The physical size of the Inno3D GeForce FX5700 Ultra is similar to that of the high-end GeForce FX 5950 Ultra. However, it is lighter and occupies one case slot. Under the heatsink are four GDDR3 memory modules that are mounted on the top side of the graphics card. The cooling system is effective as the average temperature during heavy GPU loads was 55C.

XFX 5950 Ultra vs. Inno3D 5700 Ultra

I usually advise customers not to use the first PCI slot since an add-in card can obstruct the flow of air. Cables can also be tucked out of the way when attempting to optimize airflow inside a case.

Case Shot During Testing

As you can see from the above photo, the second PCI slot is occupied by a sound card, which provides ample space for the graphics card cooling system to operate efficiently.

This is the part where I get to list the components of the system I used for testing. Although it's a high-end system, the GeForce FX 5700 Ultra is targeted at the lower mid-range. BTW, Western Digital's Raptor HDD is insanely fast!

• AMD Athlon 64 3000+ / 512MB L2 Cache
• MSI K8T Neo FIS2R Motherboard w/DualDDR Memory
• Mushkin PC3500, Level 1, DDR CAS2 RAM - (2) 512MB DIMMs - 1GB Total
• Western Digital Raptor 74GB, 10,000RPM, SATA HDD, Model WD740GD
• ViewSonic A90f+ CRT 19-Inch Monitor
• Vantec Stealth 470 Watt PSU
• Inno3D Tornado GeForce FX 5700 Ultra - 475MHz/950MHz
• ForceWare Graphics Driver Version 56.72 and 56.12
• 32-Bit Color / 85Hz Refresh Rate @ 1024x768
• Windows XP Professional with Service Pack 1 / DirectX 9.0b

Other Hardware:

• XFX GeForce FX 5950 Ultra - ForceWare Graphics Driver Version 56.72
• Sapphire Radeon 9600XT - Catalyst Version 4.3 Driver

Games Tested:

• Call of Duty
• Unreal Tournament 2003
• Unreal Tournament 2004 Demo
• Far Cry

Synthetic Benchmarks:

• 3DMark2001 SE
• 3DMark03 Build 340
• AquaMark 3
• Gun Metal
• CodeCreatures Benchmark Pro

Notes:

• All applications tested were patched to their latest version.
• Antialiasing and anisotropic filtering settings were configured using the graphics driver control panel.
• Gameplay tests were conducted with sound options set to each game's default settings.
• Gameplay tests were conducted with in-game graphics options set to their highest quality.
• FRAPS was used to to capture and compare average frame rates for some benchmarks.

Overclocking provided maximum stable frequencies of 540MHz/1020MHz, which worked flawlessly in all tests. I pushed the frequencies up to 550MHz/1050MHz at which time artifacts appeared.

Next Page: Call of Duty and Unreal Tournament 2003