iMac, Rev. B

The tested Revision B iMac had been updated with the hardware and software updates, and was running Mac OS 8.5.1. The screen was set to 800 x 600 resolution and millions of colors.

The drive was not optimized before benchmarking.

Remember that benchmarks are arbitrary. They measure certain types of performance that may or may not reflect the way you work.

Speedometer 3.06

The system was tested on 6 January 1999 with Mac OS 8.5.1 with most inessential extensions off. The iMac was tested in 24-bit video mode at 800 x 600 resolution. Results are relative to a Mac Classic, which rates 1.0. Numbers rounded off to one or two decimal places.

All tests were run at the default cache for a 32 MB system, which the Memory control panel reports as about 1024KB. Virtual Memory was set at 64 MB.

VM     CPU  graphics   disk    math
64MB  103.1   69.12   13.31   526.1

The CPU score demonstrates that the iMac is two orders of magnitude faster than the first Macintosh.

Speedometer 4.02

The system was tested on 6 January 1999 with Mac OS 8.5.1 with most inessential extensions off. The iMac was tested in 24-bit video mode at 800 x 600 resolution. Results are relative to a Quadra 605, which rates 1.0. Numbers rounded off to one or two decimal places.

All tests were run at the default cache for a 32 MB system, which the Memory control panel reports as about 1024KB. The variable was Virtual Memory, which was off in the first test, set to 33 MB in the second, and bumped to 64 MB in the third.

VM   CPU  graphics  disk    math
 off  15.22   n/a     2.52   644.7
33MB  15.26   n/a     2.64   641.9
64MB  15.13   n/a     2.56   641.9

More than anything else, these results demonstrate that when running a single program that comfortably fits within memory, use of virtual memory makes no significant difference in performance. All results are within 1% of each other.