![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
finally got around to unwrapping the cellophane off two brand spanking new his 4870's. i intended to post a crossfire comparison between x38 and p45 chipset boards, but i never got to trying many titles nor testing the card in the p45 rig. here are the specs of the x38 build:
q6600@3.38 ghz
asus maximus formula se
4x1 ocz ddr1100 5-5-5-15@1000mhz
raptor 150 raid0
vista ultimate 64
i only tried crysis once at 16x10, the rest of the tests were at 19x12. i chose to use the same stage that the [H] editors use in their video card evals. i began recording frame rates with fraps from the end of the cut scene when nomad receives the orders to take out the anti aircraft batteries, and ended following the landing of the vtol after destroying the second battery. i have god mode enabled, as well as unlimited ammo. i pretty much moved through the stage, taking out enemies and buildings with the rocket launcher. i played through the stage three times with each of my three sets of settings. my numbers aren't an average of the three runs, but the recorded numbers of the first run; using the 2 subsequent test plays to verify that there wasn't something weird going on.
1680x1050: all high 16x af. min 3, max 52, avg 28.32
1920x1200: object medium 16x af, rest high. min 0, max 40, avg 21.92
1920x1200: object/shadow medium 16x af, rest high. min 3, max 55, avg 32.86
i tried all settings at high with a resolution of 19x12 but that was utterly unplayable. the game remained in the low 20's even when there was nothing going on around you. micro stuttering was present at all times, but not to any excessive degree. even at the lowest average fps run, the micro stuttering was no more distracting than the already slow pace. i like the third set of settings best. even with the occasional micro stuttering, the game felt incredibly smooth because it felt like the 4870's kept the game from hitting minimum frame rates a large portion of the time. i was only able to make one run with shaders on very high, object/shadow on medium, and the rest on high so i didn't include them in my results, but the game play felt nearly as smooth as the third series of settings.
i made two series of runs in cod4 and after doing so, i understand why [H] dropped the game from their stable of tests. when it wasn't micro stuttering, the game was flying at impressively fluid frame rates. unfortunately, 4870's in cf micro stutter a lot in this game. i chose not to use 1600 resolutions in the cod4 test because i feel anyone contemplating buying two 4870 cards to play this game at $300+ a piece can and should spend $300+ on a monitor capable of at least 1920x1200. i chose the war pig stage and played through the entire stage on standard difficulty using fraps to record frame rates.
1920x1200: max in-game, 16x af, 8x ssaa box filter. min 9, max 417, avg 116.28
1920x1200: max in-game, 16x af, 16x ssaa box filter. min 9, max 372, avg 112.72
i tried 8x ssaa with edge detect, but the frame rates were unplayable.
i'll drag my p45 system into work tomorrow and try to run some tests. i'll probably also include a series of grid runs. hopefully someone interested in trying 4870's in crossfire will find this post useful.
q6600@3.38 ghz
asus maximus formula se
4x1 ocz ddr1100 5-5-5-15@1000mhz
raptor 150 raid0
vista ultimate 64
i only tried crysis once at 16x10, the rest of the tests were at 19x12. i chose to use the same stage that the [H] editors use in their video card evals. i began recording frame rates with fraps from the end of the cut scene when nomad receives the orders to take out the anti aircraft batteries, and ended following the landing of the vtol after destroying the second battery. i have god mode enabled, as well as unlimited ammo. i pretty much moved through the stage, taking out enemies and buildings with the rocket launcher. i played through the stage three times with each of my three sets of settings. my numbers aren't an average of the three runs, but the recorded numbers of the first run; using the 2 subsequent test plays to verify that there wasn't something weird going on.
1680x1050: all high 16x af. min 3, max 52, avg 28.32
1920x1200: object medium 16x af, rest high. min 0, max 40, avg 21.92
1920x1200: object/shadow medium 16x af, rest high. min 3, max 55, avg 32.86
i tried all settings at high with a resolution of 19x12 but that was utterly unplayable. the game remained in the low 20's even when there was nothing going on around you. micro stuttering was present at all times, but not to any excessive degree. even at the lowest average fps run, the micro stuttering was no more distracting than the already slow pace. i like the third set of settings best. even with the occasional micro stuttering, the game felt incredibly smooth because it felt like the 4870's kept the game from hitting minimum frame rates a large portion of the time. i was only able to make one run with shaders on very high, object/shadow on medium, and the rest on high so i didn't include them in my results, but the game play felt nearly as smooth as the third series of settings.
i made two series of runs in cod4 and after doing so, i understand why [H] dropped the game from their stable of tests. when it wasn't micro stuttering, the game was flying at impressively fluid frame rates. unfortunately, 4870's in cf micro stutter a lot in this game. i chose not to use 1600 resolutions in the cod4 test because i feel anyone contemplating buying two 4870 cards to play this game at $300+ a piece can and should spend $300+ on a monitor capable of at least 1920x1200. i chose the war pig stage and played through the entire stage on standard difficulty using fraps to record frame rates.
1920x1200: max in-game, 16x af, 8x ssaa box filter. min 9, max 417, avg 116.28
1920x1200: max in-game, 16x af, 16x ssaa box filter. min 9, max 372, avg 112.72
i tried 8x ssaa with edge detect, but the frame rates were unplayable.
i'll drag my p45 system into work tomorrow and try to run some tests. i'll probably also include a series of grid runs. hopefully someone interested in trying 4870's in crossfire will find this post useful.
OMG someone with brains is posting today!!! There is NO point in spending $300+ on a VGA if you dont have a 19x12 or 25x16 monitor coming! 
