Re: [hlds_linux] fps fluctuates every time
which clocksource do you use? - Original Message - From: kERPLUNK k...@datafull.com To: Steven Hartland kill...@multiplay.co.uk; Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, February 24, 2009 8:19 PM Subject: Re: [hlds_linux] fps fluctuates every time yes, i booted with noapm and nothing changes.. - Original Message - From: Steven Hartland kill...@multiplay.co.uk To: kERPLUNK k...@datafull.com; Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, February 24, 2009 5:12 PM Subject: Re: [hlds_linux] fps fluctuates every time Make sure that ALL power management option in the machine BIOS are disabled. If you have any Power management enabled then your server will be totally unpredictable, causing lag. Regards Steve - Original Message - From: kERPLUNK k...@datafull.com all of you have fps fluctuation or someone could fix this? This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. In the event of misdirection, illegible or incomplete transmission please telephone +44 845 868 1337 or return the E.mail to postmas...@multiplay.co.uk. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] server.cfg questions
have you tried compiling the kernel the way i told you? - Original Message - From: Christopher Szabo sza...@hotmail.com To: hlds_linux@list.valvesoftware.com Sent: Monday, February 23, 2009 3:42 PM Subject: [hlds_linux] server.cfg questions Hm. Still got the problem. Didnt help to change sys_ticrate to 333, 500, 1000 or 1. And i tried all sorts of rates, so thats not the problem either. :/ _ Snygga till dina bilder snabbt, enkelt och gratis med PhotoGallery http://download.live.com/photogallery ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] server.cfg questions
Try compiling your kernel with Device Drivers --- Network device support --- [*] Your NIC [*] Use Rx Polling (NAPI) and Device Drivers --- DMA Engine support --- [*] Intel I/OAT DMA support [*] Network: TCP receive copy offload And see if something changes. - Original Message - From: Christopher Szabo sza...@hotmail.com To: hlds_linux@list.valvesoftware.com Sent: Sunday, February 22, 2009 4:57 PM Subject: [hlds_linux] server.cfg questions Hi! I’m running a Counter-strike 1.6 publicserver (32 slots) and I have a problem with my settings. Everybody has a high choke, 30-100. I have tried many rate settings but none is working well. The connection isn’t the problem because it’s a datacenter and other publicservers in the same datacenter don’t have the high choke. Anyway, I looked into my server.cfg and found some things that I don’t really know what they do. max_queries_window -1 max_queries_sec_global -1 max_queries_sec -1 And if that’s not the problem, what could be? I’m using amxmodx, the latest and a few plugins like ptb, hpk, admin_listen and admin_esp. That’s all. Using sys_ticrate 1000, running kernel in 1000 HZ and using pingboost 2. Should I lower the ticrate? Dell Power Edge R300 Dell Intel X5000 Chipset Quadcore Intel Xeon X5470 3.33Ghz 1333Mhz FSB Hynix 2x2GB Dual Rank-RAM 2st Seagate 73GB SAS 15K rpm (Raid-1) _ Kolla mailen när du är på språng - skaffa Hotmail i mobilen! http://windowslivemobile.msn.com/BrowserServiceHotmail.aspx?lang=SE-SE ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Many timeouts
Well, what does the log or the screen-view tell you why it happens? - Original Message - From: Nicolas Fernandez nooky...@gmail.com To: hlds_linux@list.valvesoftware.com Sent: Sunday, February 22, 2009 8:45 PM Subject: [hlds_linux] Many timeouts Hello ppl, i'm running hlds steam in linux box, well the problem is when the all players are playing and get Timeout, the server still runing, but the players can't join in the server getting a Timeout instantly trying to join in the server. Why this happens? Thanks. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] different cpu-cores - different performance
Which Motherboard / Bios do you use? Original-Nachricht Datum: Wed, 28 Jan 2009 12:00:31 +0100 Von: J.Miribel j.miri...@nitroserv.com An: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Betreff: Re: [hlds_linux] different cpu-cores - different performance Does not help, I just recompiled the kernel without ACPI processor support -- 1 core showing up.. Best regards Gary Stanley a écrit : At 04:23 AM 1/28/2009, J.Miribel wrote: Same here. acpi=off as kernel param and I'll have only one core showing up. You can use acpi=ht to turn acpi code on only for multi core functions, but I still get some FPS drops.. Only thing you need to do is uncheck the ACPI Processor, you don't really need to disable ACPI fully. Some broken bios's actually hate ACPI being disabled. You may also wish to try idle=poll to see if that helps with performance. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux -- Pt! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] different cpu-cores - different performance
Your point to the Power Management made me recognize bad evil ACIP, so i added to my grub.conf: kernel /boot/kernel-2.6.25.8 root=/dev/sda3 apm=off acpi=off noapic and voilâ! No more drops. 4 cores with 100% 1000FPS. i can die in peace now. thanks gary and the others! - Original Message - From: Gary Stanley g...@velocity-servers.net To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com; Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 11:46 PM Subject: Re: [hlds_linux] different cpu-cores - different performance At 05:41 PM 1/20/2009, Matthias Bleile wrote: See here: http://rafb.net/p/n7UeOP48.html They are no Players present, doesn't matter, the drops are equal with or without players. If a server is used the Usage of the assigned core rises up to 5-6%, no differences here between the cores. Probably power management or scheduler latency. - Original Message - From: David A. Parker dpar...@utica.edu To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 11:25 PM Subject: Re: [hlds_linux] different cpu-cores - different performance The taskset commands look good. What is the usage on each CPU? htop is a good utility to see that kind of information quickly. Also try: mpstat -P ALL 1 1 This will give you the usage of all CPUs over a one-second interval. Adjust the first number to make it measure usage over a different interval (in seconds). Adjust the second number to make it repeat N times. - Dave Matthias Bleile wrote: Wow, many answers ^^. So: --- Could you post the taskset command lines you are using? taskset -c 0 ./hlds_run. taskset -c 1 ./hlds_run. --- --- cat /proc/interrupts. Probably sharing interrupts. See here: http://rafb.net/p/LN4JtN70.html Only big differents is the timer interrupt only occuring on CPU0. The Rescheduling Interrupts can't be it, it would mean CPU2 is in perfect health, but it isnt. --- --- Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. i can't even turn it on, if i wanted to, cause: Symbol: IRQBALANCE [=n] Prompt: Enable kernel irq balancing Depends on: X86_32 SMP X86_IO_APIC I dont use an 32Bit System. --- --- Useless. It doesn't do what it used to, and biiig performance differents between nice --0 and nice --20. --- --- Kind of useless, but can be useful in specific situations Big difference between no preemption and full preemption. without preempt im not able to get it stable 1000fps even on core0 --- --- Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? same here. core0 = stable 1000 core1 = small spikes core2 = big spikes core3 = only spikes ;) Guess thats it for now ^^. - Original Message - From: 1nsane 1nsane...@gmail.com To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 10:51 PM Subject: Re: [hlds_linux] different cpu-cores - different performance Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? On Tue, Jan 20, 2009 at 4:23 PM, Gary Stanley g...@velocity-servers.netwrote: At 04:04 PM 1/20/2009, Matthias Bleile wrote: Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session cat /proc/interrupts. Probably sharing interrupts. The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Same. Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 i486 doesn't have XMM regs, nor does it have cmov (IIRC) - not used nice, used nice --20 Useless. It doesn't do what it used to, and - Kernel with full preemption - Kernel without any preemption
[hlds_linux] different cpu-cores - different performance
Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 - not used nice, used nice --20 - Kernel with full preemption - Kernel without any preemption, pure processing power. - Turned off Multicore Scheduling, turned back on. - tried tsc, hpet and jiffies as clocksource. If i use Core1+2 its as bad as i use Core3. If i use Core2+3 its more 500FPS than 1000FPS So, heres my question: Has anyone ever encountered something similar, AND sucessfully fixed it. I dont know if its the kernels fault, or an taskset-problem or an Intel problem. Appreciate any help i can give. ps: Dont mind my bad english ;) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] different cpu-cores - different performance
well, as i wrote: 4 hlds Servers == 4 CS 1.6 Servers We got 1 Q6600 and 1 Q9550, both with the exact same issues. - Original Message - From: Matthias Bleile matthiasble...@gmx.de To: hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 10:04 PM Subject: [hlds_linux] different cpu-cores - different performance Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 - not used nice, used nice --20 - Kernel with full preemption - Kernel without any preemption, pure processing power. - Turned off Multicore Scheduling, turned back on. - tried tsc, hpet and jiffies as clocksource. If i use Core1+2 its as bad as i use Core3. If i use Core2+3 its more 500FPS than 1000FPS So, heres my question: Has anyone ever encountered something similar, AND sucessfully fixed it. I dont know if its the kernels fault, or an taskset-problem or an Intel problem. Appreciate any help i can give. ps: Dont mind my bad english ;) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] different cpu-cores - different performance
ok, now i understand hlds can mean anything. so its cs 1.6, and we got - Q6600 - Q9550 both with the same issues. thanks. - Original Message - From: Bruce Potter gd...@shmoo.com To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 10:13 PM Subject: Re: [hlds_linux] different cpu-cores - different performance What game and what type of CPU (Q6600, etc). thanks bruce On Jan 20, 2009, at 4:04 PM, Matthias Bleile wrote: Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 - not used nice, used nice --20 - Kernel with full preemption - Kernel without any preemption, pure processing power. - Turned off Multicore Scheduling, turned back on. - tried tsc, hpet and jiffies as clocksource. If i use Core1+2 its as bad as i use Core3. If i use Core2+3 its more 500FPS than 1000FPS So, heres my question: Has anyone ever encountered something similar, AND sucessfully fixed it. I dont know if its the kernels fault, or an taskset-problem or an Intel problem. Appreciate any help i can give. ps: Dont mind my bad english ;) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] different cpu-cores - different performance
Wow, many answers ^^. So: --- Could you post the taskset command lines you are using? taskset -c 0 ./hlds_run. taskset -c 1 ./hlds_run. --- --- cat /proc/interrupts. Probably sharing interrupts. See here: http://rafb.net/p/LN4JtN70.html Only big differents is the timer interrupt only occuring on CPU0. The Rescheduling Interrupts can't be it, it would mean CPU2 is in perfect health, but it isnt. --- --- Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. i can't even turn it on, if i wanted to, cause: Symbol: IRQBALANCE [=n] Prompt: Enable kernel irq balancing Depends on: X86_32 SMP X86_IO_APIC I dont use an 32Bit System. --- --- Useless. It doesn't do what it used to, and biiig performance differents between nice --0 and nice --20. --- --- Kind of useless, but can be useful in specific situations Big difference between no preemption and full preemption. without preempt im not able to get it stable 1000fps even on core0 --- --- Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? same here. core0 = stable 1000 core1 = small spikes core2 = big spikes core3 = only spikes ;) Guess thats it for now ^^. - Original Message - From: 1nsane 1nsane...@gmail.com To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 10:51 PM Subject: Re: [hlds_linux] different cpu-cores - different performance Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? On Tue, Jan 20, 2009 at 4:23 PM, Gary Stanley g...@velocity-servers.netwrote: At 04:04 PM 1/20/2009, Matthias Bleile wrote: Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session cat /proc/interrupts. Probably sharing interrupts. The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Same. Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 i486 doesn't have XMM regs, nor does it have cmov (IIRC) - not used nice, used nice --20 Useless. It doesn't do what it used to, and - Kernel with full preemption - Kernel without any preemption, pure processing power. Kind of useless, but can be useful in specific situations - Turned off Multicore Scheduling, turned back on. Adds overhead to the scheduler - tried tsc, hpet and jiffies as clocksource. TSC is your best bet on intel, but HPET is overall the better choice. If i use Core1+2 its as bad as i use Core3. If i use Core2+3 its more 500FPS than 1000FPS You are probably running into power management issues, or SMI interrupts causing pipeline stalls. Or you are just running into know CPU errata :-) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] different cpu-cores - different performance
See here: http://rafb.net/p/n7UeOP48.html They are no Players present, doesn't matter, the drops are equal with or without players. If a server is used the Usage of the assigned core rises up to 5-6%, no differences here between the cores. - Original Message - From: David A. Parker dpar...@utica.edu To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 11:25 PM Subject: Re: [hlds_linux] different cpu-cores - different performance The taskset commands look good. What is the usage on each CPU? htop is a good utility to see that kind of information quickly. Also try: mpstat -P ALL 1 1 This will give you the usage of all CPUs over a one-second interval. Adjust the first number to make it measure usage over a different interval (in seconds). Adjust the second number to make it repeat N times. - Dave Matthias Bleile wrote: Wow, many answers ^^. So: --- Could you post the taskset command lines you are using? taskset -c 0 ./hlds_run. taskset -c 1 ./hlds_run. --- --- cat /proc/interrupts. Probably sharing interrupts. See here: http://rafb.net/p/LN4JtN70.html Only big differents is the timer interrupt only occuring on CPU0. The Rescheduling Interrupts can't be it, it would mean CPU2 is in perfect health, but it isnt. --- --- Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. i can't even turn it on, if i wanted to, cause: Symbol: IRQBALANCE [=n] Prompt: Enable kernel irq balancing Depends on: X86_32 SMP X86_IO_APIC I dont use an 32Bit System. --- --- Useless. It doesn't do what it used to, and biiig performance differents between nice --0 and nice --20. --- --- Kind of useless, but can be useful in specific situations Big difference between no preemption and full preemption. without preempt im not able to get it stable 1000fps even on core0 --- --- Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? same here. core0 = stable 1000 core1 = small spikes core2 = big spikes core3 = only spikes ;) Guess thats it for now ^^. - Original Message - From: 1nsane 1nsane...@gmail.com To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 10:51 PM Subject: Re: [hlds_linux] different cpu-cores - different performance Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? On Tue, Jan 20, 2009 at 4:23 PM, Gary Stanley g...@velocity-servers.netwrote: At 04:04 PM 1/20/2009, Matthias Bleile wrote: Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session cat /proc/interrupts. Probably sharing interrupts. The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Same. Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 i486 doesn't have XMM regs, nor does it have cmov (IIRC) - not used nice, used nice --20 Useless. It doesn't do what it used to, and - Kernel with full preemption - Kernel without any preemption, pure processing power. Kind of useless, but can be useful in specific situations - Turned off Multicore Scheduling, turned back on. Adds overhead to the scheduler - tried tsc, hpet and jiffies as clocksource. TSC is your best bet on intel, but HPET is overall the better choice. If i use Core1+2 its as bad as i use Core3. If i use Core2+3 its more 500FPS than 1000FPS You are probably running into power management issues, or SMI interrupts causing pipeline stalls. Or you are just running into know CPU errata :-) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please
Re: [hlds_linux] different cpu-cores - different performance
there you go. Core 0,2 - perfect Core 2,3 - spikes every now and then Core 1,3 - spikes every now and then Core 2,3 - spikes wherever you look - Original Message - From: 1nsane 1nsane...@gmail.com To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 10:51 PM Subject: Re: [hlds_linux] different cpu-cores - different performance Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? On Tue, Jan 20, 2009 at 4:23 PM, Gary Stanley g...@velocity-servers.netwrote: At 04:04 PM 1/20/2009, Matthias Bleile wrote: Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session cat /proc/interrupts. Probably sharing interrupts. The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Same. Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 i486 doesn't have XMM regs, nor does it have cmov (IIRC) - not used nice, used nice --20 Useless. It doesn't do what it used to, and - Kernel with full preemption - Kernel without any preemption, pure processing power. Kind of useless, but can be useful in specific situations - Turned off Multicore Scheduling, turned back on. Adds overhead to the scheduler - tried tsc, hpet and jiffies as clocksource. TSC is your best bet on intel, but HPET is overall the better choice. If i use Core1+2 its as bad as i use Core3. If i use Core2+3 its more 500FPS than 1000FPS You are probably running into power management issues, or SMI interrupts causing pipeline stalls. Or you are just running into know CPU errata :-) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] different cpu-cores - different performance
exactly WHAT is caused by Probably power management or scheduler latency. i dont see anything special in this test.? what are you reffering to? - Original Message - From: Gary Stanley g...@velocity-servers.net To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com; Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 11:46 PM Subject: Re: [hlds_linux] different cpu-cores - different performance At 05:41 PM 1/20/2009, Matthias Bleile wrote: See here: http://rafb.net/p/n7UeOP48.html They are no Players present, doesn't matter, the drops are equal with or without players. If a server is used the Usage of the assigned core rises up to 5-6%, no differences here between the cores. Probably power management or scheduler latency. - Original Message - From: David A. Parker dpar...@utica.edu To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 11:25 PM Subject: Re: [hlds_linux] different cpu-cores - different performance The taskset commands look good. What is the usage on each CPU? htop is a good utility to see that kind of information quickly. Also try: mpstat -P ALL 1 1 This will give you the usage of all CPUs over a one-second interval. Adjust the first number to make it measure usage over a different interval (in seconds). Adjust the second number to make it repeat N times. - Dave Matthias Bleile wrote: Wow, many answers ^^. So: --- Could you post the taskset command lines you are using? taskset -c 0 ./hlds_run. taskset -c 1 ./hlds_run. --- --- cat /proc/interrupts. Probably sharing interrupts. See here: http://rafb.net/p/LN4JtN70.html Only big differents is the timer interrupt only occuring on CPU0. The Rescheduling Interrupts can't be it, it would mean CPU2 is in perfect health, but it isnt. --- --- Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. i can't even turn it on, if i wanted to, cause: Symbol: IRQBALANCE [=n] Prompt: Enable kernel irq balancing Depends on: X86_32 SMP X86_IO_APIC I dont use an 32Bit System. --- --- Useless. It doesn't do what it used to, and biiig performance differents between nice --0 and nice --20. --- --- Kind of useless, but can be useful in specific situations Big difference between no preemption and full preemption. without preempt im not able to get it stable 1000fps even on core0 --- --- Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? same here. core0 = stable 1000 core1 = small spikes core2 = big spikes core3 = only spikes ;) Guess thats it for now ^^. - Original Message - From: 1nsane 1nsane...@gmail.com To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Tuesday, January 20, 2009 10:51 PM Subject: Re: [hlds_linux] different cpu-cores - different performance Try running 2 servers on Core 0 and 2 and the rest on Core 1 and 3. Does that change anything for you? On Tue, Jan 20, 2009 at 4:23 PM, Gary Stanley g...@velocity-servers.netwrote: At 04:04 PM 1/20/2009, Matthias Bleile wrote: Hi everybody. So i wasted 1 week of free time and i still wasn't able to find out.. We got a Intel Quad Core and 4 hlds Servers. Each server is assigned to 1 CPU using taskset. The Server assigned to CPU0 gives excellent performance, as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27030.1232478679.session The Server assigned to CPU1 gives a slightly worse performance as seen here: http://fpsmeter.fragaholics.de/?load=hlds.84.201.7.34.27400.1232478412.session cat /proc/interrupts. Probably sharing interrupts. The Server assigned to CPU2 gives MUCH worse performance as seen here: http://fpsmeter.fragaholics.de/index.php?load=hlds.84.201.7.34.27200.1232479691.session Same. Turn off irqbalancing. It's useless and it doesn't do anything but hurt gameserver hosting. Got no picture of CPU3 but believe me, its worth crying. I tried everything i could imagine. - 32Bit OS, 64Bit OS - Debian, Gentoo - hlds_i686, hlds_i486 i486 doesn't have XMM regs, nor does it have cmov (IIRC) - not used nice, used nice --20 Useless. It doesn't do what it used to, and - Kernel with full preemption - Kernel without any preemption, pure processing power. Kind of useless, but can be useful in specific situations - Turned off Multicore Scheduling, turned
[hlds_linux] CS 1.6 - How does rcon stats calculate the output?
Hello there. First of all, i love this mailinglist :D. Second, please excuse my i try my best-english. And third would be, i always wondered how rcon stats gets its output. Im pretty sure it somehow connects to the /proc/$PID, but thats all i know. Does someone have an precise idea on how it works? Background is, im very very very close on holding the 1000FPS absolutely 100% stable with 10 clients playing, but every 10-20 minutes there is a 1second drop to 500FPS and that ruins all. Its very important i can reconstruct this behavior but i'm simply missing the basic knowledge of CS 1.6 Engine. Im using jiffies as clocksource and i have to figure out how the cs1.6 engine calculates. thanks for any help you can offer. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] sys_ticrate
then he hasn't used pingboost 3 with it. - Original Message - From: Javi Martín [EMAIL PROTECTED] To: hlds_linux@list.valvesoftware.com Sent: Thursday, October 02, 2008 8:03 AM Subject: [hlds_linux] sys_ticrate Good Morning! Yes, the 1-3 ms in Europe is a very little variance in a game-server, but in Spain, where the connections have at times 5 - 20 ms of variance, we need the best, and i want to offer the best. The best for me, is a 1000 fps server, because i test it, and i like it. I saw servers with sys_ticrate 1 - 1000 fps, and the servers has been tested by me without accelerations. Regards. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] sys_ticrate
Can't we compromise in using the phrase: You don't need to use pingboost while using 1000Hz kernel. ? Shouldn't sounds so... your making a big mistake. - Original Message - From: Kveri [EMAIL PROTECTED] To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Thursday, October 02, 2008 7:27 PM Subject: Re: [hlds_linux] sys_ticrate server is accelerated ONLY with sys_ticrate 1 (or maybe 9000 or so too) + pingboost 3 + 1000Hz kernel. You shouldn't use pingboost while using 1000Hz kernel. Kveri Javi Martín wrote / napísal(a): Good Morning! Yes, the 1-3 ms in Europe is a very little variance in a game-server, but in Spain, where the connections have at times 5 – 20 ms of variance, we need the best, and i want to offer the best. The best for me, is a 1000 fps server, because i test it, and i like it. I saw servers with sys_ticrate 1 - 1000 fps, and the servers has been tested by me without accelerations. Regards. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] sys_ticrate
Why is it that bad to run pb3? Just because of the CPU Usage? - We live in 2008 :) Or do you know any disadvantages? - Original Message - From: Kveri [EMAIL PROTECTED] To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Wednesday, October 01, 2008 7:03 PM Subject: Re: [hlds_linux] sys_ticrate DO NOT use sys_ticrate 1. Max sys_ticrate 1000 w/ -pingboost 3. But if you have 1000Hz kernel you're good without pingboost (because pingboost is a workaround for 100Hz kernels and achieves more than 100fps), I'm running 1000Hz kernel with HPET, -pingboost 0 (same as w/o -pingboost) and sys_ticrate 1000 = 995fps stable. Kveri Javi Martín wrote / napísal(a): Hello. I want to set in my servers -pingboost 3 with sys_ticrate 1 without accelerations, to get 1000 fps stable. I saw servers with this, but their owners don´t want to tell me the method XD. Any idea for this? Sorry for my bad english. Kind Regards. Javier. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] sys_ticrate
The acceleration just happens with ticrate 1. so don't use ticrate 1. use 2500 and have your 1000fps stable. if u dont have them stable start optimizing your system. that's the key my friend. - Original Message - From: bl4nkeh [EMAIL PROTECTED] To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Wednesday, October 01, 2008 10:35 PM Subject: Re: [hlds_linux] sys_ticrate You do know that there is no benefit to having anywhere over 250-300fps, right? Anything more than that doesn't really do anything at all. Javi Martín wrote: Hello The question is not the CPU load. The question is, i want the maximal FPS stable, with pingboost 3 and sys_ticrate 1 is posible, but with this we obtain accelerations. I want the method to have 1000 fps stable without this accelerations. Regards. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] sys_ticrate
Of course you get an answer. Right here - http://supportwiki.steampowered.com/wiki/Optimizing_a_Dedicated_Server at the very bottom. - Original Message - From: Gary Stanley [EMAIL PROTECTED] To: Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com; Half-Life dedicated Linux server mailing list hlds_linux@list.valvesoftware.com Sent: Thursday, October 02, 2008 1:45 AM Subject: Re: [hlds_linux] sys_ticrate At 04:35 PM 10/1/2008, bl4nkeh wrote: You do know that there is no benefit to having anywhere over 250-300fps, right? Anything more than that doesn't really do anything at all. Ask valve if anything over 100 is useful. Will you get an answer? No. If it is useless as you say, why did they let it bet set over 100 to begin with? (or even visible to players..) Basically server fps shows you how accurate gettimeofday() is. (with interrupt rate) Javi Martín wrote: Hello The question is not the CPU load. The question is, i want the maximal FPS stable, with pingboost 3 and sys_ticrate 1 is posible, but with this we obtain accelerations. I want the method to have 1000 fps stable without this accelerations. Regards. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux