Review: Mushkin 1GByte (2x512MB) Level-II PC3200 RAM

by Tarinder Sandhu on 9 December 2004, 00:00

Tags: Mushkin

Quick Link: HEXUS.net/qa4n

Add to My Vault: x

System setup and notes

Here's a quick rundown of the test system should you wish to compare benchmark results with your own.
  • Mushkin 1GByte (2x512MB) Level-II PC3200 RAM (2-2-2-5 @ DDR400, and 2.5-3-3-7 @ DDR500)
  • OCZ PC3200 Platinum Rev 2 1GByte (2-2-2-5 @ DDR400)
  • Crucial Ballistix PC4000 1GByte (2x512MB) (2-2-2-5 @ DDR400)

Other components

  • Intel Pentium 4 3.4GHz Northwood S478 CPU run at 3.0GHz (15x200MHz FSB)
  • ABIT IC7-G i875P Canterwood motherboard with PAT-enabled BIOS
  • NVIDIA GeForce 6800 GT 256MB, 8x AGP, 350/1000
  • Pioneer 105 DVD-RW
  • Western Digital 160GB (WD1600) 8MB cache hard drive
  • Dell P991 19" monitor

Software

  • Windows XP Professional SP1
  • DirectX 9.0b
  • Intel 5.1.1002 driver set
  • NVIDIA ForceWare 61.77 drivers
  • Pifast v4.1 to 10m places
  • ScienceMark 2.0
  • 3DMark 2001SE v330
  • UT2003 Retail (Build 2225)
  • DOOM 3 - 1024x768 - High Quality - Demo 1
  • HEXUS XviD media encoding test


Notes

Three sets of extreme performance memory up against each other. The basic setup consisted of a Pentium 4 3.0GHz Northwood CPU run at 15x200MHz FSB. To make things interesting, I also ran the Mushkin pair at 250MHz FSB (DDR500) with timings of 2.5-3-3-7 timings. To keep things in parity, I dropped the CPU's multiplier down to 12x, such that we're looking at:

3006MHz - 15x multiplier - 200MHz FSB - DDR400 - 2-2-2-5 timings.

3006MHz - 12x multiplier - 250MHz FSB - DDR500 - 2.5-3-3-7 timings.

What I'm attempting to answer is, given a similar overall clock speed, whether having a higher FSB and higher latencies (1:1 CPU-to-RAM ratio) is better than a comparatively lower FSB and tight timings. On to the tests.