Navigate: 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 Search Prices:

In order to go on, the next thing we need is to look at what really happens in both applications, that is 3DMark2001SE and TS5.1 and then put things into perspective.

Performance Hit
The chart shows the scores we achieved with 3DMark2001SE while TrueSpace5.1 was rendering in the background. The lower columns show the performance hit compared to when 3DMark was running by itself. The HyperThreading enabled system takes the biggest hit with 51.5% loss in performance compared to 46.6% when HT was disabled.

The numbers above show HT at a severe disadvantage, however, this is exactly the reason why we went through the little exercise with the car calculations earlier. In other words, for what we are looking at, the 3DMark2001 score does not matter at all, what we are interested in is the overall runtime of both applications and here we see a rather different picture.

The striped blue columns give the runtime in seconds for 3DMark2001. It is important to realize that there is only a minor variance between the different systems. The yellow columns in the back give the overall runtime in seconds for both applications and here it is where we see a dramatic performance increase for the HT enabled system. Once again, we show the numbers for the SMP system not only as reference but also to underscore the validity of our claims.

To summarize, what we just described: 3DMark2001SE takes a greater performance hit from HT enabled than when it is disabled, we interpret this as 3DMark being unable of taking advantage of MP or HT systems and, thus, operating more efficiently if it does not have to toggle between logical or physical processors.

TrueSpace5.1, on the other hand is, under these conditions, fully capable of playing its multithreaded nature and to very efficiently compete for CPU resources. Overall, the total runtime for both applications is 28% faster with HT enabled. Please keep in mind that the times were hand-stopped, there may be variations in the decimals, otherwise, we believe the results are quite accurate.

next page:    => Some Trivial Benchmarks For Dessert =>