Benchmarks

1. The Full Results
2. Before and After
3. K6-2 : The Middle Man
4. Super7 - Super Enough?
5. The Configurations

Back to Special Pages

1. The Full Results

I wanted to see the difference between the "before" and "after" performance, so I benchmarked the PA-2012 with the original K6-200 and with the replacement K6-III-400. I also benchmarked it with a K6-2-350 at 366MHz, just to get some comparisons.

In addition, I wanted to compare the performance of the PA-2012 and K6-III against a full Super7 100MHz system. To do this, I benchmarked the PA-2013 with the K6-III. The other 2 PCs were older 486 machines and required a complete upgrade, so my friend's boss asked me to replace them with PA-2013-based systems. The ones I got have 2MB of L3 cache on board! The PA-2013 is FIC's latest Super7 100MHz ATX board and the latest revisions can support up to the K6-III-500.

I used ZD's Winstone 98 and ran the Business Winstone suite. All tests were run with Write Allocation enabled. After each test, I defragged the hard disk and rebooted the system, so that each test would have the same starting point.

Here are the full results :-

Back to Top

2. Before and After

On the Business Winstone, the PA-2012 with the original K6-200 achieved a score of 17.8. After changing the CPU to a K6-III-400, the score increased to 26.4. This is a 48.3% improvement in overall performance!

Back to Top

3. K6-2 : The Middle Man

The K6-2-366 on the PA-2012 has a score of 21.4. This is an 20.2% increase over the K6-200, while the K6-III-400 improves the K6-2 score by a similar 23.4%. So the K6-2-366 is the mid-range performer.

Back to Top

4. Super7 - Super Enough?

This is probably the most interesting result of all. This is because it provides an indication of whether all my work is actually worthwhile. The question would be something like this; "Why bother modifying an old mobo like the PA-2012 to run the K6-III when the performance will probably be much less than a newer Super7 100MHz mobo?".

With everthing-else being equal, the PA-2013 with 2MB L3 cache and 64MB PC-100 SDRAM improved the PA-2012's score by a minimal 5.7%!

How's that for an answer?

Back to Top

5. The Configurations

My test system configuration is as follows :-

Back to Top

Back to Special Pages

Disclaimer

Adrian
Last revised: April 19, 2000.