+ Reply to Thread
Results 1 to 5 of 5

Thread: Way to Benchmark Settings

  1. #1
    Ascendant
    Join Date
    Jan 2012
    Posts
    2,570

    Default Way to Benchmark Settings

    This is something that a large number of MMOs lack, is a way to gate a good baseline of performance for certain settings and between machines.

    All it needs to be is about 3 minutes of a camera flyby through Santum, Meridian, Ember Isle, and a few good battle scenes of something like a pre-recorded scarn fight with piles of players around, a couple of new areas from the new X-Pac stuff. Show the FPS numbers broken down by flyby scene and overall.

    Currently I just walk a pre-determined route around on the PTS, so as to get the scenes as consistent as possible, but that essentially removes any other players. Since areas where there are lots of players are generally the cause of performance problems it's incredibly difficult to get any meaningful performance numbers while adjusting settings.

  2. #2
    Ascendant
    Join Date
    Jan 2012
    Posts
    2,570

    Default

    No one cares that much I take it?

    When I said Scarn up there i meant Scerrion of course . . . but I still think it would be a pretty valuable tool for tweaking settings so that you can actually play a zone event once in a while instead of standing there waiting for your screen to draw a frame while you make a cup of tea.
    Last edited by Sariina; 09-10-2012 at 07:09 AM.

  3. #3
    Ascendant
    Join Date
    Feb 2011
    Posts
    1,950

    Default

    I'm trying to understand your suggestion. You want this benchmark to be present, like a youtube video or something? I don't really understand what you're asking for, or what purpose it would serve.
    Regardless of what happens, people will always be unhappy. Be thankful for what you have and always be constructive.
    Quote Originally Posted by A Poster View Post
    good or the best
    This quote offers insight into why there are so many perceived balance issues. To some, it doesn't matter how good something is; if it isn't the best, it isn't worth using.

  4. #4
    Plane Touched Ahielia's Avatar
    Join Date
    Aug 2011
    Posts
    246

    Default

    I support this idea.

    Quote Originally Posted by Caydin View Post
    I'm trying to understand your suggestion. You want this benchmark to be present, like a youtube video or something?
    Have you never done the hardware tests of other games?
    Counter Strike had one, the camera went to certain sections, a scene played out and all the while the program recorded the max, min and average fps of the scene.
    Basically the same thing as 3DMark, Heaven and such, just within the confines of the game, using the game engine.

    or what purpose it would serve.
    It would give you a basic idea of how well the computer would perform at the given settings in a given area.

  5. #5
    Ascendant
    Join Date
    Jan 2012
    Posts
    2,570

    Default

    Quote Originally Posted by Caydin View Post
    I'm trying to understand your suggestion. You want this benchmark to be present, like a youtube video or something? I don't really understand what you're asking for, or what purpose it would serve.
    No, not like a youtube, like pre-recorded scenes in the game (but still rendered by the engine) so that you could adjust settings in your video setup, run through it, and see how your performance changes.

    The problem is you can't do this on a live server because the data changes every time. So there is no way to tell if the performance gain by lowing setting X is because setting X slows your machine down, or 20 people ported out of the area you were just testing.

    You can try this on the PTS, but the main performance problems are generally where people are (IE, boss events, home cities). What I do right now is try a setting or tweak something on my machine, log in to the PTS, and I run from the gloamwood pines portal, up through IPP, over to stillmoor and right in to the zareph return portal where I turn fraps off, port back to gloamwood pines and do the whole thing again. Even then though, if mobs move, or a rift opens, or the lighting changes . . . the comparison is not really valid (It's a fine approximation, and useful for what it is, but I can't compare these results to someone else for instance).

    It would also eliminate the problems of people saying things like "I get 120FPS in sanctum all the time on my 6x86 with a 12MB SLI'd voodoo2's . . . l2 tweak" since you would have a standard test that everyone can run to get relative performance between different machines.

    For example a benchmark test could run through:
    20 second camera fly by of NPC only Sanctum
    20 second camera fly by of NPC only Meridian
    Load the assets for each zone and test load times
    30 seconds 20 person recorded boss fight like Matron
    30 seconds of a 100 person scerrion fight in EI
    Display the load times and FPS results of each test

+ Reply to Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts