X-bit labs

Intel Pentium 4 6XX and Intel Pentium 4 Extreme Edition 3.73GHz CPU Review (page 9)

Category: CPU

by Ilya Gavrichenkov

[ 02/20/2005 | 12:05 AM ]

Testbed and Methods

The main goal of this test session is to find out the performance level provided by the new Intel Pentium 4 6XX processors and the Intel Pentium 4 Extreme Edition 3.73GHz processor, and to compare their performance with that of the predecessors and competitors. For our tests we took two top models from the 6XX family, Pentium 4 660 working at 3.6GHz and Pentium 4 650 working at 3.4GHz.

The testbeds were configured as follows:

• CPUs:
  • AMD Athlon 64 FX-55 (Socket 939, 1024KB L2, 2.6GHz);
  • AMD Athlon 64 4000+ (Socket 939, 1024KB L2, 2.4GHz);
  • AMD Athlon 64 3800+ (Socket 939, 512KB L2, 2.4GHz);
  • AMD Athlon 64 3500+ (Socket 939, 512KB L2, 2.2GHz);
  • Intel Pentium 4 570 (LGA775, 1024KB L2, 3.8 GHz);
  • Intel Pentium 4 560 (LGA775, 1024KB L2, 3.6 GHz);
  • Intel Pentium 4 550 (LGA775, 1024KB L2, 3.4 GHz);
  • Intel Pentium 4 660 (LGA775, 2048KB L2, 3.6 GHz);
  • Intel Pentium 4 650 (LGA775, 2048KB L2, 3.4 GHz);
  • Intel Pentium 4 Extreme Edition 3.46GHz (LGA775, 2048KB L3, FSB 1066MHz);
  • Intel Pentium 4 Extreme Edition 3.73GHz (LGA775, 2048KB L2, FSB 1066MHz);
• Mainboards:
  • EPoX 9NPA Ultra (Socket 939, NVIDIA nForce4 Ultra);
  • Intel Desktop Board D925XECV2 (LGA775, i925XE Express).
• Memory:
  • 1024MB DDR400 SDRAM (Corsair CMX512-3200XLPRO, 2 x 512MB, 2-2-2-10);
  • 1024MB DDR2-533 SDRAM (OCZ PC2 4300, 2 x 512MB, 4-4-4-11).
• Graphics card: PowerColor RADEON X800 XT (PCI-E x16).
• Storage subsystem: Maxtor MaXLine III 250GB (SATA150).

We ran the tests in MS Windows XP SP2 with installed DirectX 9.0c. The test systems were configured for maximum performance. Note that we increased the Cycle Time (Tras) in Athlon 64 to 10. Our practical experience shows that in this mode the memory controller of AMD Athlon 64 processor works better than in case this timing is set to the minimal value of 5.

This time we included much more applications into the tests, because we decided to start using a PC WorldBench 5 standard scripts set, which is widely used in the industry today for benchmarking purposes.