Review: Jetway PT800TWIN

by Ryszard Sommefeldt on 13 September 2004, 00:00

Tags: Intel (NASDAQ:INTC), VIA Technologies (TPE:2388), Jetway

Quick Link: HEXUS.net/qaxd

Add to My Vault: x

System Setup

Hardware

  • Jetway PT800TWIN, Socket 478, VIA PT800, Pentium 4, 26/11/2003 BIOS
  • Intel Pentium 4 3.2ES, 512KB L2, 12 x 200MHz, 2.4GHz
  • Western Digital WD360 Raptor, SATA, 36.2GB
  • NVIDIA GeForce FX5950 Ultra
  • Corsair XMS3200LLPT, 2 x 256MB, DDR400

Software

  • Windows XP Professional w/SP1
  • ATI CATALYST 4.1 and Control Panel
  • VIA Hyperion 4.51v
  • MagicTwin 4.02.105
  • HEXUS Pifast v41
  • Simplisoft HDTach 2.61
  • Kribi Bench 1.19
  • Sciencemark 2.0
  • 3DMark 2001SE v330
  • Quake3 v1.30 HQ (four demo)
  • LAME 3.92MMX MP3 Encoding(192CBR, U2's Pop album)
  • Realstorm Ray Tracing
  • X2: The Threat - Rolling Demo

The Jetways gets benchmarked against itself for a couple of reasons. Firstly, the emulated 2.4GHz Pentium 4 rules out direct CPU-level comparisons with other Pentium 4 boards reviewed at HEXUS. Secondly, the performance delta obtained when using MagicTwin is the main focus, not absolute raw performance (although that is tested, but only at 2.4GHz).

So you'll see two entries on the graphs, one for the base 2.4GHz figure, one for the figure when MagicTwin is in use and there's a load being generated by one of the stations.

A load was generated on Station 2 by logging in at each reboot and running a copy of Windows Media Player 9, playing a DivX rip of The Score, along with a single instance of IE6 SP1, Wordpad and Windows Explorer. That provided a moderate CPU load (average of 15%) and memory load, along with some disk activity. Benchmarking something like MagicTwin is unscientific at best, but we're looking to see what effect the 15% CPU load, and reduced memory capacity available to Station 1, has on performance.

Station 2 also used hardware acceleration on the FX5950 Ultra, meaning the graphics card's potential performance is reduced too. It's a moderate, general desktop load and should give us a good indication of how MagicTwin is able to split the resources. Windows XP does all the hard work for it so we're looking for as little overhead as possible.

Here's how CPU-Z saw it.



A naughty 1MHz overclock on the front side bus frequency when set to 200MHz in the BIOS.



The aforementioned 2-2-2-5 latencies are confirmed.



PT800 and VT8237 are also confirmed, along with the BIOS date and AGP settings.