|
Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home |
. | . |
Prices: Mainboards ABIT ASUS Chaintech Shuttle Soyo Tyan CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ |
LOSTCIRCUITS
|
|
| Intel Pentium4 "Prescott" To Scale or not to Scale, that is our question. | |
|
(Review by MS, May. 23, 2004) |
|
Intel Prescott Starting at: |
Back to Cinebench 2003
Cinebench encompasses a total of 4 different benchmarks, that is Raytracing, C4D Shading, OpenGL Soft and Hardware rendering and a ratio value for the speeding up of OpenGL rendering from software to hardware T&L.
C4D Shading
Prescott clearly wins this one by defeating even the mighty Galatin-based ExtremeEdition. Interestingly, though, the Northwood (periwinkle) appears to scale better than the Prescott (blue) that flattens out at the higher multiplier. The full spectrum of the different systems we ran is shown on the left in reduced format.
OpenGL Software
OpenGL Software Rendering on the other hand sees the Northwood on top and there is no apparent difference in scalability. The full spectrum of the different systems we ran is shown on the left in reduced format.
OpenGL Hardware
OpenGL Hardware rendering is again dominated by the Prescott and the Prescott scales slightly better. The full spectrum of the different systems we ran is shown on the left in reduced format.
OpenGL Speed-up
Even though it is somewhat irrelevant for any standalone analysis, the OpenGL Speedup score is very interesting for comparing similar processors / speedgrades on the same platform with respect to their efficacy of data access. Data access in this case appears to encompass main system memory as well as the differenet levels of cache. Access latencies vs. cache size and clock speed appears to be a major factor for the outcome of the results as well. This is the first benchmark where we really observed a true scaling advantage of the Prescott vs. the Northwood stepping of the P4. More correctly, though, one should say that the performance drop off of the Northwood with higher multipliers is worse than that of the Prescott. Note that the metric is giving arbitrary values and, therefore, we felt that using an non-zero offset for the scale given in the graph can be justified.
next page: => MDK2 =>
All advice and educational articles on LostCircuits are free, but if you feel you can, please make a small donation to us!