Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Gary Stanley
At 04:51 PM 8/28/2008, Nephyrin Zey wrote: >I believe, and I could be wrong, that all recent versions of glibc call >gettimeofday as a virtual syscall, which means the context switching >doesn't occur. IIRC, Only on x86_64. i386 doesn't move gtod into the shared page for userland. >I'm not sure

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Gary Stanley
At 04:51 PM 8/28/2008, Nephyrin Zey wrote: >I believe, and I could be wrong, that all recent versions of glibc call >gettimeofday as a virtual syscall, which means the context switching >doesn't occur. IIRC, Only on x86_64. i386 doesn't move gtod into the shared page for userland. >I'm not sure

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Gary Stanley
At 04:45 PM 8/28/2008, [EMAIL PROTECTED] wrote: >Gary Stanley Wrote: > > Have you tried an older kernel? I don't see those issues (at all). > > However, I have plenty of hacks in place in kernel/time.c to make > > gettimeofday() return for speed, no accuracy (saves a couple mpy and > > divl cycles)

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Gary Stanley
At 04:45 PM 8/28/2008, [EMAIL PROTECTED] wrote: >Gary Stanley Wrote: > > Have you tried an older kernel? I don't see those issues (at all). > > However, I have plenty of hacks in place in kernel/time.c to make > > gettimeofday() return for speed, no accuracy (saves a couple mpy and > > divl cycles)

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Nephyrin Zey
I believe, and I could be wrong, that all recent versions of glibc call gettimeofday as a virtual syscall, which means the context switching doesn't occur. I'm not sure that fixing the overuse of gettimeofday() will solve all these problems, I think there are just some serious issues with the wa

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread octodhd_hlds
Gary Stanley Wrote: > Have you tried an older kernel? I don't see those issues (at all). > However, I have plenty of hacks in place in kernel/time.c to make > gettimeofday() return for speed, no accuracy (saves a couple mpy and > divl cycles), and i don't use 2.6.26 series on my development stuff.

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Gary Stanley
At 03:36 AM 8/28/2008, Nephyrin Zey wrote: >I've complained before about how srcds chugs massive amounts of CPU, but >now that I've enabled SourceTV it's gotten absolutely absurd. Here is my >server idling, while my monitoring system polls it once a minute for CPU >usage. The server is *empty*, wit

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Gary Stanley
At 03:36 AM 8/28/2008, Nephyrin Zey wrote: >I've complained before about how srcds chugs massive amounts of CPU, but >now that I've enabled SourceTV it's gotten absolutely absurd. Here is my >server idling, while my monitoring system polls it once a minute for CPU >usage. The server is *empty*, wit

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Nephyrin Zey
Bah, you're right, I still had my one plugin set to muck with FPS settings. Here's the same deal with all plugins prevented from ever loading: CPU InOut Uptime Users FPSPlayers 12.50 0.00 0.00 1 2 253.74 0 stats CPU InOut Uptime Users FPSPlayers 12

Re: [hlds_linux] srcds CPU usage, again.

2008-08-28 Thread Jon Kessler
Neph what OS are you running,, and why (if I am reading your post right) is your stats reporting a FPS of 3831.42 ? have you tried not setting an affinity to any one core? ~kennycom - Original Message - From: "Nephyrin Zey" <[EMAIL PROTECTED]> To: "Half-Life dedicated Win32 server ma

Re: [hlds_linux] srcds CPU usage, again.

2008-08-28 Thread Peter Lindblom
Our xeon 3060 with 4Gb mem. FC6, is running 4 dod:s servers, 32 slots + 20 slots + 16 slots + 12 slots with sourceTV. The scrim-server with sourceTV have been running for 2 month now and is now using 8% cpu idling. The other 3 servers is using 3-4% idling each ! So I have no problem with 100% cp

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread ics
Your assumption on the bandwidth usage is about right. We have 2 around half a day full 30 slot TF2 servers and in 2 months we have used 1,3 TB of bandwidth (out). If i add external map download bandwidth to that, it lifts that up a bit more. However, there is also a third server (CSS) running

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread gameadmin
AMD Athlon(tm) 64 X2 Dual Core Processor 3800+, SourceTV demo recording, no players connected arena_badlands uname -a: Linux orion 2.6.20-16-generic #2 SMP Fri Aug 31 00:55:27 UTC 2007 i686 GNU/Linux Running ubuntu feisty 32-bit 11:28:02 stats 11:28:02 CPU InOut Uptime Users FPSPl

Re: [hlds_linux] [hlds] srcds CPU usage, again.

2008-08-28 Thread Ryan Mannion
I'm interested in hearing more performance data points. I can't find any useful information on the web aside from really bizarre speculation from people who have no business commenting on server performance. On a Xeon 3060 (2 2.4GHz cores) with 2GB running CentOS 5, I can run, at most two full 32-

[hlds_linux] srcds CPU usage, again.

2008-08-28 Thread Nephyrin Zey
I've complained before about how srcds chugs massive amounts of CPU, but now that I've enabled SourceTV it's gotten absolutely absurd. Here is my server idling, while my monitoring system polls it once a minute for CPU usage. The server is *empty*, with no bots, with just SourceTV on. SourceTV