CPU bottleneck or GPU bottleneck? How to actually tell

Page 1 of 2 · 25 posts

1. MCMaya C.

   Maya C.Original post

   @maya_c · 23 Mar 2026, 17:54

   #1

   I have read the FAQ and it does not cover this case. What confuses me is that it behaves differently depending on the order I do things in.

   Let me know if more detail would help.

2. MHMarco H.

   Marco H.

   @marco_h · 24 Mar 2026, 12:14

   #2

   For what it is worth, worth testing on a second account to rule out something profile-specific.

3. PKPilot KX

   Pilot KXMod

   @pilot_kx · 24 Mar 2026, 21:24

   #3

   From experience, you get most of the benefit from the first adjustment; everything after that is marginal.

4. DPDamian Piszczek

   Damian PiszczekAdmin

   @damianpiszczek · 25 Mar 2026, 09:28

   #4

   The usual fix here is worth testing on a second account to rule out something profile-specific.

5. BLBlazenyx

   Blazenyx

   @blazenyx15 · 26 Mar 2026, 13:31

   #5

   This is the bit people miss — the wiki number is theoretical; real results are always a bit lower.

6. GHGhosthawk

   Ghosthawk

   @ghosthawk · 26 Mar 2026, 22:32

   #6

   Adding to the above: the averages hide the problem; look at the one percent lows instead.

   That fixed it for me, at least.

7. DRDriftspark

   Driftspark

   @driftspark44 · 27 Mar 2026, 13:14

   #7

   Slightly different take — keep a note of what you changed, otherwise you cannot undo it cleanly later.

   Report back once you have tried it.

8. SESam E.

   Sam E.

   @sam_e · 28 Mar 2026, 11:31

   #8

   I went through this last month. the wiki number is theoretical; real results are always a bit lower.

9. SOSolarfox

   Solarfox

   @solarfox77 · 29 Mar 2026, 09:10

   #9

   Counterpoint: check whether an overlay is running before blaming the game itself.

10. BLBlazelynx

    Blazelynx

    @blazelynx68 · 29 Mar 2026, 11:55

    #10

    This is the bit people miss — region and server load matter more than most guides acknowledge.

11. DPDamian Piszczek

    Damian PiszczekAdmin

    @damianpiszczek · 30 Mar 2026, 10:23

    #11

    Slightly different take — verify the files before anything else; it catches more problems than people expect.

12. BLBlazelynx

    Blazelynx

    @blazelynx32 · 30 Mar 2026, 14:47

    #12

    Speaking as someone who got it wrong initially, you get most of the benefit from the first adjustment; everything after that is marginal.

    Let us know which part worked.

13. GHGhosthawk

    Ghosthawk

    @ghosthawk · 30 Mar 2026, 16:54

    #13

    For what it is worth, it depends heavily on your specific setup, so take any single number with a grain of salt.

    Good luck, it is a common snag.

14. GHGhosthawk

    Ghosthawk

    @ghosthawk · 31 Mar 2026, 09:30

    #14

    This is the bit people miss — the defaults are conservative on purpose, so there is usually headroom to gain.

    Good luck, it is a common snag.

15. ECEchofen

    Echofen

    @echofen · 1 Apr 2026, 07:05

    #15

    From experience, temperature is the usual culprit when performance degrades over a session.

    Report back once you have tried it.

16. THTheo H.

    Theo H.

    @theo_h · 1 Apr 2026, 12:27

    #16

    Adding to the above: the trick is to change one thing at a time and actually note the result before touching anything else.

17. MFMaya F.

    Maya F.Mod

    @mayaf · 2 Apr 2026, 08:50

    #17

    Short version: the difference is real but small enough that you may not notice it while playing.

18. DRDriftlynx

    Driftlynx

    @driftlynx · 2 Apr 2026, 16:11

    #18

    Agreed on most of this, except keep a note of what you changed, otherwise you cannot undo it cleanly later.

19. PGPriya G.

    Priya G.

    @priya_g · 3 Apr 2026, 00:42

    #19

    In my testing, the defaults are conservative on purpose, so there is usually headroom to gain.

20. JZJonas Z.

    Jonas Z.

    @jonas_z16 · 4 Apr 2026, 02:11

    #20

    In my testing, this changed for me after the last patch, so older threads may steer you wrong.

    Let us know which part worked.