Installation/Operation

Physical installation into our test bed, a Zalman GT1000 case, gave us no hint of any size or physical layout issues to advise you of.  Clean, easy fit for a reasonably sized VGA card.  Software wise, driver installation and getting the card set went without any difficulties.

Temperature

As we have seen pretty much everywhere else, modern VGA cards run hot.  The Catalyst Control Panel sensor gave us an idle GPU temperature reading of 79 degrees Celsius.  With that starting point I loaded the GPU up to see how hot we could get the Visiontek HD4870 up to.  No matter how long we fragged we never saw the temps cross 85 degrees Celsius.  This is not the first time we have seen this exact trend but with no overheating of the card itself, I have no real reason to be dubious of the temp readings.  With the cooler completely encases in acrylic, our handy IR Thermometer was of no use.

Sound Level

The 70mm cooling fan on Visiontek's HD4870 is a variable rate unit that adjusts its temperature based on the readings of an internal diode.  The Visiontek card would periodically kick into "Dust Buster" mode but this almost seemed random occurring once every 90 minutes or so for less than one second.  The sound level during game play and at all other times was tolerable and with it running inside a closed case, was undetectable over normal room sounds.

Performance

Our test bed is based around an ASUS Maximus Formula X38 based mainboard.  The OS is Windows Vista Home Premium installed on a Western Digital SATA II HDD.  This is a bit of a change from our last test bed as one of our HDDs died making me appreciate RAID 1 (or my lack of it last time) all that much more.  Comparison cards today are going to be a Visiontek Radeon HD3870x2 and a Force3D HD4850.

Testing will be done with real world game play using frame rates recorded by the FRAPS program.  Results are graphed in three minute segments from three game titles, Crysis, Unreal Tournament 3 and a new game for us here, Assassin's Creed.  Inside each of these games, we have saved spots where we can record game play from a set location making the course of action easier to duplicate between test products.  Screen resolution and other settings will be listed out before each performance graph.  Full screen anti aliasing will be set to 0x in our testing.  ATi's Catalyst 8.6 drivers were used for all cards tested.

Test Bed

• ASUS Maximus Formula mainboard

• Intel Core 2 Duo 6600 (2.4GHz) @ 3.0GHz

• ThermalRight Chill Factor Thermal Paste

• Kingwin RVT 12025 D Cooler

• (2) 1GB OCZ Reaper PC2-6400 Memory (4-4-4-15) @ DDR2-867

• 250MB Western Digital Caviar SE SATA II HDD

• Generic 16x DVD-RW drive

• Kingwin Mach 1 1000w PSU

• Windows Vista 32 Home Premium with all service packs installed

• Open rack layout

BACK                    NEXT

Pg 1 - Introduction
Pg 2 - The Card
Pg 3 - Temperature, Sound, Test Bed
Pg 4 - Benchmarks
Pg 5 - Image Quality, Overclocking, Conclusion