Techware Labs Header
RSS
Home | Reviews | Articles | Downloads | Guides | Staff | Advertising | Links
Mainboards | Networking | Video | Cases | Storage | Other

Intel's New Family of 800 MHz FSB Processors

Review by Michael Tran on 6.05.03
Managing Editor: Harry Lam
CPU Provided by Intel

 

Performance (Continued):

 

Real World Tests:

 

Comanche 4 Demo:

Processor (1024x768x32) FPS Theoretical % Actual %
2.4GHz 800MHz FSB 38.67 0.0% -
2.6GHz 800MHz FSB 39.51 8.3% 2.2%
2.8GHz 800MHz FSB 40.08 16.7% 3.6%
3.0GHz 800MHz FSB 40.42 25.0% 4.5%

When we compare the trends exhibited in Comanche 4 Demo tests with the 3DMark 2003 tests, we can see that Comanche 4 does not stress the video card as much.  Despite an overall CPU impact of about 17%-28%, we notice that there is a difference of only 1.75 frames per second (which is relatively insignificant).  Since the scores barely increase with increasing CPU speed, the benchmark is video card limited.

Processor (1280x1024x32) FPS Theoretical % Actual %
2.4GHz 800MHz FSB 30.79 0.0% -
2.6GHz 800MHz FSB 31.00 8.3% 0.7%
2.8GHz 800MHz FSB 31.07 16.7% 0.9%
3.0GHz 800MHz FSB 31.11 25.0% 1.0%

When the resolution is increased to 1280x1024x32, the non-existent performance difference at the last resolution is reduced even further.    In the last resolution, there was only a 1.75 frames per second difference between 2.4GHz and 3.0GHz, but at this resolution, that difference plunged to only 0.32 frames per second.  With this insignificant difference between CPU speeds, we can conclude that this benchmark at this resolution is video card limited.

Processor (1600x1200x32) FPS Theoretical % Actual %
2.4GHz 800MHz FSB 23.67 0.0% -
2.6GHz 800MHz FSB 23.68 8.3% 0.04%
2.8GHz 800MHz FSB 23.69 16.7% 0.08%
3.0GHz 800MHz FSB 23.69 25.0% 0.08%

Since the differences in performance have practically hit a plateau, this test is obviously video card limited.

 

Jedi Knight II:

Processor (1024x768x32) FPS Theoretical % Actual %
2.4GHz 800MHz FSB 130.4 0.0% -
2.6GHz 800MHz FSB 137.6 8.3% 5.5%
2.8GHz 800MHz FSB 144.4 16.7% 10.7%
3.0GHz 800MHz FSB 148.9 25.0% 14.2%

Due to the fact that the CPU impact is from 57%-66%, this game is much more CPU dependent than video card dependent.  The actual percentage is relatively close to the theoretical percentage, therefore the test is CPU limited with a slope of about 5% per 200MHz increase.  At this resolution, the performance is a bit overkill and additional features such as anisotropic filtering or FSAA should be enabled.

Processor (1280x1024x32) FPS Theoretical % Actual %
2.4GHz 800MHz FSB 124.6 0.0% -
2.6GHz 800MHz FSB 128.9 8.3% 3.5%
2.8GHz 800MHz FSB 131.4 16.7% 5.5%
3.0GHz 800MHz FSB 134.8 25.0% 8.2%

Again, the CPU is still the limiting factor but it performance seem to be decreasing with an average slope of 2.7% per 200MHz increase.  Since the improvements decreased within this test, the impact of CPU decreased to 32%-42%.  At this resolution, the performance is still excessive and additional features such as anisotropic filtering or FSAA should also be enabled.

Processor (1600x1200x32) FPS Theoretical % Actual %
2.4GHz 800MHz FSB 97.2 0.0% -
2.6GHz 800MHz FSB 97.2 8.3% 0.0%
2.8GHz 800MHz FSB 97.2 16.7% 0.0%
3.0GHz 800MHz FSB 97.2 25.0% 0.0%

This test is video card limited at this resolution because the results showed no improvement.  Running at 97.2 FPS, this game is very playable but enabling addition options may decrease performance below 60 FPS.

« Previous Page Next Page »