pre-rendered frames in nvidia drivers?

Discussion in 'Videocards - NVIDIA GeForce Drivers Section' started by alexander1986, Sep 12, 2016.

  1. alexander1986

    alexander1986 Master Guru

    Messages:
    201
    Likes Received:
    21
    GPU:
    RTX 2060
    Hey, so I've hear setting pre-rendered frames to 1 in nvidia drivers will reduce input lag and in some cases increase fps (in cases where the CPU is bottlenecking it might be decreased fps though, atleast from what I read)

    so, whats the truth here?

    I have the system you can see in my profile, should I put this on auto , or 1 ?

    or something else? and why? thanks in advance!
     
  2. Glottiz

    Glottiz Ancient Guru

    Messages:
    1,922
    Likes Received:
    1,164
    GPU:
    TUF 3080 OC
    turn on fps meter, set prerendered to 1 and see if you get more fps. then move your mouse around and see if you notice less input lag. if you don't notice any difference then what does it matter if anyone here tells you to use 1 or 2 or whatever.

    rule of thumb if you have to ask about this better just leave it at auto.
     
  3. alexander1986

    alexander1986 Master Guru

    Messages:
    201
    Likes Received:
    21
    GPU:
    RTX 2060
    yeah, that sounds good I will check later, but does anyone know the logic behind setting it to 1, I mean whats the difference?
     
  4. nav-jack

    nav-jack Master Guru

    Messages:
    253
    Likes Received:
    0
    GPU:
    980 ti 1520/2000
    you'd need to use presentmon to see if there are any real technological differences beyond our own flawed perception of things. i still set it to 1 in competitive games tho
     

  5. laimis911

    laimis911 Member Guru

    Messages:
    127
    Likes Received:
    4
    GPU:
    GTX1070
    if you want to feel the difference then you have to lock you frame rate to 30, set in NV control panel vsync to adaptive (half refresh) or set the game to 30fz via games menu ,when it is locked go to a game and move camera around with a controler analog stick, you will notice that there is much faster response compared to pre-rendered frames 2,3 or auto, but you will not notice the difference if you are playing at higher frame rates, pre-rendered frames 1 best works with 30fps or less where the input latency is highest .
     
  6. Ribix

    Ribix Guest

    Messages:
    4
    Likes Received:
    0
    GPU:
    Geforce GTX 560 Ti 1GB
    Maximum pre rendered frames limits the number of frames that can be in the context cpu queue. Lower numbers will have less input lag but can lower performance. You can use gpuview to view the effects of maximum pre rendered frames has on the context cpu queue.
     
  7. yobooh

    yobooh Guest

    Messages:
    260
    Likes Received:
    15
    GPU:
    Gigabyte 970 G1
    In some games 1 could give problems...
    I set it to 2 in global config.
     
  8. Xtreme512

    Xtreme512 Master Guru

    Messages:
    795
    Likes Received:
    44
    GPU:
    RTX 4080 Super
    I always put it on 1 globally for years. But if you have old CPU etc. or having troubles, set it to 2. I guess I will put it on 2 globally too.
     
  9. dr_rus

    dr_rus Ancient Guru

    Messages:
    3,882
    Likes Received:
    1,015
    GPU:
    RTX 4090
    1 gives me issues in Metro 2033 Redux for example, can't say that my CPU is old, can I? So yeah, leave it at 2 globally and only set 1 for those games which benefit from this.
     
  10. Dragondale13

    Dragondale13 Ancient Guru

    Messages:
    1,527
    Likes Received:
    244
    GPU:
    GTX 1070 AMP! • H75
    I fired up BSI the other night with pre-rendered frames on 1 globally and movement was a bit sluggish.Didn't change it to 2, just set it back to app controlled and it was fixed.
     

  11. khanmein

    khanmein Guest

    Messages:
    1,646
    Likes Received:
    72
    GPU:
    EVGA GTX 1070 SC
    in the game choose 1 or 2 & NVCP just leave it auto or play around the value.
     
  12. MrBonk

    MrBonk Guest

    Messages:
    3,385
    Likes Received:
    283
    GPU:
    Gigabyte 3080 Ti
    The guy who made TZFix for Tales of Zestiria did testing before he got 60hz working.
    In ToZ 2 provided the best mix of lowered input latency and frame stability and 1 had the lowest frame stability (Though in most cases you will be hard pressed to tell the difference) and lowest IL.

    I can feel the difference between 3 and 2 and even 1.

    In some games 3 to 1 can drop almost 30ms of input lag. And when using 1/2 refresh rate sync for perfect 30hz performance. 3 is almost unplayable. 1 makes a huge difference.

    Here's some more definitive testing using a high speed camera and USFIV
    http://www.displaylag.com/reduce-input-lag-in-pc-games-the-definitive-guide/

    It definitely makes a difference in input latency. If that's a priority. I'd definitely use 1.
     
  13. alexander1986

    alexander1986 Master Guru

    Messages:
    201
    Likes Received:
    21
    GPU:
    RTX 2060
    thanks for this answer, the game im playing in this case is a realllly old game, quake1 (quakeworld mod) so its not graphically intensive at all

    , can have over 2000 fps solid in it, but this game is the least forgiving when it comes to input delay and stuff, need flawless input for this game :D

    gonna play around a bit with 1/2/auto i think, thanks again!
     
  14. theahae

    theahae Active Member

    Messages:
    67
    Likes Received:
    12
    GPU:
    GTX 1060
    I think OP got it mixed up? If you have an older cpu you should use higher value of pre-rendered frames (eg 3) and not lower. Since higher values are most suited for weaker cpus. Can anyone confirm ?
     
  15. khanmein

    khanmein Guest

    Messages:
    1,646
    Likes Received:
    72
    GPU:
    EVGA GTX 1070 SC
    yeah so far i tested 3 is really unplayable on all NBA games & watch dogs

    USFIV & SFV i set auto so far everything ok with fast sync too.
     

  16. aufkrawall2

    aufkrawall2 Ancient Guru

    Messages:
    4,352
    Likes Received:
    1,815
    GPU:
    7800 XT Hellhound
    I set it to 1 globally, never had any problems with it (only exception Gothic 2 DX11 mod which is highly experimental).
    In most games there isn't a real performance loss either, often it's just ~1%.
     
  17. Generally, I leave the maximum prerendered frames setting on "Use 3D Application Setting", but there are certain games (for example AC games, some open-world games) that do not feel as smooth (more variable frametime fluctuation) using the default setting so, decreasing to 2 or increasing to 4 to 6 depends on the game. It really depends on the game so, experiment with the setting on game by game basis.
     
  18. TheRyuu

    TheRyuu Guest

    Messages:
    105
    Likes Received:
    0
    GPU:
    EVGA GTX 1080
    The prerendered frames setting is unlikely to affect input lag in games unless you're using vsync[1].

    Remember that if you're going to do testing like this you should make sure that it is objectively better than the default (hint: it's unlikely to be). You can measure this objectively with PresentMon[1] using frame time analysis. Don't fall into the trap of thinking something "feels better".

    [1] http://esreality.com/post/2640619/input-lag-tests-ql-csgo/
    [2] https://github.com/GameTechDev/PresentMon
     
  19. aufkrawall2

    aufkrawall2 Ancient Guru

    Messages:
    4,352
    Likes Received:
    1,815
    GPU:
    7800 XT Hellhound
    This is simply wrong, vsync just makes it even worse.
    This prerender queue does not affect DB/TB behavior.
     
  20. Mda400

    Mda400 Maha Guru

    Messages:
    1,089
    Likes Received:
    200
    GPU:
    4070Ti 3GHz/24GHz
    As the OP guessed correctly, only if your CPU is bottlenecking the GPU is when you change to other values than 1. So set it to 1 globally then higher values if a certain game is bottlenecked by your CPU.


    This is true. You can think of it as a hardware's backbuffer (CPU) instead of a software's backbuffer (Vsync).
     
    Last edited: Sep 15, 2016

Share This Page