Testing Configuration

To put the processors to the test, we’ll further add to our pool of results by using the same standard configuration that was used to test previous AM3 processors. We declined to test these new chips with DDR2 and DDR3 is rapidly becoming the mainstream choice, and these are mainstream processors.

AMD DDR3 Testing Configuration

• AMD Phenom II X2 550 Black Edition (3.1GHz) & Athlon II X2 250 (3.0 GHz)
• Noctua NH-U12P Heatsink with Noctua NF-P1 Fan
• Gigabyte GA-MA790FXT-UD5P motherboard
• 2×2048MB Corsair XMS3 DDR3-1600 Memory
• Western Digital 320GB SATA2 Hard Drive (WD3200AAKS, 7200RPM)
• HIS AMD Radeon HD4850 512MB GDDR3 Video Card
• OCZ 760W Power Supply

Intel DDR2 Testing Configuration

• ASUS P5K-E Motherboard (P35/ICH9 Chipset)
• 2×1024MB Corsair XMS2 PC2-8500 Memory
• Western Digital 320GB SATA2 Hard Drive (WD3200AAKS, 7200RPM)
• HIS AMD Radeon HD4850 512MB GDDR3 Video Card
• OCZ 760W Power Supply

Intel DDR3 Testing Configuration

• DFI Lanparty LT X48-T3RS Motherboard (X48/ICH9R Chipset)
• 2×2048MB Corsair XMS3 DDR3-1600 Memory
• Western Digital 320GB SATA2 Hard Drive (WD3200AAKS, 7200RPM)
• HIS AMD Radeon HD4850 512MB GDDR3 Video Card
• OCZ 760W Power Supply

For comparative data, the system memory is running at 1066MHz on both the Phenom and Core 2 processors. Timings are set to 5-5-5-15-2T via the EPP SPD profile in the dual channel kit. The Athlon II X2 processor cannot natively run 1066MHz memory, so the memory is set to 800MHz with the same timings. After conducting some preliminary testing, leaving the Phenom memory controllers ‘unganged’ for dual 64bit operation seemed to yield the best results in a variety of benchmarks.

For DDR3 testing of the two new processors, the system memory was set to 1333MHz to run on both the Phenom II and Core 2 processors. Our memory used 9-9-9-24-1T timings (per recommended settings), 1.8v vDIMM, and unganged dual channel mode.

OS/Driver Configuration:

We’ve taken the full step to Windows Vista for processor and graphics card testing due to the availability of DX10. Vista does pose some challenges for benchmark sessions as some of its features can create inconsistencies between testing runs. Although some reviewers prefer to ‘train’ their systems to make the most of features like SuperFetch, we have disabled these features to ensure a higher degree of predictability. Other ’scheduled’ or unnecessary background tasks have also been disabled (as listed below).

OS: Microsoft Windows Vista (64-bit), Service Pack 1
OS Disabled Features: “SuperFetch”, “Windows Search” service, Defender, Sidebar, Scheduled Defrag, System Restore and UAC.
OS Performance Features Enabled: “High Performance” PM setting enabled (except for power consumption testing)
Intel Chipset Drivers: Intel INF 9.0.0.1008
ATI/AMD Graphics Drivers: ATI Catalyst Version 8.12 Drivers

Benchmark/Applications:

The following benchmarks and revisions were used for testing:

• PCMark Vantage – Version 1.00
• 3DMark Vantage – Version 1.01
• Cinebench 64-bit – Release 10
• Crysis – Version 1.2
• X264 Benchmark – Version 2.0 (results posted from x264.exe v0.59.819M)
• Sisoft Sandra 2009 – SP2 (build 15.72)