Solved

Can you help me find my bottleneck(s) (if there are any)

Hello Guys

Could you please help me find potential bottlenecks in my rig? I'm not that great at matching components so please tell me what I could change/improve.

MB: Asus P8Z68-V PRO (https://www.asus.com/ch-de/Motherboards/P8Z68V_PRO/specifications/)
CPU: Intel i5-3570K 3.4 GHz (https://ark.intel.com/products/65520/Intel-Core-i5-3570K-Processor-6M-Cache-up-to-3_80-GHz)
GPU: 2 x AMD Radeon R9 290 4GB in bridgeless Crossfire (https://www.asus.com/Graphics-Cards/R9290DC2OC4GD5/specifications/)
RAM: 32 GB RAM DDR3-2133 (Overkill, I know... http://www.corsair.com/en-gb/vengeance-pro-series-16gb-2-x-8gb-ddr3-dram-2133mhz-c11-memory-kit-cmy16gx3m2b2133c11)
Storage: Samsung SSD 850 EVO 1TB (for OS, programs and games) + regular HDD's (for files, movies, pictures)
PSU: Corsair HX850

In case it helps, here are some Fire Strike benchmarks. I tried to overclock a little bit as you can see in the result on the right side. http://www.3dmark.com/compare/fs/11794397/fs/11804230#

Overclocking the CPU more seems to result in Problems, but that might be because I'm not that skilled yet when it comes to OC'ing.

Another Thing i noticed is that games seem to perform a lot better (or at least cause less problems) in windowed fullscreen than in actual Fullscreen - which I think is weird because as far as i know Crossfire only works properly when the games are run in actual fullscreen...
Reply to Kollosso
11 answers Last reply Best Answer
More about find bottleneck
  1. There are no bottlenecks, it may depend on the game that there may be a possible bottleneck which you can check by just using MSI Afterburner or a similar software to monitor stats in gaming. If maybe your RAM/CPU/GPU have big differences in usages then the one(s) with the higher usage will be the bottleneck.
    Reply to Ali Tauseef
  2. Best answer
    Depends what you're doing with your computer.

    Overall you've got a good system! But assuming you're gaming, your bottleneck would be the crossfire cards. I'd try to avoid crossfire entirely if I were you and buy the best single-card that you can afford. Too many issues with lots of games not taking full advantage of both cards.

    If you're looking for a significant upgrade and you're at 1080p gaming, the new 580 should suit you very well. Great $/performance. Anything more (1440p, 144Hz+) and you'll need to go to a GTX 1070 / 1080.

    Also, what Harddrive do you have? If you don't have a SSD or M.2 then I'd first and foremost recommend to upgrade that over anything! A world of difference.

    Finally, anything over 16GB of RAM is more than enough and likely won't bottleneck. PSU's never bottleneck unless it hasn't got a high enough wattage for your components.
    Reply to Subressor
  3. There will always be some sort of bottleneck, it's just the way pc's run games. There should only be a concern if your PC isn't performing up to your standards, whether that's FPS wise, or graphical quality wise.

    Here are some steps to identify the bottleneck out of the CPU and GPU/s:
    1. Download & install MSI Afterburner
    2. Change settings in MSI afterburner to display the FPS, CPU load/percentage, and GPU load as the On-screen display
    3. Open your desired game and start to play.
    4. Turn down in game resolution as low as it can while maintaining the same aspect ratio. By doing this, you are directing the boundary towards the CPU.
    5. the FPS you are recieving now is how well your CPU can perform. If this FPS is below your performance goal, try overclocking more, or turn down (if available) settings like view distance and FOV.
    6. If FPS is fine here, turn the resolution back up to whatever's your monitor's native resolution is. Turn the graphics quality your preferred settings (in terms of looks). If you are receiving FPS underneath your performance goal, the GPU is the bottleneck. You can turn down graphics quality or upgrade GPU.
    7. If the FPS is at or above your performance goal, there is no bottleneck to be concerned of.

    MSI afterburner will give you an indication of how CPU or GPU dependant the game is, and how it performs on multiple gpus and CPU cores.
    Reply to Rexper
  4. Thank you all for your help.
    I installed MSI Afterburner and turned on the overlay as Rexper suggested.

    I tested one game so far (Hitman on Ultra @1080p) and turns out the CPU Usage was on 100%, which i kind of doubt but I was going for an overclock anyways. So i went into the Bios and accidentaly started OC Tuner, which is supposed to automatically overclock your system. Since it gives you somewhat of a basis to tweak later I decided to let it finish its job - which it unfortunately didn't. All i got was a black screen for more than an hour while the PC itself was running. Has anyone ever experienced that?

    Also, does anyone have experience with overclocking the 3570K? What are your settings?

    I am for sure going to try Afterburner on other games to see if the CPU actually is the Problem (which again, I doubt). In the meantime I ordered a RX 580 to see if my gaming experience actually improves this way.
    Reply to Kollosso
  5. Re. CPU on 100%, I find that quite improbably. Try looking at task manager to see what programes are doing what with your CPU.

    Re. Black screen, sounds like a bad overclock. Go back to the BIOS, start again from scratch.

    Maybe try looking at guides and what others have done to see what extra performance you can squeeze
    http://www.tomshardware.co.uk/forum/278263-29-safely-efficiently-overclock-3570k

    Nearly all games rely mainly on the graphics for FPS wise, so I think the new card should give you a good increase. Some games however (like SC2) have a lot of physics and the CPU can play a big part. I'm not sure where Hitman falls.

    Your CPU however is still very much relevant, a good 70% of the stalwart 6700K. More than enough for todays titles
    https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-3570K+%40+3.40GHz
    Reply to Subressor
  6. A little update on things:

    Today I manually OC'ed my CPU which resulted in a stable 4.4GHz. So tomorrow i will get to continue my tests with Afterburner to determine if the cards are bottlenecking me. After that my rx 580 should arrive which i then again will test.

    I hope this will then finally give me the performance i expect.
    Reply to Kollosso
  7. Okay, so I am just confused at this moment.

    I ran a Firestrike benchmark with my newly Overclocked CPU and the score was just horrible.
    So I thought "okay then, so I must suck at OC'ing stuff" and reverted everything to what i had before.

    I tested again with Firestrike and i get significantly less FPS than before.
    See the results here. As far as I'm concerned I replicated my best test conditions as good as i could.
    Best result ever on the left, result now on the right:
    http://www.3dmark.com/compare/fs/11804230/fs/12456629

    Crossfire is actually working, I checked while playing Hitman. Both cards showed activity.
    What confuses me even more, is that when running Hitman in fullscreen, where crossfire actually works, I only had like 2 or 3 FPS more than in windowed mode with only 1 card active.

    WTF am I missing here? :??:
    Reply to Kollosso
  8. Just looking at the physics comparison (which is almost purely CPU dependant); it looks like your CPU's impact hasn't changed much. So the only conclusion I would draw is that there's something going on with your GPU's that's causing the difference.

    Maybe 1 isn't activating or playing properly? I know both benchmarks say both cards are "on", but maybe something is still going wrong. #CrossFireProblems. That's the only other source I can think of.
    Reply to Subressor
  9. Subressor said:
    Just looking at the physics comparison (which is almost purely CPU dependant); it looks like your CPU's impact hasn't changed much. So the only conclusion I would draw is that there's something going on with your GPU's that's causing the difference.

    Maybe 1 isn't activating or playing properly? I know both benchmarks say both cards are "on", but maybe something is still going wrong. #CrossFireProblems. That's the only other source I can think of.


    Since my new RX 580 arrived today and I will have to open my PC anyways I'll try each GPU on its own then to see if they both work properly (also for reselling purposes). I think I'll shy away from crossfire setups in the future anyway. I've had nothing but trouble with it... :(

    Is there a way i can tell for sure that crossfire is working properly? While playing games the cards took turns (high usage on card 1 / no usage on card 2 - quickly followed by the opposite) as they are supposed to do, aren't they?
    Reply to Kollosso
  10. Kollosso said:


    Is there a way i can tell for sure that crossfire is working properly? While playing games the cards took turns (high usage on card 1 / no usage on card 2 - quickly followed by the opposite) as they are supposed to do, aren't they?


    I don't use a crossfire/SLI setup, but that sounds like exactly what they're NOT suppose to do haha. That sounds to me like these some configuration issues (primary card selection etc.)

    Both cards should have a stable working load. I heard from a friend that normally his main card is at 100% and the 2nd card stays at around 70%, but varies from game to game.

    My personal opinion is to try and avoid at all costs; but its your system after all :-)
    Reply to Subressor
  11. I installed the RX 580 and tested it with Fire Strike again.
    It gets surpassed in the benchmark quite a bit by my crossfire setup in its best days, but in game it's a whole other story.

    It just outperforms my old setup by miles. There is no stuttering, just pure fluid gameplay.

    Thank you so much fot this tip! I think this is the best card you can get at this price at the moment!

    Thanks to everyone who helped, especially Subpressor :D
    Reply to Kollosso
Ask a new question Answer

Read More

RAM Bottleneck Components