AMD has positioned the Phenom II X6 1090T at the very top of the Thuban hierachy, which simply means that living the life of a Phenom II X6 1090T is really tough. This is especially so given that the competitor’s processors have long been known for having more Instructions Per Clock (IPC) than AMD’s past CPUs.

AMD has ranked the Phenom II X6 1090T squarely against the Intel Core i7-860 and Core i7-930 CPUs for their own in-house testing. We’d test the AMD Phenom II X6 1090T against what we have on hand. The candidates are listed below:

1 AMD Phenom II X6 1090T Tested

The systems in which the candidates are tested can been seen in the following list:

AMD Phenom II X6 1090T Setup
  • AMD Phenom II X6 1090T
  • XFX ATI Radeon HD 5870
  • Antec TruePower Quattro 1200
  • Kingston Technology SSDNow V+ SNV225-S2/256GB
  • Western Digital Cavair Black WD5001AALS
  • Asus Crosshair IV Formula (0707 BIOS)
  • GEIL PC3-16000 (2000MHz) CL9 ULTRA Dual Channel Kit @ DDR3-1600MHz 8.8.8.24 
AMD Phenom II X4 955 Setup
  • AMD Phenom II X4 955
  • XFX ATI Radeon HD 5870
  • Antec TruePower Quattro 1200
  • Kingston Technology SSDNow V+ SNV225-S2/256GB
  • Western Digital Cavair Black WD5001AALS
  • Asus Crosshair IV Formula (0707 BIOS) 
  • GEIL PC3-16000 (2000MHz) CL9 ULTRA Dual Channel Kit @ DDR3-1600MHz 8.8.8.24 
AMD Phenom II X2 555 (Unlocked) Setup
  • AMD Phenom II X2 555 (Unlocked)
  • XFX ATI Radeon HD 5870
  • Antec TruePower Quattro 1200
  • Kingston Technology SSDNow V+ SNV225-S2/256GB
  • Western Digital Cavair Black WD5001AALS
  • Asus Crosshair IV Formula (0707 BIOS) 
  • GEIL PC3-16000 (2000MHz) CL9 ULTRA Dual Channel Kit @ DDR3-1600MHz 8.8.8.24 
Intel Core i7-980X Setup
  • Intel Core i7-980X
  • XFX ATI Radeon HD 5870
  • Antec TruePower Quattro 1200
  • Kingston Technology SSDNow V+ SNV225-S2/256GB
  • Western Digital Cavair Black WD5001AALS
  • MSI Eclipse SLI (7520v1C BIOS)
  • ADATA DDR3-1800+ CL8 Triple Channel Kit @ DDR3-1600MHz 8.8.8.24 
Intel Core i7-920 Setup
  • Intel Core i7-920
  • XFX ATI Radeon HD 5870
  • Antec TruePower Quattro 1200
  • Kingston Technology SSDNow V+ SNV225-S2/256GB
  • Western Digital Cavair Black WD5001AALS
  • MSI Eclipse SLI (7520v1C BIOS)
  • ADATA DDR3-1800+ CL8 Triple Channel Kit @ DDR3-1600MHz 8.8.8.24 
Core i7-860 Setup
  • Core i7-860
  • XFX ATI Radeon HD 5870
  • Antec TruePower Quattro 1200
  • Kingston Technology SSDNow V+ SNV225-S2/256GB
  • Western Digital Cavair Black WD5001AALS
  • Asus P7P55D Deluxe (1408 BIOS)
  • GEIL PC3-16000 (2000MHz) CL9 ULTRA Dual Channel Kit @ DDR3-1600MHz 8.8.8.24
Test Environment & System Details
 AMD Phenom II X6 1090T Tested
3 AMD Phenom II X6 1090T Tested
Software Used
  • ATI Catalyst 10.3 Suite Windows 7 (64 bit)
  • CPUID CPU-Z 1.54 (64-bit)
  • Colin McRae: Dirt 2 Demo
  • Futuremark 3DMark Vantage 1.0.2
  • GIMPS Prime95 v25.11
  • Lavalys EVEREST Ultimate Edition v5.50
  • Maxon CINEBENCH 11.5
  • Microsoft Windows 7 Ultimate (64-bit)
  • OCCT Perestroïka 3.1.0
  • RARLAB WinRAR 3.93
  • Resident Evil 5 Benchmark
  • SiSoftware Sandra Lite 2010 SP1d 16.41
  • Super PI Mod v1.5
  • Tech ARP x264 HD Benchmark v3.18
  • QUAKE 4 1.4.2 Multiplayer Demo
Memory Controller
While glancing through AMD’s documentation on the Phenom II X6 1090T CPU, we noticed a paragraph suggesting that the “Northbridge HT Link” multiplier be set at 11x for DDR3 1333MHz and 12x for DDR3 1600MHz memory. Our testing suggests that the impact of memory controller speed on overall system performance is not small, and that is why our testing on the AMD platform is conducted using (manually keyed-in) recommended IMC speeds. 

If you do find yourself in a situation necessitating the set up of a Thuban based PC, do not forget to raise the memory controller frequency manually if “Auto” settings are not up to mark. Those who wish to tweak the most out of their Thuban CPUs may wish to raise Northbridge frequency up to the limits of stability.

Power Saving and Multipliers
All the six CPUs tested today have a number of power saving features. Microprocessor architecture has become rather substantial over the years, and hence it is possible for significant power savings to be had putting certain processor subsystems offline. The problem is that, regardless of architecture, processors are ultimately electrical systems subject to various electrical influences and governed by the laws of Physics. A finite amount of time is required to switch offline systems back online (blame Miller) and that introduces variability we wish to avoid when benchmarking. During the entire course of this review, all benchmarks were conducted on systems with power saving features disabled. Another cause of benchmarking inconsistency is described in the following paragraph.

Intel Turbo Boost Technology and AMD Turbo CORE Technology both fall under what I call “dynamic multiplier management.” Judging by the amount of CPU activity, the CPU multiplier will vary within electrical and thermal limits, much to the delight of a day-to-day PC user. For benchmarking consistency however, dynamic multipliers are about as predictable as Hot Lotto (or TOTO if you prefer the local flavour). Along with power saving features, dynamic multipliers have been disabled for all CPUs under test.