> First you can't reliably measure cpu load on HT enabled CPU's because > maximal summaric performance depends on what instructions are being > executed on sibling cpus in the same time.
Somewhat the point I was offering. Without UT2K4 sucking up tons of CPU time, the SRCDS isn't adversely affected by it. > Second what does it mean 25% for single virtual cpu? That single > phisical cpu is maxed out ? What happens when two virtual cpu's have > reached 25% each ? Phisical cpu is still maxed out but single virtual is > slowed approximatly by half. Take this into account. 25% meaning 1 of 4 virtual CPUs running at "100%". Since UT2K4 (Like everything else) is SMP, it can only use one CPU, theoretically capping it at 25% of the total cpu available. Yes, I know virtual CPUs don't really work the same as a true quad-cpu systems. But if sibling 0 and 1 are both running 100%, it should be safe to say they each have roughly 50% of physical CPU0, and therefore 25% each of the total. If none of that made sense about what I'm trying to say, just know that UT2K4 always ran at 99% CPU even though in-game there wasn't any lag. It just wasted cycles, and without it, I see performance gains elsewhere. The rest of your post is quite similar to what I use for usage (even before shutting down Ut2k4.. Things just seem smoother without it). _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
