The IIci uses RAM for video, which it takes from the first bank of
memory on the motherboard. Word on the street has always been that
using a separate video card improves CPU performance, on assertion I
wanted to test.
Another is that the 32 KB cache card makes about a 15% difference in
performance.
The tested machine had 20 MB of RAM and a 540 MB Apple-branded
Quantum 540S hard drive, which was 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
Speedometer 3 was not run on this system.
Speedometer 4.02
The system was tested on 12 May 1999 under System 7.5.5. Computer
attached to a 14" Apple color monitor with 8-bit video, except as
noted. 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 disk
cache settings without the level 2 cache card installed.
cache video CPU graphics disk math
32KB 8-bit 0.38 0.39 1.64 1.25
64KB 8-bit 0.38 0.39 1.63 1.24
128KB 8-bit 0.38 0.39 1.62 1.25
256KB 8-bit 0.38 0.39 1.35 1.25
The cache setting should have little influence on non-disk tests,
which these numbers bear out. With this particular setup, cache size
makes no appreciable difference up to 128KB. At 256 KB it drops almost
20%, which is unexpected.
For the next test, the 32 KB Apple cache card is installed.
cache video CPU graphics disk math
32KB 8-bit 0.40 0.46 1.68 1.37
64KB 8-bit 0.40 0.46 1.69 1.37
128KB 8-bit 0.40 0.46 1.68 1.37
Speedometer 4 shows no perceptible difference at these cache
settings. CPU scores are about 5% higher than without the cache,
graphics scores are up about 17%, math scores about 10%, and even disk
scores are up a couple percentage points.
The next test is with a 64KB Daystar FastCache card.
cache video CPU graphics disk math
32KB 8-bit 0.40 0.47 1.73 1.36
64KB 8-bit 0.40 0.47 1.74 1.37
Results are not significantly higher than with Apple's 32KB card,
although disk performance is slightly better -- about 3%. So the next
test is run with Daystar's PowerMath turned on (in the FastCache
control panel).
cache video CPU graphics disk math
32KB 8-bit 0.40 0.48 1.74 2.24
64KB 8-bit 0.40 0.48 1.74 2.24
I don't know how PowerMath patches the SANE routines, but the gain
is over 60% on math scores. This also slightly boosts graphic
performance.
For the following tests, the IIci was used with an Apple portrait
monitor and a Sonnet Presto 040. The point of this test was to measure
the benefit of using a video card instead of internal video. The 14"
8-bit results are shown for comparison. (Tests performed with Sonnet
Presto accelerator installed.)
video video CPU graphics disk math
14" 8-bit 1.49 1.28 1.57 10.04
card 2-bit 1.37 0.58 1.81 10.04
int. 4-bit 1.32 1.29 1.61 10.03
Although the CPU score is about 4% higher and the disk score is
about 12% higher, the significant penalty incurred using the Radius
video card seems to make for a poor tradeoff. Unless the IIci is being
used with an accelerated video card, the serious degradation in graphic
performance more than offsets the slightly improved CPU and disk
scores.
For benchmark results with the Sonnet Presto 040 accelerator,
click here.
Go to the Mac IIci
profile.
