NVIDIA nForce 590 + AMD FX-62

Processors 213 Page 21 of 25 Published by

teaser

Page 21

Far Cry
Far Cry's story centers on Jack Carver, who has put a mysterious past behind him in favor of a less stressful life chartering boats in the South Pacific... or so he thinks. After delivering a female journalist to an uncharted island, Jack's boat is attacked and he subsequently finds himself stranded against a group of mercenaries, at which point his adventure begins. Graphically Far Cry is amazing; the action takes place in a huge, brightly colored environment with dense jungle style shrubbery, peacefully quiet beaches and large indoor areas. The excellent shadow effects simply bring the game to life with reflections on walls and even on your own weapons as you creep through the trees. Its like something straight out of a movie, helped along by the lighting and tight, crisp textures.  

For Far Cry we did things a bit different then normal. First off, the game has been patched to version 1.3. Secondly where possible we made sure that the graphics cards were forced to run Shader Model 3.0 if possible. Next to that we are using our own Guru3D.com constructed timedemo to prevent driver cheats.

The results (frames per second) that you see below are a lot lower then in normal conditions as we modified configuration settings and make it as rough as it can get on the graphics card. All in all, at this time and moment, this is one of the best tests we can offer you to benchmark DirectX 9 compatible graphics cards.

As you can see, not a huge lead here. I can explain this though. I've been telling you for many months now that this game is CPU limited and we need a faster CPU to make the GPU produce better scores.

With that being said, nice performance. We always go for the highest possible image quality during our tests. So our regular testing uses a personalized configuration file. The Far Cry scores you can see here are a little lower then you'll have at home, as we enable all possible graphics settings from the game engine. This goes for things like maximum texture sizes to looking at water and actually seeing sand banks underneath the water surface. So that's like the highest game setting available plus a few additions from our side and that does affect the framerate.

The framerate is measured with an intensive shader rich timedemo with a custom Guru3D.com written configuration file with everything maxed out.  The two results based on the FX-62/NF590 combo are kicking my high-end FX-57 system right in the nuts.

3DMark 05 Business Edition
3DMark 06
Business Edition

The latest in the 3DMark benchmark series built by Futuremark Corporation (formerly known as MadOnion.com). More than 5 million benchmark results have been submitted to Futuremarks Online ResultBrowser database. It has become a point of great prestige to be the holder of the highest 3DMark score. A compelling, easy-to-use interface has made 3DMark very popular among game enthusiasts. Futuremarks latest benchmark, 3DMark03, continues this tradition by providing a Microsoft DirectX 9 benchmark.

The introduction of DirectX 9 and new hardware shader technologies puts a lot of power in the hands of game developers. Increasingly realistic 3D games will be available over the next year and a half. The use of 3D graphics will become more accessible to other applications areas and even operating systems. In this new environment, 3DMark03 will serve as a tool for benchmarking 3D graphics.

This benchmark is not based on any game. Please remember this, never buy a graphics card based solely on the 3DMark score. I'm not bashing the 3D Mark suite here, it's good software but definitely not the sole basis for you to make an informed decision on to buy a graphics card, especially after what happened in 2003.

Let's have a look at 3DMark 03 and 05 scores. Default is again the standard 1024x768 reference test run from FutureMark with all options set to default.

Very respectable performance for all cards. 3DMark is designed in a way that excludes the CPU in the score calculation and tries to focus purely on DX9 graphics card performance.

Half-Life 2

Your mission is to save the planet from total alien supremacy. See, that petite incident in Black Mesa was just the beginning: now those pesky Xen invaders and a new threat called the Combine have spread across the whole Earth, causing massive amounts of death and destruction. Its up to you to set things right.

The source engine provides a gritty realism that surpasses (marginally) even Doom 3s "Super-real" prowess. While maybe not as visually spectacular as Doom 3, HL2s lighting seems a lot more "natural". Let me put it like this, Doom 3s lighting can seem like someone has inserted a laser light show onto Mars making it almost too spectacular, where as HL2s lighting is just "accepted" by the eye as lights reflect, and create shadows with precision streaming through windows with an unnerving realism.

For HL2 we recorded our own timedemo. We opted for the riverboat level where complex shaders will make things rough on the graphics card.

Above is the average framerate displayed versus it's resolution with no AA or AF settings enabled. Half-life 2 always has been in favor of Radeon cards bigtime but with the latest ForceWare drivers NVIDIA was able to close the gap. Normally we see massive CPU limitation. It's really easy to recognize. Once the scores go flat line for all graphics cards you are running into a limitation.

With CPU limitation the CPU can deliver geometric data fast enough to the graphics processor driver and therefore it can't render faster then it's capable of. Obviously with this new rig we have some more CPU and memory power to play arround with.

Again bare in mind, the graphics card and rest of the configuration is the similar on all tests, except the CPUs. This is how CPU bound HL2 is .. between the "slowest" X2 4000+ processor and the overclocked FX-62 at 3.13 GHz there is a 50 FPS differential in 10x7.

Share this content
Twitter Facebook Reddit WhatsApp Email Print