Macintosh Benchmarking Factors
The listed factors have been found to affect each category of benchmark.

Video
  • Resolution
  • Color Depth
  • RAM/Cache speed
  • CPU speed
  • Bandwidth to VRAM (frame buffer)
  • QuickDraw acceleration (INIT/hardware)
  • GWORLD memory?
  • Cursor resting over the test area
  • Softwares which cache the ROM, such as DayStar's QuadraBoost

SCSI
  • Disk software driver
  • SCSI bus width
  • SCSI revision/features
  • SCSI controller bandwidth to host
  • Parity?
  • Blind writes?
  • CPU speed (bus speed?)
  • CPU architecture
  • Hard disk tech/bridge type (type of bridged media)
  • Physical prescense of other devices, even powered off.

CPU
  • NewerTech VSO IIfx 2.0 control panel inhibits some QuickDraw tests from running. Setting the speed then soft-restarting without the control panel allows these tests to run at the desired clock speed.
  • Slight but statistically significant impact discovered for certain benchmarks when the MediaVision ProAudioSpectrum16 is installed and drivers are loaded (IIfx in tested case; more testing needed.)

FPU

RAM
  • Softwares which cache the ROM, such as DayStar's QuadraBoost
  • Bus speed (CPU speed)
  • Enabling virtual memory cuts the read speed roughly in half. Memory write speed seems unaffected.
  • Especially in machines with L2 cache enabled, test procedures should be run in the same order and with the same number of repetitions for most accurate comparison.