Elmer, <There have now been 10 benchmarks posted showing a Celeron performing equal to or better than a Pentium 233MMX.>
Elmer, your persistence in Celery advocacy become annoying.
If you could take a close look at the benchmark count,
you would uncover the following facts:
1. Winstone 98 is comprised of 9 actual applications. In
ALL OF THEM the Celery fail against p5-233. How it compares
to p5-200 is carefully hidden. Count - p5 leads 9:0.
2. The BAPCO Sysmark32 is a set of 8 ANOTHER real business
applications (some apps are older releases of apps in
Winstone98, but the coding/libraries are different).
In all 8 applications the Celery lost, count - 17:0.
3. In the High-end WS98, the Celery wins 4 and lost 3 apps;
total count is now 20:4;
4. All other 7 single-task benchmarks are synthetic workloads
that may not represent the real-world performance, see for
example
sysdoc.pair.com
Therefore, in the best case the total count is 20:11, P55-233
leads. Again, since the p200 numbers are intentionally hidden,
it is hard to see how the Celery stands against P5-200MMX.
Based on NINE REAL BUSINESS APPLICATIONS (as per Winstone98
collection),
tomshardware.com
the Celery falls behind the vanilla K6-200.
5. There are apparent signs of tweaking in system configurations
in order to make the Celery look better. For example,
intel.com
the P55-233 score are 15.7 (vs 14.7 for Celery), while
intel.com
the same benchmark scores at 18.2!
What is the difference? Very simple: besides a slower hard
drive, the Celery system was configured with ATI Rage Pro
cards. Looking at ATI web site
atitech.ca@work/xworkagp.html
you may easily find out that ATI Rage Pro card
"optimizes graphics performance of Pentium IIr systems"
and therefore the regular Pentium-MMX will be at
disadvantage in these tests.
In conclusion, stop bragging about your super-Celery.
Period. |
