Sekra 10 Posted January 8, 2011 (edited) I still have to say that this test will only show how your server will run only with a shitload of AI and we all know that once the AI thread is "capped" you cant go further. So what you are testing is sort of the "top speed on a straight track" but to win a race you need good cornering too. This tests only the capability to run a certain amount of AI on the server which is a part of the testing too, but for instance if you play only PvP missions that have no AI but a bunch of scripts? The results might be quite different. I do remember seeing someone do tests on a server on server FPS vs amount of AI. Just makes me wonder, what is the number of AI on this mission? What happens if you put only 50% of AI on this mission? or 200%? Edited January 8, 2011 by Sekra Share this post Link to post Share on other sites
Kochleffel 10 Posted January 8, 2011 Tested our DAO Server´s Today, I7 920 @ 3,2Ghz 8gb Ram Windows Server 2008 HT on: lowest 15 highest 18 HT off: lowest 16 highest 18 we run usually HT on, and 3 Gameserver´s in BOX (it makes no Difference if u run 1 or 3 Server´s tested to.) Share this post Link to post Share on other sites
-EC- Maciek 10 Posted January 8, 2011 (edited) Updated, Just makes me wonder, what is the number of AI on this mission? What happens if you put only 50% of AI on this mission? or 200%? AI number is around 500. Server on: 200% - 8 fps 100% - 16 fps 50% - 46 fps DDR speed: 800 MHz - 14 fps 1066 MHz - 14 fps 1333 MHz - 15 fps Edited January 9, 2011 by [EC] Maciek Share this post Link to post Share on other sites
Karvanoppa 10 Posted January 9, 2011 (edited) Hello. I just tested our servers and here is specs and results. Server #1 =Olut= BIG A2+OA ACE2b =Olut= BIG A2+OA System: Dell Poweredge Xeon X3420@2,4GHz - 8G ram (don't know witch type or speed) Ubuntu 9.10 - Kernel 2.6.31-22server 64bit Server #2 =Olut= Test platform Intel E6400@2,13GHz - 3G DDR2 800MHz This are Debian based virtual server with Openvz container. Setup is 2cpu and 2900MB ram. Ubuntu 10.04 - Kernel 2.6.32-4-pve Results: Server1: Two arma servers online. Another server playing dominations when i test this. = minimun FPS 7 with 10 average. Server2: Minimun FPS and avarage are same and that is 8FPS. Thanks Edited January 10, 2011 by Karvanoppa Share this post Link to post Share on other sites
n00b1 10 Posted January 12, 2011 This may seem like a retarded question but what kind of influence does the server performance at this level have on the client performance? I just built a new Sandy Bridge system and the game runs like butter on Single Player and online with only a few players (even on Warfare mode). But on a well populated server the framerate is killed and there is nothing I can do to improve it! Share this post Link to post Share on other sites
gL33k 0 Posted January 13, 2011 damn it , i just look at the graph. is 18fps enough for AI behaviour ?? Share this post Link to post Share on other sites
[kh]jman 49 Posted January 13, 2011 (edited) Yes 18 fps is fine. Is that on full AI load on the test mission? Edited January 13, 2011 by [KH]Jman Share this post Link to post Share on other sites
-EC- Maciek 10 Posted January 13, 2011 1) I know from my expirience that, when u get fps on server lower than 8, it will start slow down graphics. 2) we play one mission 8 ppl alnd 500 AI (9 fps on server) Share this post Link to post Share on other sites
[kh]jman 49 Posted January 13, 2011 http://www.kellys-heroes.eu/forum/viewtopic.php?f=5&t=2701#p22901 Lowest 10 fps Share this post Link to post Share on other sites
n00b1 10 Posted January 13, 2011 But does this slow down client framerate too or just make people judder about? :confused: Share this post Link to post Share on other sites
gossamersolid 155 Posted January 15, 2011 Who's with me in saying that I hope BIS works on better multi-core support and performance updates for the ArmA 2 server software. I find it really annoying that my 4 core server tops out at 30% CPU usage on an arma 2 server, but then the Arma 2 server starts choking with low FPS. I have more resources and I wish it'd use them. Share this post Link to post Share on other sites
Karvanoppa 10 Posted January 16, 2011 This is not needed to upgrade in list, but i tested just for intresting my game pc speed for this. Specs: Intel Q9550 @ 3,8GHz | 4GB 1066 DDR2 Results: lowest 16fps avarage 18fps Share this post Link to post Share on other sites
gossamersolid 155 Posted January 16, 2011 9FPS Min Xeon Quad @ 2.27ghz 4GB DDR3 Win Server 2008 R2 x64 Share this post Link to post Share on other sites
visceralsyn 10 Posted January 16, 2011 Goss, Something is wrong there, My Xeon E3110 Wolfdale @ 3.0ghz ( dual core ) with 4GB of DDR²@800mhz running Mandriva 2010.1 ( 2.6.31.13 )got 11fps. ...Syn... Share this post Link to post Share on other sites
Hud Dorph 22 Posted January 16, 2011 Goss, Something is wrong there, My Xeon E3110 Wolfdale @ 3.0ghz ( dual core ) with 4GB of DDR²@800mhz running Mandriva 2010.1 ( 2.6.31.13 )got 11fps. ...Syn... Why ?? you have more GHz. - dual vs quad dont mean much as Arma2-server has enough in 2 cores. Share this post Link to post Share on other sites
djczing 10 Posted January 17, 2011 Max 11, Min 9, Avg @ 10 FPS Redhat Fedora (2.6.27 kernel) linux Q6700 @ 2.6Ghz (no OC) 4gb DDR2 XFX Mobo Share this post Link to post Share on other sites
gossamersolid 155 Posted January 17, 2011 Goss, Something is wrong there, My Xeon E3110 Wolfdale @ 3.0ghz ( dual core ) with 4GB of DDR²@800mhz running Mandriva 2010.1 ( 2.6.31.13 )got 11fps. ...Syn... Like Dorph said, once you have two cores, it's purely clock speed. Share this post Link to post Share on other sites
mosh 0 Posted January 17, 2011 Gaming computer below in specs got 15 FPS or so, varied a little each test 14 was min, 18 max out of all tests. I got 1 or 2 more FPS by running from SSDs (Agility slightly faster than Intel, even with OS on Agility) vs WD Blacks (faster even being on OS Black drive than non OS other types of HD)... biggest difference was obviously loading times and speed of admin functions with server on SSD. I do think I might have tweaked 1 or 2 more FPS by shutting down services, optimizing, etc... but didn't really care, just tested for my curiosity. Main server - AMD 6400 X2 (3.2) - 4 GB DDR2 - WD Black HD - Windows 7 64 min 11 max 14 average 13 I tried optimizing this one as much as I could, and that was about best I could tweak out of it... though it is also a gaming computer, not just server. Share this post Link to post Share on other sites
larsiano 12 Posted January 30, 2011 my result: Battlegroup-=XTRA=- Server -=XTRA=- 2 x E5420 2.5Ghz quadcore CPU Asus 5000V mobo 6 x 2GB DDR2 667MHhz FB-DIMM + noisy mem cooler 2 x 60GB Vertex 2 @ RAID0 on Adaptec 5420 SASRAID card Windows 7 x64 Ultimate UK -=XTRA=- min 11 max 12 average 11 Share this post Link to post Share on other sites
=wfl= sgt bilko 10 Posted January 31, 2011 (edited) [GCA] Linux Test Server: i7-920 @ 2.66GHz (quad core with HT enabled) 8GB DDR3 Ubuntu 9.10 (2.6.31-16 x86_64) With a2oa server v1.57.77737 -cpucount=2 FPS (min/max/avg): 13/16/13 Max mem: 137MB Max Out: 152kbps -cpucount=4 FPS (min/max/avg): 14/16/15 Max mem: 574MB Max Out: 164kbps Comment: Heli fly by looks stuttering/warping. Edited January 31, 2011 by =WFL= Sgt Bilko Share this post Link to post Share on other sites
-EC- Maciek 10 Posted January 31, 2011 Will update tomorrow :) Share this post Link to post Share on other sites
DarkEclip 10 Posted February 3, 2011 (edited) [22eR] Quebec: Intel® Xeon CPU 3.20GHz, 4 cores 75.00$can /m 4GB DDR3 Linux 2.6.18-194.32.1.el5PAE sur i686 With a2oa linux server v1.57.77737 @ACE2, @WARFX, @JSRS -cpucount=4 -exthreat=7 -maxmen=2048 FPS stuck at 6 FPS Max mem: 436MB Max Out: 152kbps I think we have a big problem :( Edited February 3, 2011 by DarkEclip Share this post Link to post Share on other sites
LondonLad 13 Posted February 5, 2011 (edited) [server Browser Name] [16AA] Private Server #2 [server Specs] HP Blade BL460 (G1) 2 * Intel Xeon® CPU E5420 @ 2.5GHz (2 Physical Processors, 8 Virtual) - Quad Core Processors 16GB RAM PC2-5300 DDR2 2 * 146GB SFF SAS HDD's (RAID 1) Windows Server 2008 Standard x64 (64-bit OS) Dedicated 100MB (1-to-1 connection) [ARMA Server Version] a2oa server v1.57.76903 Startup: start /W arma2oaserver.exe -config=server.cfg -port=2302 (no exThreads=x used) [Result - Test 1] During the start (before mission selection): FPS 50 Once Mission starts it went from FPS 15, to the lowest score of FPS 9 (fluctuated more between 10 & 11) Pre map-screen: FPS 50, 59MB RAM Map screen: FPS 50, 259MB RAM Ingame (Start): FPS 15, 294MB RAM 1st Troop Movement: FPS 12, 304MB RAM RAM started off at 59MB and went up to around 460MB (by the end of the test) ---- [ARMA Server Version] a2oa server v1.57.76903 Startup: start /W arma2oaserver.exe -config=server.cfg -port=2302 -exThreads=1 [Result - Test 2] During the start (before mission selection): FPS 50 Once Mission starts it went from FPS 38, to the lowest score of FPS 10 (was stable most of the time at FPS 11) Pre map-screen: FPS 50, 59MB RAM Ingame (Start): FPS 38, 284MB RAM 1st Troop Movement: FPS 12, 306MB RAM RAM started off at 59MB and went up to around 4xxMB (by the end of the test) Comment: Choppers that fly-by seem to stutter/warp/judder as they move 2nd set of troops that come from left to right (closer to the woods) seem to stutter/warp/judder slightly as they move (not all time) ---------- Post added at 22:51 ---------- Previous post was at 21:29 ---------- [server Browser Name] [16AA] Private Test #1 [server Specs] Intel Core2 Duo E6600 @ 2.4GHz 4GB PC2-5300 RAM Windows XP Professional (32-bit) Dedicated 100MB (1-to-1 connection) [ARMA Server Version] a2oa server v1.57.76903 Startup: start /W arma2oaserver.exe -config=server.cfg -port=2302 exThreads=1 [Result - Test 1] During the start (before mission selection): FPS 47 Once Mission starts it went from FPS 9, to the lowest score of FPS 7 (mostly at 9 and 8) Pre map-screen: FPS 47, 82MB RAM Map screen: FPS 46, 328MB RAM Ingame (Start): FPS 7, 367MB RAM RAM started off at 59MB and went up to around 667MB (by the end of the test) - There was a sudden surge in RAM towards the end of the test where it jumped from around 510MB upto 667MB *It would seem 'exThreads=1' is still something that needs to be manually set on the server start-up parametre for the best usage (for multicore systems). I had stopped doing that since one of the more recent patches, as I thought it was now automatically detected* Edited February 5, 2011 by LondonLad Share this post Link to post Share on other sites