Our first game here is Far Cry.  This game was released before Half Life 2 and Doom 3, and the graphics shocked quite a few people, with realistic grass and water effects.  This game is still visually stunning, so how does this card handle it?

    We can see that the game is basically CPU limited at this resolution as the increase in frame rate for the overclocked system is 20% and the clock speed increase is 25%.  Also the fact that the two cards perform exactly the same seems to indicate this, but lets turn up the detail settings to see if this is true, or if the video cards are in fact equal in this game.

    Again we see the same results as the previous test, only the overclocked CPU/video card showing any difference.  Does turning up the resolution make any difference?

    Still nothing new to report, all three test setups should have no problem playing this game so far.  For the last test, will there be any divide between the two cards?

    Finally we see a seperation between the two cards.  At this point the x1800XL is running at about 45fps on average with the x800XL rendering at 38fps on average.  This translates into a 17% difference in average frame rate, or about 7fps.  Overclocking provides a 6% increase in frame rate, or about 2fps, nothing to write home about but still a difference.  Overall Far Cry is playable on both cards at basically all settings we tried, though some may find the x800XL might not be fast enough for extremely heavy action/shader areas.  Lets now look at a slightly newer game to see if these two cards are indeed CPU dependent in D3D games.

    Next we will look at Half Life 2.  Simply put this game has very nice graphics, from facial detail to water detail to the physics of the game.  So lets see how these two cards perform in this popular game.

    As we look at this second by second graph, what can we see?  Looking at the average frame rates, we see the there is a jump from about 75fps from the x800XL to about 90fps for the x1800XL.  The x800XL spends about 23 seconds below the 60fps mark, mostly in two sections (6-14 seconds and 48-56 seconds), though it only dips below the 40fps mark for one second.  The x1800XL on the other hand only spends four seconds below the 60fps mark, and this is just below this mark, at the 56fps mark.  Overclocking gives a 16% improvement to slightly over the 100fps average mark, with the minimum frame rate not reaching below 70fps for more than about 4-5seconds in total.  Overall at this setting all three cards are more than playable.  Lets turn up the quality settings.

    Here we see very little difference in the performance of the x1800XL card in both its forms.  The x800XL declines somewhat in its average from about 75fps to 65fps.  This makes the increase from upgrading cards greater, to 32% from 20% in the previous test.  The x800XL takes a few trips below the 30fps mark in this test, though is still quite playable for most if not all.  The x1800XL follows the same trend as in the previous test, not going under 60fps for more than a few seconds in total and the same for the 70fps mark when overclocked.  What happens when we up the resolution?

    Here the x800XL makes a slight comeback from the previous result, while overclocking the CPU/GPU doesn't make as much of a difference.  The x800XL follows its results for its first Half Life 2 test it provides a slightly (5fps) lower average frame rate.  The x1800XL stays around the 85fps average again providing a frame rate that only very rarely goes under the 60fps mark.  Overclocking nets a 13fps increase or about a 15% improvement in average frame rate.  Will upping the quality settings make any more of a difference between the two cards?

    This test shows a difference between the two cards that we saw in our Doom 3 tests.  The x800XL gives you 40fps on average, while the x1800XL provides about 65fps on average.  The difference between the two cards is 64% or about 25fps.  The x1800XL drops below the 40fps mark for a total of about 4-5 seconds, which is pretty good, and is what the x800XL gave at 1024*768 with the same quality settings.  Overclocking gives a 10% increase in frame rate, which should be attributable to the GPU overclock.  This game is playable at all of our test settings with the x1800XL and all but the highest with the x800XL.  Now lets look at some DirectX 9.0c or Pixal Shader 3.0 based games.

    The first of the three new games is Splinter Cell 3: Chaos Theory.  This game is graphically a big improvement over the previous two games, with massive changes taking place in the engine, out were DirectX 7 and 8 code and in was support for AA and DirectX 9.0c features.  Gameplay hasn't changed, but graphics definitely have improved.  Lets see how this new card handles this game.  As a point we also included results with the PS3.0 features enabled in our tests without AA.

    Here we see four different results.  The x800 provides a very playable 65fps on average, with the frame rate never dropping below the 40fps mark.  Next is the PS3 based x1800XL, which provides an average just over 70fps, which means that you can upgrade the quality settings to the absolute maximum with this card upgrade.  With these features turned off the card gains another 10fps, and only drops below 60fps briefly.  Overclocking gives another 10fps, showing some CPU limitedness, with the minimum never going below 65fps.  Lets see how turning on AA and ansiotropic filtering changes the results?

    As we move up the quality settings, we loose the PS3.0 features as not all are supposed to play well with AA.  The results show a generation gap, as the x800XL is about 20fps slower than the x1800XL.  All three cards are still playable in my books, as I finished the game with an average frame rate nearer to 25fps and it still worked.  The only card to go under the 30fps and basically the 40fps mark is the x800XL, with the x1800XL only going under 40fps once.  So lets see what turning up the screen size does to the frame rate.

NEXT