12. Same 1280*1024 resolution at max settings but with sound this time:
XP 2000
Sempron 2400

13. As always the final and most demanding resolution is 1600*1200; no sound first
XP 2000
Sempron 2400

14. And with sound:
Xp 2000
Sempron 2400

It is clear that a processor with a 333 MHz FSB is faster than one with a 266 MHz one; the question is how much faster is Sempron 2400 over T-bred 2000?
The answer is: not by much! Unlike the Thunderbird 1200C which is much faster in every test than 1200B the difference here is pretty low!
One thing to keep in mind is that in Aquamark 3 the trend continues: not much difference in tests without sound; over 5% in all tests with sound!

15. We go back in time with 3dmark 2000, a DirectX 7 test that is completely CPU bound as seen in the results - the scores increase as the resolution increase and not the other way around:
1024*768 default settings is first
XP 2000
Sempron 2400

16. 3dmark 2000 - 1280*1024 resolution

17. 3dmark 2000 at 1600 *1200
XP 2000
Sempron 2400

At every resolution Sempron is in front by 4-5%, so again within the limit of error! It seems that increasing the FSB to 333 does not benefit much to the K7 architecture!

18. Next is 3dmark 2001 - one of the most balanced test that takes everything into account: CPU, RAM, GPU and the platform which makes them all "sing in tune"
1024-768 resolution:
XP 2000
Sempron 2400

Finally FSB 333 enables Sempron the first clear cut victory! 8.5% better than its smaller brother with 266 FSB.

19. Switching resolution to 1280*1024 and max settings the advantage of the increased FSB remains: Sempron is 8% better than T-bred!

20. At the 1600*1200 resolution and max details there is still a 5% advantage courtesy of the increased FSB!
XP 2000
Sempron 2400

21. We enter the DirectX 9 era with 3dmark 2003. Actually only the mother nature test requires a DirectX 9 GPU, the rest run just fine on any DirectX 8 card!
As always the first is the 1024*768 resolution at default settings:
XP 2000
Sempron 2400

22. The delta between the 333 MHz CPU and the 266 MHz one is pretty small: 3%.
Increasing the resolution to 1280*1024 max details shrinks that small gap to only 2.5%!

23. At 1600*1200 the difference becomes negligible: 1.4%
XP 2000
Sempron 2400

It seems that 3dmark 2003 makes no difference between a CPU with a 266 MHz FSB and a 333 MHz one! All results are well within the margin of error!
However ....

24 & 25 & 26. 3dmark 2003 includes tests with sound:
- no sound
- 24 sounds
- 60 sounds
There is a huge advantage of the Sempron over the Athlon XP in all 3 tests:
no sound: 23%
24 sounds: 15%
60 sounds: 16%
So, clearly, having a CPU with a faster FSB does have its advantages!

27. 3dmark 2005 is the last version to be tested. Exceptionally GPU bound it is a good slideshow with any nvidia AGP card than a real test.
1024*768 - at this resolution with standard settings there is a 4.7% increase by going with a 333 MHz FSB CPU at the same clock speed.
XP 2000 first
Sempron 2400 second

28. Increasing the resolution to 1280*1024 and the level of detail to maximum and the difference between the 2 CPU's has become mostly academic:

29. At the 1600*1200 resolution the geforce 7800gs is begging for mercy as would the 7600gt or any nvidia AGP card for that matter!
3dmark 2005 requires a far more powerful system with at least a geforce 8800GTX to break past the video card bottleneck!