Medal Of Honor Warfighter Performance, Benchmarked

Benchmark Results: Ultra Quality Preset

The Ultra preset adds 4x MSAA anti-aliasing and the highest detail levels available.

Of course, nobody is going to play this game on one of these cards at 1280x1024. For the same of comparison, though, we're able to see how much more demanding the Ultra preset is than Medium. The Radeon HD 7770 and GeForce GTX 650 Ti manage to stay above 30 FPS, but they don't have a lot left in them to handle higher resolutions.

At 1680x1050, AMD's Radeon HD 7770 and Nvidia's GeForce GTX 650 Ti dip below 30 FPS a couple of times. We'd still consider them viable options, though. More attractive, perhaps, are the Radeon HD 7850 1 GB and GeForce GTX 660, which sustain performance in excess of 40 FPS at all times.

AMD's Radeon HD 7850 1 GB suffers most pointedly from its single gigabyte of memory, which hurts its performance with high levels of anti-aliasing and Shadow Quality applied. We'd now consider the Radeon HD 7770 and GeForce GTX 650 Ti unplayable.

We also add the Radeon HD 7850 1 GB in CrossFire and a pair of GeForce GTX 660s in SLI, along with a GeForce GTX 680. Nvidia's fastest single-GPU board keeps up with AMD's Radeon HD 7970, while the Radeon HD 7850 1 GB cards in CrossFire achieve a slightly higher average frame rate than the single-GPU boards, but again are hampered by significant frame rate drops. The GeForce GTX 660s in SLI easily deliver the most attractive performance in our test.

At high resolutions, graphics memory capacity is needed most. That's why the Radeon HD 7850 1 GB setup in CrossFire is hammered to the point of being unplayable (succumbing even to a single GeForce GTX 660). Again, you've have to cut anti-aliasing and shadow detail to get this setup back in the realm of acceptable performance.

Just because the 2 GB GeForce GTX 660 beats a paid of Radeon HD 7850s doesn't make it playable, though. Really, you need a GeForce GTX 670, 680, or Radeon HD 7970 to average more than 40 FPS and maintain a sub-30 FPS minimum.

Two GeForce GTX 660s in SLI give you a good average frame rate, but we did notice instances where performance would drop precariously during our time gaming on them. Unfortunately, we've found multi-card configurations to be inconsistent at times, and this is no exception.

Create a new thread in the US Reviews comments forum about this subject
This thread is closed for comments
54 comments
    Your comment
    Top Comments
  • greghome
    No 7850 2GB to see if it's a memory bottleneck ? :/

    and you're missing the 7870 and 7950 in them. just sayin'
    22
  • JJ1217
    You put a 7850 1 GB, so now no one is going to buy a 7850 to play this game, as they'll get the wrong results due to memory bandwidth constraints. People who know about video ram will have no issue with this, but what about those looking for a good cheap video card to run games well? You pretty much just destroyed any chance of someone getting a 7850 for this game, due to the terrible gathering of results.

    Expected more from T.H to be honest.
    17
  • ojas
    mohit9206its great to see that entry level cards like 650, 7750 and 7770 are all a viable option even at 1080p at high setting !!! am so proud of my 7750 .. hehe..btw i dont agree with toms on the fact that a game becomes "UNPLAYABLE" if its minimum fps drops below 30.thats just a load of bulls**t.

    Try playing the game (or any game) on a constant 60 and you'll see.

    Of course the level of comfort (as far as fps is concerned) varies from person to person, I personally don't enjoy it when the frame rates drop below 40, and sub 30 is intolerable.

    I guess what Don meant by unplayable was intolerable. And i guess most here, including me, would agree.
    12
  • Other Comments
  • mayankleoboy1
    Nice review! :)
    In CPU benchmark, it would have been better to see the continuous FPS graph , rather than just the single values of 'Average' and 'minimum' .

    Also, CPU frequency scaling is needed
    4
  • rmpumper
    No one is playing this crap. Why waist your time testing it?
    -18
  • esrever
    Interesting that the 1gb on the 7850 starts showing signs of weakness at higher settings even at 1080p. The minimals went lower than the 7770 :o

    I think nvidia's gpu boost is causing the nvidia cards to have higher average and lower minimals since it can render higher fps when less things are going on but they can only have so much performance when the rendering gets tough. I think GPU boost is a pointless feature because of that since why would anyone want high maximal fps and low minimal fps?
    10
  • greghome
    No 7850 2GB to see if it's a memory bottleneck ? :/

    and you're missing the 7870 and 7950 in them. just sayin'
    22
  • JJ1217
    You put a 7850 1 GB, so now no one is going to buy a 7850 to play this game, as they'll get the wrong results due to memory bandwidth constraints. People who know about video ram will have no issue with this, but what about those looking for a good cheap video card to run games well? You pretty much just destroyed any chance of someone getting a 7850 for this game, due to the terrible gathering of results.

    Expected more from T.H to be honest.
    17
  • JJ1217
    Woops didn't mean memory bandwidth, meant amount of memory ^.^
    8
  • EzioAs
    321797 said:
    No 7850 2GB to see if it's a memory bottleneck ? :/ and you're missing the 7870 and 7950 in them. just sayin'


    I'm curious as well, though in my opinion it's most probably a memory bottleneck at 1080p wilth ultra settings. BF3 already uses more than 1GB with max image settings with 4xAA as well so if Warfighter uses an updated Frosbite2 engine, it's highly plausible.

    On the other hand, I'm not fully satisfied that they didn't test the game with the 7870. And how about 560ti and 6870(the 2 very popular card from last-gen), I think at least a couple mid-range card from last gen should be tested
    6
  • greghome
    EzioAshow about 560ti and 6870(the 2 very popular card from last-gen), I think at least a couple mid-range card from last gen should be tested


    i miss my 6950 on benchmarks.......
    Story of my hardware life.

    First Year, Wow Top of the line
    2nd Year, Still in benchmarks
    3rd Year, Still performing good enough
    4th Year......I need an uphrade
    4
  • the3dsgeek
    Can you please do a performance benchmark comparison of NFS most wanted? its running like shit on my GTX670
    7
  • ojas
    Liked the way you ran benchmarks, covered all major resolutions with all major detail levels across a wide spectrum of cards.

    Anyway, didn't really read your game review, but Rock, Paper, Shotgun was extremely critical of the game, and i understand their sentiment, because BF3 is similar in some respects.
    http://www.rockpapershotgun.com/2012/10/29/wot-i-think-medal-of-honor-warfighter/

    P.S. Why you no benchmark Sleeping Dogs? It brings my GTX 560 down to 40 fps minimums at 1024x768 at the highest settings...It may be a CPU bottleneck though, have to look into that fully.
    4
  • ttg_Avenged
    Doesn't even touch my GTX 680..
    -12
  • ojas
    the3dsgeekCan you please do a performance benchmark comparison of NFS most wanted? its running like shit on my GTX670

    Lol that's because it's a sucky console port.
    7
  • mohit9206
    its great to see that entry level cards like 650, 7750 and 7770 are all a viable option even at 1080p at high setting !!! am so proud of my 7750 .. hehe..
    btw i dont agree with toms on the fact that a game becomes "UNPLAYABLE" if its minimum fps drops below 30.
    thats just a load of bulls**t.
    -6
  • captainblacko
    Im shocked at the Pentium G860's FPS. that's pretty impressive for a £52 CPU!
    2
  • Iastfan112
    I always give a big sigh when I see them acknowledge that the multiplayer is likely a CPU bottleneck....yet we're not going to make any sort of attempt to illustrate where it exists. It'd be lovely to know, for instance, does the 4170's four "cores" help it compared to the i3?

    I understand there would be a significantly greater margin of error compared to the repeatable SP benches but the information would still be pertinent and useful.
    7
  • ojas
    mohit9206its great to see that entry level cards like 650, 7750 and 7770 are all a viable option even at 1080p at high setting !!! am so proud of my 7750 .. hehe..btw i dont agree with toms on the fact that a game becomes "UNPLAYABLE" if its minimum fps drops below 30.thats just a load of bulls**t.

    Try playing the game (or any game) on a constant 60 and you'll see.

    Of course the level of comfort (as far as fps is concerned) varies from person to person, I personally don't enjoy it when the frame rates drop below 40, and sub 30 is intolerable.

    I guess what Don meant by unplayable was intolerable. And i guess most here, including me, would agree.
    12
  • mayankleoboy1
    ojasTry plating the game (or any game) on a constant 60 and you'll see.Of course the level of comfort (as far as fps is concerned) varies from person to person, I personally don't enjoy it when the frame rates drop below 40, and sub 30 is intolerable.I guess what Don meant by unplayable was intolerable. And i guess most here, including me, would agree.


    Playing on intel IGP + P4 for many years made me accustomed to 30FPS. :P
    6
  • Onus
    Interesting. I too have to wonder about the 1GB HD7850. The results don't appear to extrapolate cleanly to my 2GB HD7870.
    I've noticed you've used the DDR3 version of the HD6670 in recent tests, and would really like to see the GDDR5 version instead. For those who can't quite afford a HD7750, it seems to me that even the most entry level card for games should be one with GDDR5. Particularly in this case, it looks like this change might cross the line back into "playable" on some settings.
    It is also rather remarkable that an old Athlon II X2 240 can play this game as well as it does. Even though objective measurement might not be possible, I think some subjective observations on its ability to handle Multi-player would be useful.
    4
  • ojas
    jtt283Interesting. I too have to wonder about the 1GB HD7850. The results don't appear to extrapolate cleanly to my 2GB HD7870.I've noticed you've used the DDR3 version of the HD6670 in recent tests, and would really like to see the GDDR5 version instead. For those who can't quite afford a HD7750, it seems to me that even the most entry level card for games should be one with GDDR5. Particularly in this case, it looks like this change might cross the line back into "playable" on some settings.It is also rather remarkable that an old Athlon II X2 240 can play this game as well as it does. Even though objective measurement might not be possible, I think some subjective observations on its ability to handle Multi-player would be useful.

    I think their GPU chart puts the 6670 GDDR5 two tiers above the GDDR3...at par with a 9800GT.

    Also Tom's: Dishonored and Hitman: Absolution. One i know is resource intensive, the other one simple looks great, so i'm interested. :P
    3
  • Katsu_rap
    Can't believe my 560ti isn't in the benches. It's not too long ago that I bought it and I believe many people who bought it a couple months ago aren't planning for another video card upgrade just yet.

    I'm not really complaining but you know, where's the value of mid-range cards if the next gen cards and new games comes out and they aren't even tested?
    5