The ping shown is from city to city in the UK (although current ISP vs Blaster problems make it an unreliable benchmark atm). I think it is possibly higher by default in 3.1.1.1d as compared with 3.1.1.0c - although this may be due to the higher system CPU usage (thereby slowing everything else on the stack down) as much as anything different in the networking code. On the same network 100Base/T I get 5ms (much the same as you do). It goes without saying that once the box hits critical CPU usage (around 96%) pings jump to 200-300ms.
I guess the short answer (rather than waffling on as I have been doing :)) is that I havent tested it enough to answer your question. All I can say is that 3.1.1.1d without pb and sys_ticrate gives pings of around 60-80. 3.1.1.1d with pb 3 and sys_ticrate 10000 gives pings of around 20ms. I havent tested variations on those values and possibly wont have the time too do so. It works for me - I'd be interested to hear if it works for anyone else.
Regards
Simon
NB from memory - 3.1.1.0c values without sys_ticrate and pb 3 gave much the same pings (around 50-60ms) - although a slightly lower CPU usage (still fairly high though).
P�r Olsson wrote:
Simon Alman wrote:
Hmmm pings can be a whole mess of complication - heres my thoughts on linux:
With the default Half Life code running I would agree that linux has higher latencies, however using pingboost 3 uses a different timing system and brings pings back down to "normal" levels.
Without pingboost I would expect about 60ms as the best you are going to get, with pingboost 3, I get between 18-25 ms (running hlds_3.1.1.1d) instead.
On the 3.1.1.0c boffix.v2 I get 3-6 ms ping on boxes in the same city and from 15-50 nationwide. All this on sys_ticrate 1000 and pb 3. Do you think your running the (on this list at least) much hated 3.1.1.1 series causes your high minimum? Without pb 3 and sys_ticrate I get about 25 ms more on average.
Cheers, P�r
_______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
