The Official Benchmarks Thread

245

Comments

  • GBT13GBT13 Join Date: 2012-08-05 Member: 154798Members
    CPU: Intel I5 2500K @ 3.8Ghz
    GPU: Nvidia Geforce GTX570
    OS: Windows 7 Ultimate 64 Bit
    RAM: 8 Gb @ 1600 Mhz
    Minimum FPS: 27
    Maximum FPS: 58
    Average FPS: 48
    Build: 216
    Resolution: 1920 X 1080
    Ping: ~45 ms

    <a href="http://imgur.com/PgYQB" target="_blank"><img src="http://i.imgur.com/PgYQB.png" border="0" class="linked-image" /></a>
  • NeoRussiaNeoRussia Join Date: 2012-08-04 Member: 154743Members
    edited August 2012
    CPU: Q6600 G0 @3.0ghz
    GPU: Nvidia Geforce GTX460 1GB
    OS: Windows 7 Ultimate 64 Bit
    RAM: 4 Gb @ 1600 Mhz
    Minimum FPS: 8
    Maximum FPS: 72
    Average FPS start of game: 30
    Average FPS end of game: 10
    Build: 216
    Resolution: 1920 X 1080
    Ping: 30 ms

    CPU use - 20-40% first core, 10-20% other cores
    GPU use - 20% and less
  • TyphonTyphon Join Date: 2002-11-01 Member: 1899Members
    CPU: Intel Core i3-2120 @ 3.3Ghz
    Ram: 8 GB DDR3 1333
    GFX Card: AMD 6870 1Gb
    OS: Windows 7 Ultimate 64 bit

    Build: 216
    Resolution: 1680 x 1050
    Vsync: Off
    Ping: ~75 ms

    (ignoring readyroom and an occasional glitch)

    Game 1
    Minimum FPS: 24
    Maximum FPS: 65
    Average FPS: 46

    Game 2
    Minimum FPS: 22
    Maximum FPS: 61
    Average FPS: 44
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    edited August 2012
    Hey, Dghelneshi how do you chart the frame times? (I'm a total noob at excel!) I guess you need to make it only chart the difference between the last cell. I assume the points where the frame time is really low is when you died and were in spectator. Interesting how there is a spike when you die and then when you respawn.

    EDIT: Figured it out ^_^
  • DghelneshiDghelneshi Aims to surpass Fana in post edits. Join Date: 2011-11-01 Member: 130634Members, Squad Five Blue, Reinforced - Shadow
    edited August 2012
    Yes, every time you die or respawn you get a huge hitch since the game needs to reset and load a lot of stuff at the same time. Frame times are more interesting than FPS in many cases since they capture spikes like that (which have occured outside of deaths, too), though I should have made a bar chart instead of a line chart (and 1px width).
    Too bad Excel can't handle more than 32000 data points. I had to screenshot, change the beginning and end points of the data, screenshot again and then fit them together.
  • ScardyBobScardyBob ScardyBob Join Date: 2009-11-25 Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
    <!--quoteo(post=1964734:date=Aug 18 2012, 10:40 AM:name=Dghelneshi)--><div class='quotetop'>QUOTE (Dghelneshi @ Aug 18 2012, 10:40 AM) <a href="index.php?act=findpost&pid=1964734"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Yes, every time you die or respawn you get a huge hitch since the game needs to reset and load a lot of stuff at the same time. Frame times are more interesting than FPS in many cases since they capture spikes like that (which have occured outside of deaths, too), though I should have made a bar chart instead of a line chart (and 1px width).
    Too bad Excel can't handle more than 32000 data points. I had to screenshot, change the beginning and end points of the data, screenshot again and then fit them together.<!--QuoteEnd--></div><!--QuoteEEnd-->I agree, but for the purposes of this thread, avg, min, and max are enough. Without other data correlated to the frametime graphs, the info is not particularly useful.
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    edited August 2012
    Here are mine:

    Specs:

    CPU: Core i5 750 2.66GHz
    Graphics: ATI 4870
    RAM: 4GB
    OS: Vista SP2 64-bit

    All graphic settings low 1680x1050

    All-in server 18 players, 150ping


    Pretty much unplayable for me :(


    Min FPS: 0 Max: 49 Avg: 27.775

    99th Percentile frame time: 65ms


    <img src="http://i.imgur.com/WUrbs.jpg" border="0" class="linked-image" />

    <img src="http://i.imgur.com/d6UZJ.jpg" border="0" class="linked-image" />
  • DghelneshiDghelneshi Aims to surpass Fana in post edits. Join Date: 2011-11-01 Member: 130634Members, Squad Five Blue, Reinforced - Shadow
    edited August 2012
    <!--quoteo(post=1964782:date=Aug 18 2012, 11:36 PM:name=ScardyBob)--><div class='quotetop'>QUOTE (ScardyBob @ Aug 18 2012, 11:36 PM) <a href="index.php?act=findpost&pid=1964782"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I agree, but for the purposes of this thread, avg, min, and max are enough. Without other data correlated to the frametime graphs, the info is not particularly useful.<!--QuoteEnd--></div><!--QuoteEEnd-->
    That's odd, because I actually find min/max/avg completely useless. It doesn't actually tell you how well the game runs. Frametime and FPS graphs can tell you how much of that "average FPS" was caused by actual gameplay and how much was just readyroom or spectating. They also tell you how much the games slow down in the later stages compared to the beginning of a match. It's also odd that Alien spectators seem to have less fps than Marine spectators.
  • IronHorseIronHorse Developer, QA Manager, Technical Support & contributor Join Date: 2010-05-08 Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
    edited August 2012
    Wilson, OC that thing, use windows 7 64bit and maybe grab some extra/faster ram!
    That default clock speed is what is killing ya. (just look at tom hanks' OC results)

    Also, you guys can create an actual benchmark currently, it just wouldnt be backwards compatible between the patches.
    Just record a demo and run it in realtime (theres a hotkey for this, but i forget) and observe the results, with or without fraps.
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    <!--quoteo(post=1964835:date=Aug 19 2012, 02:05 AM:name=ironhorse)--><div class='quotetop'>QUOTE (ironhorse @ Aug 19 2012, 02:05 AM) <a href="index.php?act=findpost&pid=1964835"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Wilson, OC that thing, use windows 7 64bit and maybe grab some extra/faster ram!
    That default clock speed is what is killing ya. (just look at tom hanks' OC results)<!--QuoteEnd--></div><!--QuoteEEnd-->

    I had it OCed for a while but it was a bit unstable and I was too lazy to sort it out so I just put it back to the defaults. I just hope the actual 1.0 release is playable for me with these settings. I don't feel like I should need to overclock just to play this game. It looks worse than the source engine and the performance is like 75% worse.
  • IronHorseIronHorse Developer, QA Manager, Technical Support & contributor Join Date: 2010-05-08 Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
    Highly disagree in regards to the source engine comparison (esp with texture quality and lighting) but yea the performance is not as good as an engine thats literally the baseline of scale-ability after 10 years of development from hundreds of people (valve, no less) and refined over another 8 years.

    Idk if things will be good enough to run the game with those specs at 1.0, we all hope so, but in the meantime you could always OC again with some cheap 60 buck water cooling? I'm running an early i7 @ 4 ghz and it's default clock was 2.66. Smooth gameplay (~60fps) resulted from that change. But i understand that this isn't feasible for anyone but the enthusiast. :-/ Hope things improve for you and the game's success!
    /derailing
  • ImbalanxdImbalanxd Join Date: 2011-06-15 Member: 104581Members
    edited August 2012
    I'm no professional, well actually technically I am at the least qualified, regardless, the kind of performance issues people are getting at the moment cannot be "optimised". Optimisation is a very powerful word around here, and has all these positive connotations surrounding it, but its really not as big a deal as most people seem to think. Simple game optimisations are going to give you about a 10% performance increase, maybe even as much as 20% if they are really amazing. That means my average FPS can potentially rise from 27 to a maximum around 32, hey maybe I can be really optimistic and say 35 FPS.

    What I'm looking for is an increase of around 100%, to get my average FPS close to 60, but that is only acceptable when the deviation from the average is low. What I'm REALLY looking for is an increase closer to 400%, that way my actual frame rate would be above 50 in combat and out. That kind of improvement does not come from "optimisations". That kind of improvement comes from system overhauls, and discovering glaring horrendous bugs like "what's this? oh my we are rendering the world 10 times each frame by mistake!".

    The kind of optimisations that people think are coming are not, simply because they are not possible. If you're getting 20 FPS now, you can look forward to 25 at release.
  • playerplayer Join Date: 2010-09-12 Member: 73982Members
    I don't know, I find it hard to believe the developers wouldn't have some kind of certainty about the performance that can be yielded from Lua when they first started this project, but on the other hand it doesn't look very good at this moment in time. I'm sitting tight until v1.0.
  • TyphonTyphon Join Date: 2002-11-01 Member: 1899Members
    <!--quoteo(post=1964880:date=Aug 19 2012, 05:21 AM:name=Imbalanxd)--><div class='quotetop'>QUOTE (Imbalanxd @ Aug 19 2012, 05:21 AM) <a href="index.php?act=findpost&pid=1964880"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I'm no professional, well actually technically I am at the least qualified, regardless, the kind of performance issues people are getting at the moment cannot be "optimised". Optimisation is a very powerful word around here, and has all these positive connotations surrounding it, but its really not as big a deal as most people seem to think. Simple game optimisations are going to give you about a 10% performance increase, maybe even as much as 20% if they are really amazing. That means my average FPS can potentially rise from 27 to a maximum around 32, hey maybe I can be really optimistic and say 35 FPS.

    What I'm looking for is an increase of around 100%, to get my average FPS close to 60, but that is only acceptable when the deviation from the average is low. What I'm REALLY looking for is an increase closer to 400%, that way my actual frame rate would be above 50 in combat and out. That kind of improvement does not come from "optimisations". That kind of improvement comes from system overhauls, and discovering glaring horrendous bugs like "what's this? oh my we are rendering the world 10 times each frame by mistake!".

    The kind of optimisations that people think are coming are not, simply because they are not possible. If you're getting 20 FPS now, you can look forward to 25 at release.<!--QuoteEnd--></div><!--QuoteEEnd-->
    I fully expect that during their optimization sweep there are going to be a handful of forehead-slapping design flaws found which when fixed will dramatically improve performance. Things like not caching the result of an expensive but frequently requested query, or resorting a list after every insert instead of inserting all values then sorting, etc.
  • DghelneshiDghelneshi Aims to surpass Fana in post edits. Join Date: 2011-11-01 Member: 130634Members, Squad Five Blue, Reinforced - Shadow
    edited August 2012
    <!--quoteo(post=1964926:date=Aug 19 2012, 04:52 PM:name=Typhon)--><div class='quotetop'>QUOTE (Typhon @ Aug 19 2012, 04:52 PM) <a href="index.php?act=findpost&pid=1964926"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I fully expect that during their optimization sweep there are going to be a handful of forehead-slapping design flaws found which when fixed will dramatically improve performance. Things like not caching the result of an expensive but frequently requested query, or resorting a list after every insert instead of inserting all values then sorting, etc.<!--QuoteEnd--></div><!--QuoteEEnd-->
    They apparently had something like this in b217. *spoiler*
    By the way, I once made a very simple function to calculate pi <b>twice</b> as fast as before just by simple techniques such as rearranging the variables in the calculation (e.g. "(i/n)*(i/n)" instead of "(i*i)/(n*n)"). Simple optimization can be very efficient.
  • ScardyBobScardyBob ScardyBob Join Date: 2009-11-25 Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
    <!--quoteo(post=1964880:date=Aug 19 2012, 02:21 AM:name=Imbalanxd)--><div class='quotetop'>QUOTE (Imbalanxd @ Aug 19 2012, 02:21 AM) <a href="index.php?act=findpost&pid=1964880"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I'm no professional, well actually technically I am at the least qualified, regardless, the kind of performance issues people are getting at the moment cannot be "optimised". Optimisation is a very powerful word around here, and has all these positive connotations surrounding it, but its really not as big a deal as most people seem to think. Simple game optimisations are going to give you about a 10% performance increase, maybe even as much as 20% if they are really amazing. That means my average FPS can potentially rise from 27 to a maximum around 32, hey maybe I can be really optimistic and say 35 FPS.

    What I'm looking for is an increase of around 100%, to get my average FPS close to 60, but that is only acceptable when the deviation from the average is low. What I'm REALLY looking for is an increase closer to 400%, that way my actual frame rate would be above 50 in combat and out. That kind of improvement does not come from "optimisations". That kind of improvement comes from system overhauls, and discovering glaring horrendous bugs like "what's this? oh my we are rendering the world 10 times each frame by mistake!".

    The kind of optimisations that people think are coming are not, simply because they are not possible. If you're getting 20 FPS now, you can look forward to 25 at release.<!--QuoteEnd--></div><!--QuoteEEnd-->
    Well, considering that playable server playercount went from 4 during the early alpha/engine test to roughly 16-18 players, I'd say performance improvements on the level your talking about (100-400%) are both possible and have been done. The issue for me is that these can take time to do and UWE is set to release v1.0 this summer.
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    <!--quoteo(post=1964840:date=Aug 19 2012, 02:30 AM:name=ironhorse)--><div class='quotetop'>QUOTE (ironhorse @ Aug 19 2012, 02:30 AM) <a href="index.php?act=findpost&pid=1964840"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Idk if things will be good enough to run the game with those specs at 1.0, we all hope so...
    /derailing<!--QuoteEnd--></div><!--QuoteEEnd-->

    My pc is almost the same as the recommended requirements. i5 750 has turbo boost that brings it up to 3.2 ghz. I sure do hope that the recommended requirements can run the game significantly better than they do now come 1.0
  • leeleatherwoodleeleatherwood Join Date: 2012-08-14 Member: 155623Members
    edited August 2012
    <img src="http://donaldleatherwood.com/temp/computer/ns2benchmark.png" border="0" class="linked-image" />

    Min Max Avg
    37 104 67.586

    i5 2500k @ 5.0ghz
    ASUS 7970 DirectCUII-TOP (no OC, forgot to load up Trixx after reboot)
    ASUS Maximus V Gene
    16GB DDR3 @ 2133mhz
    OCZ Vertex 3 SSD
    Windows 7 Pro 64 bit SP1

    Build: 216
    Resolution: 1920 x 1080
    Vsync: Off
    Ping: ~60ms
    Everything maxed in video options
    18 Player server, it was full.
  • MocMoc Join Date: 2012-08-07 Member: 154887Members
    edited August 2012
    <img src="http://i.imgur.com/Io5g0.jpg" border="0" class="linked-image" />

    Min: 0
    Max: 102
    Avg: 63.323

    OS: Win 7 Professional 64-bit
    CPU: i7-2600 @ 3.4ghz
    MB: ASUS P8P67-M-PRO
    RAM: Corsair DDR3 8GB 1600MHz
    GPU: EVGA GTX 580 3GB

    Build: 216
    Resolution: 1920 x 1080
    Vsync: Off (I assume)
    Ping: ~20ms
    Everything maxed in video options
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    I overclocked my CPU to 3.57GHz and ran a new benchmark on a server with better ping and less players (and less buildings etc.). Here are the results:

    Specs:

    CPU: Core i5 750 3.57GHz
    Graphics: ATI 4870
    RAM: 4GB
    OS: Vista SP2 64-bit

    All graphic settings low 1680x1050

    HBZ server 14 players, 60 ping

    FPS:

    Min: 0
    Max: 90
    Avg: 41.464


    99th Percentile frame time: 48ms

    FPS:
    <img src="http://i.imgur.com/hPq4M.jpg" border="0" class="linked-image" />

    Frametimes (ms):
    <img src="http://i.imgur.com/SKRRH.jpg" border="0" class="linked-image" />
  • Soul_RiderSoul_Rider Mod Bean Join Date: 2004-06-19 Member: 29388Members, Constellation, Squad Five Blue
    Wilson, I don't understand why your performance is so poor?

    I have a i3-530 (dual core with hyperthreading), 4gb 1300mhz RAM, a GTS 250 512 gfx card, medium settings and 1920x1080, and I get better scores than your overclocked score.

    I overclock to 3.4ghz, but keep the ram at about 1360mhz, I can't see how your performance, especially with that lower resolution, more powerful cpu and gfx card, can be so low.

    The only difference is I am still on windowsXP, could vista be the source of your problems?
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    edited August 2012
    Soul_rider, I don't think it's vista. Everyone ###### on it, but I've been using it for years and I've never had any problems in any other games. Maybe I'm just unlucky with servers or something...I noticed that in the console it said "Interpolation changed to XYZ" a few times during the game, perhaps the admin changing settings was causing hitching or something. Can you post up some benchmarks of your performance on an actual server for me to compare?

    EDIT: I just had a thought. Perhaps it's my wireless adapter causing the game to lag and lowering FPS. I recently changed to a new router and wireless adapter and now I get lag spikes in multiplayer games. It's hard to notice the lag spikes in ns2, but perhaps it's causing the low fps. I'll try and setup my old router again and retest it to see if I notice any improvements.
  • Soul_RiderSoul_Rider Mod Bean Join Date: 2004-06-19 Member: 29388Members, Constellation, Squad Five Blue
    Dude, I know people will probably tell you otherwise, but many years ago I made a very important discovery that remains true to this day, do not game on wireless!!

    Packet drops are far more common on wireless, and in fact, my flatmate just had to plug in her laptop, because our 6 week old router hass decided to only give 100kb connection speeds at random points in time for no apparent reason.

    Wireless is just too unstable for gaming. Plug yourself in and I guarantee you will see a performance increase.
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    That's actually untrue. I've gamed on wireless for years with no problems. The problem isn't wireless, it's the way windows handles wireless adapters. It constantly searches for new wireless networks every minute and this can cause huge ping spikes and packet loss etc. AFAIK there is no way to turn it off. I have tried multiple solutions but none of them have worked fully. The only proper solution I found was to get a wireless ethernet adapter. It plugs into your ethernet port and windows just treats it like a wired connection. After I got that I had perfect lag free gaming.

    Recently my internet got upgraded though and I got a new router and wireless n usb adapter. I now experience lag in games again, but I've not really been actively playing much so I've not been bothered to try and fix it or buy new hardware. I will try going to back to my old hardware and retest to see if there is a significant difference.
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    Here is a final benchmark with my old wireless router. This is probably a best case scenario for me. When I joined the server it only had 10 players (and very little buildings) and then filled up to 16.

    Specs:

    CPU: Core i5 750 3.57GHz
    Graphics: ATI 4870
    RAM: 4GB
    OS: Vista SP2 64-bit

    All graphic settings low 1680x1050

    HBZ server 16 players, 60 ping

    FPS:

    Min: 14
    Max: 73
    Avg: 39.461

    99th Percentile FPS: 20

    99th Percentile frame time: 49.5ms


    FPS:

    <img src="http://i.imgur.com/qL5r2.jpg" border="0" class="linked-image" />

    Frametimes (ms):

    <img src="http://i.imgur.com/6g8MS.jpg" border="0" class="linked-image" />


    Soul_Rider, I would really like to see your benchmarks cause I doubt you get much better than this on a full server. The thing that really stands out to me is the frametimes, just for comparison here are my frametimes from dod source (always over 100fps):

    <img src="http://i.imgur.com/GZtU1.jpg" border="0" class="linked-image" />
  • ScardyBobScardyBob ScardyBob Join Date: 2009-11-25 Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
    edited August 2012
    I have and can game over wireless. However, I've noticed NS2 is particularly prone to lag when doing so.
  • Soul_RiderSoul_Rider Mod Bean Join Date: 2004-06-19 Member: 29388Members, Constellation, Squad Five Blue
    I know many people who swear by gaming over wireless, but I used to manage a LAN gaming centre in MK that tried a large scale wireless gaming implementation that back fired massively.

    Admittedly wireless networks have improved over the years, but we found there were just too many minor problems and issues to be able to provide professional grade gaming on a wireless network.
  • Soul_RiderSoul_Rider Mod Bean Join Date: 2004-06-19 Member: 29388Members, Constellation, Squad Five Blue
    edited August 2012
    Don't know how to make graphs..

    Specs boosted to match yours, give or take...
    i3 530@3.6ghz
    DDR3@1440mhz - 4GB
    nVidia GTS250 - 512mb (stock)
    Medium details
    ns2_mineshaft (yes that killer map)
    Recorded for 27m39secs as marine, from readyroom to readyroom, we got hammered!
    Started with 15players, then went up to full 18, OAG #2

    Frames, Time (ms), Min, Max, Avg
    57220, 1672187, 0, 92, 34.219

    Recorded all the data, but i cannot for the life of me get a table created without bloody open office crashing and burning, too many lines of data! If anyone can tell me how to get a chart created without crashing my system, please let me know :)

    Interestingly on mineshaft I get lower fps when watching the ip than i do while being alive in marine spawn :p
  • koewikoewi Join Date: 2007-08-25 Member: 61984Members, Reinforced - Shadow
    <!--quoteo(post=1966303:date=Aug 24 2012, 03:30 AM:name=Wilson)--><div class='quotetop'>QUOTE (Wilson @ Aug 24 2012, 03:30 AM) <a href="index.php?act=findpost&pid=1966303"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->It constantly searches for new wireless networks every minute and this can cause huge ping spikes and packet loss etc. AFAIK there is no way to turn it off. I have tried multiple solutions but none of them have worked fully.<!--QuoteEnd--></div><!--QuoteEEnd-->

    Did you try <a href="http://www.martin-majowski.de/wlanoptimizer/" target="_blank">http://www.martin-majowski.de/wlanoptimizer/</a> to change this?
    I experience some lag when i play CS 1.6 on my wireless connection, but this tool fixes it completely.
  • WilsonWilson Join Date: 2010-07-26 Member: 72867Members
    Soul, upload your spreadsheets and I will make a quick graph for you. Your average FPS is actually lower than mine (as you would expect), did you just think that it would be higher than that?

    koewi, yes, I'm using that tool right now in fact but it doesn't fix the issue completely for me. I still get lag spikes although not as bad as when windows manages it itself. I think many people just have ###### routers (like the one my ISP gave me) and usb wireless adapters that cause many problems when gaming. When I went to my old router and wireless adapter everything works completely fine with no ping spikes. I just use the new one to take full advantage of my internet speed. If I start playing online games more frequently again I guess I will invest in a new wireless n router and ethernet adapter. They are so expensive though for the decent ones, it doesn't seem worth it to me right now.
Sign In or Register to comment.