PERFORMANCE
We tested each system for performance under Windows 3.1, DOS 6.0, and SCO Unix 3.2.4. The DOS and Windows suites consisted of a combination of BYTE low-level tests and NSTL application tests.
The application tests use actual business programs to provide a real-world measure of system performance. The DOS performance suite includes WordPerfect 6.0, Lotus 1-2-3 release 2.4, and FoxPro 2.5. The Windows suite uses WordPerfect 6.0, Microsoft Excel 5, FoxPro 2.6 for Windows, and Word for Windows 6.0. All applications execute macros that exercise common areas of each application. For instance, the Word for Windows test includes subtests that measure activities including file I/O, search-and-replace functions, changing fonts, scrolling by page and line, checking spelling, print previewing, and printing to a file. Windows tests ran
in 1024- by 768-pixel resolution with 256 colors. DOS tests ran in standard VGA resolution (640 by 480 pixels with 16 colors).
The BYTE DOS low-level tests isolate performance of subsystems: CPU, FPU, memory, video, and hard disk. These tests provide important data for analyzing the results of the application-level tests. The BYTE Windows low-level tests exercise the Windows GDI (Graphical Device Interface) to determine how well a system can execute basic Windows graphics tasks, such as drawing a line, displaying text, and executing BITBLT operations.
We also tested each system's performance under SCO Unix 3.2.4. The test suite consists of BYTE's low-level Unix tests and SPEC92, which provides scores for floating-point and integer performance. The Unix tests cover a spectrum of typical scientific and engineering tasks, such as electronic-circuit analysis, architectural analysis, and compilation, as well as general Unix system operations.
The SPEC92 suite that we use covers a variety of
application-based and low-level benchmarks representative of engineering and scientific activities. The integer test (SPECint) contains six CPU integer-intensive benchmarks, mostly written in C. Floating-point benchmarks (SPECfp) contain 14 CPU-intensive floating-point benchmarks, mostly written in FORTRAN. These tests primarily measure the performance characteristics of the processor, cache, and main memory units in processor-intensive applications. They do not attempt to measure display, network, or disk performance. SPEC results are scored as ratios of the time taken for the test compared to the time taken by a DEC VAX 11/780 computer. A score of 42.3 means that the tested system ran that test 42.3 times faster than the VAX. The higher the score, the faster the system for that test.
For Windows and DOS, we scaled all application test scores from 1 to 10, using the best results in this group of systems as a 10. Note: Because this month's tests have been updated with the latest versions of application
software, test results are not directly comparable to those of previous systems Lab Reports.
CONFIGURATION
Our testing was open to all Pentium-class systems and bus architectures. We required that all Pentiums have 32 MB of RAM. Hard drives had to be at least 1 GB, using a controller that had no more than 1 MB of cache memory. The video subsystem requirements included at least 1 MB of video memory and support for 1024- by 768-pixel resolution at 256 colors. Each system was equipped with a CD-ROM drive.
Contributors
Scott Higgs, Project Manager/NSTL, has tested hardware for NSTL for six years. He spent last year in Europe, where he helped establish a testing facility in France.
Alan Joch, Senior Editor/BYTE, coordinates the combined testing between the BYTE Lab and NSTL.
Jim Kane, Project Manager/NSTL, led the testing for this report. He has evaluated high-end systems, peripherals, and network hardware at NSTL during the past three years.
FEATURES
We considered the foll
owing features to be among the most important for Pentium systems:
-- at least one-year parts-and-labor warranty
-- 32 MB of system memory
-- number of adapter slots and drive bays available
-- display resolution of at least 1280 by 1024 pixels
How We Tested
