I'm confused as to why you only use an
Intel benchmarking utility to score all the CPU's. On occasion, I find it
misleading.
According to your site, the 1.2GHz Celeron
beats out the 1GHz Athlon 4
and the 1.1GHz Duron. This confuses
me because, according to Tom's Hardware's comparison of the
1.2GHz Celeron and the 1.1GHz Duron (which your site has a link
to), the Duron comes out on top of the Celeron.
I have noticed other situations similar to this, such as Tech Report
(again linked from your site), which shows
the 1.4GHz Athlon edging past the
1.8GHz P4. While this is a much
subtler example due to the small gap between the two, on your site it
still shows a contradiction.
So, why use an Intel-designed benchmark? Why not a third party
benchmark utility?
Intel's iCOMP Index is not a benchmarking
utility, but a scoring system based on a suite
of benchmarks. The iCOMP Index has long history, developed by
Intel first for their i386 processors, then extended through their i486,
Pentium, Pentium MMX, Pentium Pro, Celeron,
Pentium II, and Pentium III CPU lines. During this period, Intel
wisely saw the value of educating its customers about the performance
differences among its various processor families, and the iCOMP Index
did this very well.
When AMD and other processor manufacturers began to make significant
inroads into the CPU market with x86-compatible chips, The CPU Scorecard
was started to provide relatively consistent speed comparisons between Intel's
processors and everybody else's. The iCOMP Index was chosen as the
standard scoring system because it was familiar to PC buyers and had a
proven testing history behind it.
The CPU scores we provide are not directly derived from the same suite
of benchmark utilities used by Intel to produce their iCOMP Index
scores. The scores are merely placed on the same relative scale, based on
comparative tests published by the various
processor manufacturers and other reputable
sources. While the iCOMP Index was designed by Intel, the benchmark
utilities used to derive the scores are entirely provided by
third-party entities.
There are two things to be aware of when comparing iCOMP Index
scores:
Each score is an overall indicator, representing a
combination of integer, floating-point, 3D, multimedia, and internet
performance. Depending on your application (office software, gaming,
audio/video editing, etc.), some processors may perform slightly
better or worse than their iCOMP score might suggest.
As further testing information is compiled from various system
comparisons using different memory configurations, benchmark utility
versions, and incremental manufacturing improvements to the processors
themselves, the derived scores may change slightly from time to time.
Due to what may be small differences in performance, one processor can
rank ahead of another at one point, and yet be ranked lower later on
when new comparison tests are compiled.
In the specific benchmark tests referenced in your question, you will
notice that one processor did not consistently perform better than its
competitor in all the published tests. Our iCOMP Index-based scores
should only be taken to represent a particular CPU's performance in a
majority of computing tasks.
Note that, while this speed index is an important factor in helping
decide which is the faster processor and computer for you, other system
configuration options (memory speed, motherboard chipset, graphics card,
hard drive, etc.) can also significantly influence real-world performance.
Your results, therefore, may vary.
The CPU
Scorecard assumes no risk or liability for damage or loss due to
the use of the information or advice provided here. All responses are
based on the best available information at the time of writing. However,
users of this information who wish to apply it to their computer
situations do so at their own risk.
