In the Fourier test, we ran simultaneous copies of the floating-point-intensive
Fourier transform program. This program analyzes the spectral content of a WAV file; its results are largely dependent on raw floating-point power.
The second processor test measures how fast a system can build (i.e., compile and link) a large amount of source code by running two simultaneous instances of version 4.3 of Visual C++. This benchmark produces results that also depend on hard disk I/O performance, because the tested systems must access their hard disks often for both reading and writing when running the test.
The Access test is the lone application-based benchmark in this suite. Using Microsoft's Visual Test 4.0 compiler, it measured how well these PII systems could generate multiple business-database transactions. It executes two iterations of test runs.
The IMark video tests were conducted at 640- by 480-pixel resolution with 256 colors, using each vendor's preinstalled graphics driver.
The single-threaded BYTEmark test exercises the capabilities of a system's CPU, FPU,
and memory system.
Usability and Features
We examined each of the 17 reviewed machines for usability by focusing on system design and documentation. System design is important; therefore, we rate how easy it is to remove and reinstall the cover, for example. Ease of upgradability (for adding RAM and mass-storage devices) adds points to the usability score. We look for accessible memory slots and drive bays that offer room for cabling. Systems are penalized for having I/O connectors that are blocked by an expansion slot.
We give top honors to systems with vendor-specific manuals that are comprehensive, include easy-to-use diagrams, and offer up-to-date technical information. Manuals with a glossary and an index get better scores. Those that lack specifications for the motherboard, hard disk, and graphics subsystem get rated lower.
Each vendor completes a lengthy questionnaire that we send to provide a detailed description of each system's features and support options. We weight each
feature and calculate an overall features score. Features related to performance include the size and write policy of the secondary cache. The use of hard disk controllers, network adapters, and graphics adapters on the motherboard frees up expansion slots. The amount of video memory supplied with the system is critical in determining the maximum display resolution. We weight the highest noninterlaced display resolution supported by each system, along with the maximum number of colors at that resolution.
Finally, it's important to note that warranty and support policies are what frequently separate major system vendors from second- and third-tier vendors.
Evaluations in this report represent the judgment of BYTE editors, based on tests conducted by NSTL, Inc., as documented in a recent issue of NSTL's monthly PC Digest. To purchase a copy of the full report, contact NSTL at 625 Ridge Pike, Conshohocken, PA 19428; (610) 941-9600; fax (610) 941-9950; o
illustration_link (38 Kbytes)
Test Results
