 1920x1080. In fact, with ultra settings at 1080p I would often dip into the 20s a LOT in multiplayer.

Those settings really need sli or crossfire. Lower your settings, jesus christ.
Click to expand...

DID YOU READ THIS:

"Judging by literally ALL THE ati 7970 reviews/graphs out on the internet; I should not be getting a minimum of 39 fps, on some of the expansion multi-player maps, @ 1920 by 1200, no aa, all high, hbao on, motion on, & default fov. By the way, if the average fps is allot higher (which is the case most of the time); you definitely feel the slow down at low 40s-39 fps."





On high. ON FUCKING HIGH SETTINGS. ZERO AA. LIKE COMMON? 600 dollars for this crap? Same exact min fps as I got on the 5870 (but I was obviously on a lower resolution with this card). IF there are people out there, who are willing to properly read everything I wrote, & give in their input; that would be fucking awesome.
 
Just buy a GTX 580 if you think it will be faster. Don't know what else to tell you.

I've used and own both, and I know what your outcome will be, but please indulge.
 
I have a pretty good idea of what minimum framerates are like on both cards, but I play at 2560x1440.

1325889231KTNbsOX8Vr_3_2.gif


you will see in this graph that the minimum framerates are proportional to each other on both cards, and the framerate advantage remains the same throughout the entire graph (roughly).

Do you see how the peaks and valleys are shaped the same on both cards? Also, the minimum framerates here are affected by the menu screen which is capped at 35fps.

You just have to lower some of your settings. Start with 4x MSAA. Even a heavily overclocked 7970 can't run that test at 60 fps fluid, you need xfire for that. However it is still far faster than the gtx 580.
 
xoleras said:
I have a pretty good idea of what minimum framerates are like on both cards, but I play at 2560x1440.

You just have to lower some of your settings. Start with 4x MSAA.
Click to expand...

Are you reading what I am typing? Cause it does not look like it.

First go here and read: http://hardforum.com/showthread.php?p=1038288723&posted=1#post1038288723




Then tell me if my performance/setting/min fps below is "normal" for this card:

"Judging by literally ALL THE ati 7970 reviews/graphs out on the internet; I should not be getting a minimum of 39 fps, on some of the expansion multi-player maps, @ 1920 by 1200, no aa, all high, hbao on, motion on, & default fov. By the way, if the average fps is allot higher (which is the case most of the time); you definitely feel the slow down at low 40s-39 fps."
 
Dude I'm running 3 7970s in tri-fire and with every setting massed in Bf3 on a 64 player server I'll dip down into the high 70s low 80s then back upto 120 to 180 fps, if I go down to 1 card I get the same results you get. I'm running a 2600k @ 4600 MHz and 8 gigs of ram.
 
Overclock your CPU. It makes a marked diffference on multiplayer (as opposed to single-player).
 
IdiotInCharge said:
This. Same with Bad Company 2 before it.
Click to expand...

Also, turn off HBAO (all the way off, not just to SSAO). I was running 1 OCed 7970 at 2560x1600 with mostly Ultra, no MSAA and HBAO turned off, and it ran fine on every multiplayer map except Shaqi. You should be golden 1920. HBAO is a showstopper, it was even on 5870CF at 1920.

If that still doesn't do enough for you, back down on the Effects setting. The order of "bang for buck" when trading detail for FPS in BF3 is 1st MSAA, 2nd AO, 3rd Effects. At least with the 5870 and 7970s I've tried.

As far as you feeling misled or misinformed.... well you clearly didn't do enough research. I almost ran out and ditched my 5870s for 6970s when the official benchmark reviews came out for BF3, but after some digging, and finding some more thorough reviews that also compared 5870s, I realized that multiplayer is a whole different ballgame. And no-one got payed off to use single player, every half decent site's review I read clearly stated that they used single player to benchmark because it was more precise, easier to reproduce results from trial to trial, and that their numbers should only be used to scale results for MP, not directly compare.

Seriously though, OC your card. You'd be amazed at what a 20% OC can do to your minimum framerate (assuming you've got a decent CPU in there).
 
Forceman said:
Overclock your CPU. It makes a marked diffference on multiplayer (as opposed to single-player).
Click to expand...

It makes a difference alright, as in ZERO difference. I tried it. I even tried no ht and 4.5ghz. NO difference in fps @ 1920 by 1200. GO look at the cpu tested reviews, using the 7970; google it. No difference at 5ghz 2600k vs default clock (1 fps difference on some resolutions).
 
Actually with the way you came onto the other thread on your first day here, it is easier to think of you as a troll than the other way around.


Stop for a second, look calmly around, and see how you jumped on the other thread calling the OP and someone else trolls, on your very first day... first impressions mean a lot, remember that.
 
It's your CPU.

A 2600k at stock will bottleneck a 7970 in some situations, such as in a 64-player game in Battlefield 3. Overclock it to 4.5+ ghz, and you'll see a marked difference.

CPU_01.png
 
icanseeu1985 said:
It makes a difference alright, as in ZERO difference. I tried it. I even tried no ht and 4.5ghz. NO difference in fps @ 1920 by 1200. GO look at the cpu tested reviews, using the 7970; google it. No difference at 5ghz 2600k vs default clock (1 fps difference on some resolutions).
Click to expand...

That's because all the reviews are done in single player, as you mentioned. It absolutely makes a difference in multiplayer (as has been demonstrated by several people here, and a few reviews that tried testing it); if it didn't for you it can only be because your graphics settings are so high that you are bogging down the graphics card.
 
(1) Absolutely NO FLAMING, NAME CALLING OR PERSONAL ATTACKS. Mutual respect and civilized conversation is the required norm.
 
xoleras said:
Yeah, CPU speed should make a tremendous difference in bf3 multiplayer.
Click to expand...

jbz7890 said:
It's your CPU.

A 2600k at stock will bottleneck a 7970 in some situations, such as in a 64-player game in Battlefield 3. Overclock it to 4.5+ ghz, and you'll see a marked difference.

CPU_01.png
Click to expand...

http://www.guru3d.com/article/radeon-hd-7970-cpu-scaling-performance-review/9


SO, because that is single player, ocing the cpu makes no diff.
Tell me why then, when I oc it to 4.4ghz, I SEE ZERO DIFFERENCE (in the min fps on that shariq map, on the roof, 64 players, and same settings).



IT does not make sense, that 100 extra cpu mhz (to 4.5ghz), would magically make my min fps go from 29 to 40; using 1920 by 1200, 4xaa, all on ultra, hbao on, motion blur on, and default fov.












I am gonna oc to 4.5ghz anyway, and see. THanks for the suggetions. I will report back with the result.
 
Last edited:
ocing to 4.5ghz. Minimum fps @ that roof on shariq peninsula & looking at the entire map (64 players); 33. At default cpu clocks, it was 29. So it did improve a bit. This is the most extreme/intensive spot on the bf3 multiplayer maps (or on of the most extreme. Some spots on the gulf of omen are right up there).



The graphical settings I used on the default cpu clock and 4.5ghz cpu oc was; 1920 by 1200, 4xaa, all the texture stuff on ultra, the rest on high, hbao on, motion blur on, & default fov.
 
icanseeu1985 said:
ocing to 4.5ghz. Minimum fps @ that roof on shariq peninsula & looking at the entire map (64 players); 33. At default cpu clocks, it was 29. So it did improve a bit. This is the most extreme/intensive spot on the bf3 multiplayer maps (or on of the most extreme. Some spots on the gulf of omen are right up there).



The graphical settings I used on the default cpu clock and 4.5ghz cpu oc was; 1920 by 1200, 4xaa, all the texture stuff on ultra, the rest on high, hbao on, motion blur on, & default fov.
Click to expand...

I dunno...I can look over the same map and see 40-45 fps. with a card half as good with 2XAA instead of 4.

I have had a 7970. You're just full of it...there's no way it's that low.



Uploaded with ImageShack.us
 
That's_Corporate said:
you call me dumb, but your posts are getting deleted by mods

riiiiiiiiiiiiiight... whatever you say, chief
Click to expand...

Cause I was a bit rude, & I did not mean to be. Some guy was trying to piss me off, and he got to me; so I responded rudely. But I should have ignored him.


Why would I lie about my performance? Like I made an account to get this solved. I spent 600 dollars on this video card.
 
spacin9 said:
I dunno...I can look over the same map and see 40-45 fps. with a card half as good with 2XAA instead of 4.

I have had a 7970. You're just full of it...there's no way it's that low.



Uploaded with ImageShack.us
Click to expand...

WHy would I be full of it? You are not looking at the entire map. turn left, exactly 45 degrees. Look at almost the ENTIRE Map. I spent 600 dollars, I want to see if my performance is normal. Why would I LIE? Why would I be full of it? I just don't understand. I want some help. Is that not what this place is about? Helping others? I have invested like 1 hour making my case, why would I waste my time lying? I am not against amd or against nvidia. I am just trying to figure out if my performance is normal. I SPENT 600 on the sapphire 7970.



My cpu oced to 4.5ghz = it changed my minimum fps of +4. So I got 4 more to my absolute min fps, with aa or without aa @ 1920 by 1200.



2600k sandy
p8p67 pro rev 3.0 b3
8 gigs mushkin ram
 
Last edited:
icanseeu1985 said:
Cause I was a bit rude, & I did not mean to be. Some guy was trying to piss me off, and he got to me; so I responded rudely. But I should have ignored him.


Why would I lie about my performance? Like I made an account to get this solved. I spent 600 dollars on this video card.
Click to expand...

well next time you ask for help, IN THE AMD FORUM, how about you not blatantly blast the company

it seems you want the card to be crap, since you formed your opinion before taking the advice you're asking for
 
That's_Corporate said:
well next time you ask for help, IN THE AMD FORUM, how about you not blatantly blast the company

it seems you want the card to be crap, since you formed your opinion before taking the advice you're asking for
Click to expand...

I did not do anything wrong, nor did I say anything false. Reviews should have included multiplayer portions of games (especially if the game is more graphically intensive online). Just sensitive fanboys get offended. Meanwhile I am not against amd or nvidia. I just typed the truth.
 
If the performance is the same with AA on and off at that point, then perhaps there is some other reason why you are getting such a low FPS at that point (perhaps a rendering issue with the game engine). Have you tried playing in other areas with AA on and off to see what a difference it makes? You seem to be hanging a lot of your complaint on that one FPS reading, instead of the total experience. Personally, I think those settings are just too high for a single card, even a 7970, to run smoothly.
 
I'm looking into getting 2 7970s, and I game on a single 1920x1080 120hz.

Why 2? Because I understand that 1 card will give you image quality at that rez. TWO cards, however, will give you fill rate (FPS). People who claim that 1 7970 is overkill for single monitors at 1920 have never heard of 120hz with 4x msaa.
 
I just tried this myself at 2560x1440
2600k @ 4.3
7970 @ 1200core/1700memory

here are my settings
4CNYf.png


frame rate dropped down to 37, with quick random dips to 35
n3KTe.jpg
 
That is good ^^^^^ 1200 on the core, 4.3 cpu oc. GOod stuff. Compared to what I am getting on my cpu oc and 1050/1575 gpu oc; EVERYTHING is running good on my end. Thanks guys. This is all I wanted to know.
 
xSystemFailurex said:
I'm looking into getting 2 7970s, and I game on a single 1920x1080 120hz.

Why 2? Because I understand that 1 card will give you image quality at that rez. TWO cards, however, will give you fill rate (FPS). People who claim that 1 7970 is overkill for single monitors at 1920 have never heard of 120hz with 4x msaa.
Click to expand...

I hear ya.. but it's not limited to 120hz. I swapped out my unlocked 6950's because they could just barely NOT deliver a solid 60fps with 4xMSAA and High post with HBAO at 1920x1200. Most reviews said "just use post High and forget about MSAA you'll hardly notice the diff... " I was like.. are you effin' nuts!!?? W/O MSAA it's jaggie city with white dot's all over. I don't know how they could make that recommendation with a straight face. Overkill.. wtf is that?? :p
 
Evil Scooter said:
I hear ya.. but it's not limited to 120hz. I swapped out my unlocked 6950's because they could just barely NOT deliver a solid 60fps with 4xMSAA and High post with HBAO at 1920x1200. Most reviews said "just use post High and forget about MSAA you'll hardly notice the diff... " I was like.. are you effin' nuts!!?? W/O MSAA it's jaggie city with white dot's all over. I don't know how they could make that recommendation with a straight face. Overkill.. wtf is that?? :p
Click to expand...

AA still gives a bigger performance hit than most other IQ options in single card mode Unfortunately :(
 
xoleras said:
AA still gives a bigger performance hit than most other IQ options in single card mode Unfortunately :(
Click to expand...

Not most.. any IIRC. It's a huge hit. But I'm so accustomed to running with AA maxed out in all my games at this point anything less looks like monkey butt.
 
You must log in or register to reply here.
Back
Top