Jump to content

beep beep today's test is complete

 

Stock config (PCIE 3.0 x16, CPU on stock, GPU on slight OC)

Spoiler

PCIE_3.0_x16.thumb.PNG.522b13800ab56e2d91bbf709e034bc23.PNG

 

PCIE 2.0 x16

Spoiler

PCIE_2.0_x16.thumb.PNG.349d61b278344305cb91bd7f2bf4d467.PNG

 

PCIE 1.1 x16

Spoiler

PCIE_1.1_x16.thumb.PNG.73309b40e507f76065766d6b6d305d45.PNG

 

PCIE 1.1 x4

Spoiler

PCIE_1.1_x4.thumb.PNG.b872733dd72d030d0c8fb81c6b69500a.PNG

 

PCIE 1.1 x1

Spoiler

PCIE_1.1_x1.thumb.PNG.77f75d8e93efc1d2fa9d99024387b8b4.PNG

 

 

all the frametime is smooth tbh, I expected low bandwidth to have spiky frame time but nope, it's very flat

 

though pcie 1.1 x4 had one small frametime bump, not a spike but weird bump where it stuttered weirdly, but other than that, nothing much. kinda surprised to see PCIE 1.1x16 run to well, it's basically PCIE 3.0x4

 

Edit: oh and GPU usage is shown to be capped even though it was pcie bus that was the bottleneck, which was unexpected

  1. Moonzy

    Moonzy

    for funsies, i did some more tests

     

    all CPU core at 1.5GHz

    Spoiler

    CPU_1.5.thumb.PNG.2185545777b7055ae59f4f192e6072ec.PNG

     

    only assigning 2 threads to SOTR (1c/2t) at stock freq

    Spoiler

    2_thread.thumb.PNG.ca0ce9df7c805572adb40c03feeb5b03.PNG

     

    set GPU to lowest i can go (58% power, -500 core clock (1900 -> 1300), -500 mem clock

    Spoiler

    GPU_58_1.3.thumb.PNG.1bfe8cf7c1e3b67639a850c46988f767.PNG

     

    honestly, even with 1.5ghz all core, the frametime is flat

     

    the only one with spiky frametime is the 2 thread config, it also has missing/pop-in textures as well as long ass loading times 🤔

    i guess higher thread count = more stable frametime. and single core performance can go to heck but still have stable frametime as long as you have threads

  2. Moonzy

    Moonzy

    after waking up with a fresh mind and looking at it again

     

    apparently pcie3.0 x16 and 2.0 x16 (3.0 x8) have a 10% discrepancy in "GPU Bound" stat, altho avg fps is bout the same

     

    maybe i should find a way to record frametime to compare both scenarios side by side better

×