The IIfx uses a 40 MHz 68030 CPU with a 32 KB SRAM cache, so
performance is faster than that of any other 68030-based Mac. The drive
is an 80 MB Apple-branded Quantum ProDrive LPS formatted with Apple HD
SC Setup 7.3.5.
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 19 October 1998 under System 7.5.5 with all
inessential extensions off. Computer attached to a 640 x 480 monitor
and tested in 8-bit video mode except as noted. Results are relative to
a Mac SE or Classic, which rates 1.0. Numbers rounded off to two
decimal places.
The first set of numbers compares performance at different cache
settings.
cache video CPU graphics disk math
32KB 8-bit 11.45 8.72 1.97 21.04
64KB 8-bit 11.53 8.57 1.93 21.00
128KB 8-bit 11.45 8.70 1.89 21.00
256KB 8-bit 11.45 n/a n/a n/a
Speedometer locked up trying to run the graphics test with a 256 KB
cache, which also prevented it from running the disk and math tests.
The cache setting have little influence on non-disk tests. With this
particular setup, cache size makes no appreciable difference, but that
may be due to limited throughput of the drive itself since the IIfx
reports exactly the same data as the slower IIcx using the same (transplanted) hard drive.
The second set of numbers compares performance at different video
settings on the Apple 8•24GC video card, which was run in
unaccelerated mode.
cache video CPU graphics disk math
64KB 1-bit 11.45 10.44 1.91 21.09
64KB 2-bit 11.45 9.09 1.89 21.09
64KB 4-bit 11.45 8.58 1.91 21.09
64KB 8-bit 11.53 8.57 1.93 21.00
64KB 24-bit 11.45 8.44 1.91 20.95
With the 8•24GC card, performance in 4-bit, 8-bit, and 24-bit
modes is virtually identical. Dropping to 2-bits increases performance
6% over 8-bit mode, while using 1-bit mode bumps performance almost
22%.
Speedometer 4.02
The system was tested on 19 October 1998 under System 7.5.5 with all
inessential extensions off. Computer attached to a 640 x 480 monitor
and tested in 8-bit video mode. Results are relative to a Quadra 605,
which rates 1.0. Numbers rounded off to two decimal places.
The first set of numbers compares performance at different cache
settings.
cache video CPU graphics disk math
32KB 8-bit 0.71 n/a 0.75 2.87
64KB 8-bit 0.71 0.45 0.74 2.97
128KB 8-bit 0.71 n/a 0.74 2.99
256KB 8-bit 0.71 0.45 0.79 2.99
The cache setting should have little influence on non-disk tests,
which these numbers bear out. As above, with this particular setup,
cache size makes no appreciable difference, probably due to the limited
performance of the hard drive.
Speedometer 4 shows no perceptible difference at different video
settings.
MacBench 3
The system was tested on 30 November 1998 under System 7.5.5 with
all inessential extensions off. Computer attached to a 14" color
monitor and tested in 8-bit video mode at 640 x 480 resolution on an
8•24GC video card. The disk cache is the variable on the first set
of tests. Results are relative to a Power Mac 6100/60, which rates 10.
Numbers rounded off to two decimal places.
test CPU math disk graphics
32KB 2.39 0.37 12.54 4.09
64KB 2.39 0.37 13.33 4.08
128KB 2.39 0.37 14.02 4.06
256KB 2.39 0.37 14.74 4.04
512KB 2.39 0.37 17.69 4.06
1024KB 2.39 0.37 18.54 4.10
2048KB 2.39 0.37 21.37 4.07
These figures demonstrate that cache size makes no appreciable
difference to CPU, math, or graphics performance - but a large enough
disk cache can significantly improve hard drive performance. At least
with MacBench 3, going from the minimum 32KB cache to 2 MB improved
throughput by 70%. In fact, each increase in cache size improves
performance. This stands in stark contrast to Speedometer, where cache
size made no appreciable difference.
What about Mac OS 8.1?
On 31 August 2000, I tested my IIfx with Mac OS 8.1 (installed on
an external drive using Born Again),
a 256 KB disk cache, and Speedometer 4.0.2. This was the same system
tested earlier with Mac OS 7.5.5. I retested on 1 Sept. 2000 after
realizing that the System 7.5.5 setup included Speed Doubler.
setup CPU graphics disk math
int. drive/System 7.5.5 0.71 0.45 1.39 2.98
ext. drive/System 7.5.5 0.71 0.45 1.41 2.98
int. drive/Mac OS 8.1 0.71 0.45 1.35 2.99
ext. drive/Mac OS 8.1 0.71 0.45 1.47 2.99
With the tested setup, Mac OS 8.1 offers virtually identical
performance, except for 10% better hard drive scores with this external
drive. Considering the added features of OS 8.1, if you're still using
a IIfx and have enough memory, it seems worth the upgrade.
Go to the IIfx
profile.
