Testbed and Testing Methodology
We test assembled system cases at a constant ambient temperature of 23°C maintained by an air conditioner. As we assume that most users prefer low-noise computers, we set the speed of the CPU and system fans (those connected to the mainboard’s 3-pin connectors) into Silent mode (the quietest mode in the mainboard’s BIOS). If a system case has its own speed controller, we switch it to minimum and maximum speed to check out both modes. We do not change the default configuration of air flows determined by system case design.
The following components are installed into each system case:
- Intel Core 2 Duo E6850 processor at 3.00 GHz
- Zalman CNPS9500 AT cooler
- ASUS P5E Socket 775 mainboard
- Western Digital Raptor WD740GD hard disk (10,000 RPM)
- Western Digital Raptor WD740ADFD hard disk (10,000 RPM)
- Western Digital Raptor WD1500ADFD hard disk (10,000 RPM)
- Western Digital Raptor WD1500AHFD hard disk (10,000 RPM)
- 2GB DDR2-800 SDRAM Patriot PDC24G6400LLK
- HIS HD 3870 IceQ3 Turbo H387Q512NP graphics card (Radeon HD 3870)
- Seasonic M12D SS-850EM power supply (850 W)
- Microsoft Windows 7 Home Premium 32-bit SP1
If not stated otherwise, the HDDs are listed in the order of their placement from the top main HDD bay downwards without any gaps.
The temperature of the CPU is measured with Core Temp 0.99.8. HDDs, GPU and mainboard temperatures are measured with CPUID Hardware Monitor. The speed of the fans is measured with an optical tachometer Velleman DTO2234. There are the following test modes:
- IOMeter (IOMeter’s Access Time test running on all the HDDs to load them fully)
- Linpack (Linpack-based Intel Burn Test 2.5 runs in the stress test mode, loading both CPU cores; we show you the peak temperature of the hottest CPU core in the diagrams)
- MSI Kombustor (full-screen mode, DirectX 9 rendering, 1280x1024 with 8x MSAA, Xtreme burn-in)
Every temperature is read after the system has worked for half an hour in the current test mode.
The following table shows the temperatures of the components if the system is assembled without an enclosure (“open testbed”).
The noise level is evaluated subjectively.