Pentium III 500 Platform Comparison

Pages: 1 2 3 4 5 6 7 8

I/O Subsystem Performance

One of the biggest bottlenecks in any system is the I/O performance. This is because all other subsystems measure their cycle times in nanoseconds, while the I/O subsystem cycle time is milliseconds. To put this in perspective, one nanosecond is to a millisecond as one second is to 11.5 days!

Once again, Winbench99 and Passmark Performance Test were used for this comparison. Though HDTach was considered, at the time these tests were run the IDE drivers for the i820 chipset were not available and HDTach was reporting some very lopsided results in favor of the BX chipset. In order to keep the comparison fair, no busmaster drivers were installed on any of the test systems.

Winbench 99

AX6C/DRDRAM

AX6BC/SDRAM

AX63/SDRAM

AX63/VCSDRAM

AVS Express 3.4

3850

5290

5720

5720

Front Page 98

31400

30000

29100

31600

Microstation SE

8510

9740

9170

9720

Photoshop 4.0

2450

2730

2740

2820

Premiere 4.2

4070

3840

4440

4500

Sound Forge 4.0

4110

4980

5200

5290

Visual C++ 5.0

6210

6390

5750

6370

High-End Overall

4770

5240

5490

5690

Business Overall

2310

2570

2600

2720

Passmark

AX6C/DRDRAM

AX6BC/SDRAM

AX63/SDRAM

AX63/VCSDRAM

Cached Read

3.6

3.4

4.3

4.8

Cached Write

3.0

3.2

3.9

4.1

Random Read

16.4

15.1

19.5

20.8

Disk Mark

61.0

57.3

73.2

78.2

In this set of tests, we can see that the situation is almost exactly reversed from the graphics subsystem test. Here the Pro133 chipset is faster in most cases than either the i820 or the BX chipset, even with SDRAM. When combined with Virtual Channel SDRAM, the Pro133 chipset truly excels in this area. At the other end of the spectrum is the i820 chipset, indicating either that Intel’s Hub architecture is in need of some tuning in this area. Though not shown here, the Membench tests did show the Pro133 chipset a clear winner in the area of disk reads, but that the i820 was faster with disk writes.


Pages: « Prev   1 2 3 4 5 6 7 8   Next »

Be the first to discuss this article!