[SOLVED] 2950x +2080ti = slow

well i have now done a ton of research and test of the various modes the cpu is capeable of

i have testes wildlands, FC5, FC Primal, Destiny 2 forsaken all at 3440x1440 ultra

Memory access mode Dynamic or local makes quite a difference throughout most games, with (L)ocal being fastest in 16 core mode
disabling multi threading resulted in a few extra fps in both 8 and 16 core mode

gaming mode limits cpu to 8 cores, but most cores run at 4.4ghz more or less all the time resulting in 98 fps in FC5 (overclocked to 4.5ghz i got 100fps)

using creator mode 16 cores local access, no multithreading have the cores peak at 4.3 but usually levels out around 4.2 and i think that is why framerate drops to 91 fps

my ram is 3200, 4 blocks and quad channel is enabled and working


FC5 have 4 cores go to 100% so bottlenecking....

FCP same

D2f 4 cores to 80% not bottle necked....not sure whats holding this game back framerates in gaming 8 core no multithreading, swings between 95 and 155 fps but hovers around 115-120 most of the time (there is no benchmark mode)

wildlands hovers around 80% mark too with a couple high spikes on cpu , but as long as local access mode is on , then it gives a max of 58 fps (44 in dynamic mode)

theres a huge difference between cpu setting, but gaming never really gets super fast

i even tried moving gpu around to different slots to see if that made a difference...on my motherboard it does not...i did read the MSI MEG motherboard likes gpu in the 4th slot

i tried moving ram over to slot 1-3-5-7 (manual suggest slot 2-4-6-8) no difference

i ran sony vegas to render a clip i got, as long as nvenc is on theres little difference between 16 core 32 threads and 8 cores 8 threads...its 5:16 vs 5:25

i tried Ashes of the Singularity Escalation and at max settingts and it avg 82 fps abut 20fps faster (25% than 1080ti), this game likes threadripper and 2080ti (guess its part of dx12 to spread out cpu needs differently than dx11)

all drivers updated, im at a loss that such a powerful cpu is bottlenecking a 2080ti, maybe it all simply comes down to ghz and IPC

im seriously considering downgrading to a 9900k, take the hit when rendering(it aint much), but gain fps when gaming

so heres the results of the 9900 vs 2950x

i went out and bought a 9900k and gigabyte z390 master ....same ram...same cooling loop, with xspc coldplate....2080ti at 1770mhz, 3 nvme ssd's...no spinning storage....3200mhz ram

first number is amd in the best mode for what ever im testing...so some is 16c some is 8, some with hyperthreading some not....last number is 9900k@5ghz all cores

3440x1440 resolution, max settings

farcry primal is a draw at 103 vs 104 fps

FC5 98avg vs 106

D2 winding cove 100-144 vs 118-165 much more consistent though around 130+ fps....amd had lots of dips

wildlands....58 fps vs 62

sony vegas rendering hvec clip in 4k(nvenc).....5.29 vs 5.37

ashes of a singularity 85 vs 79.....this games runs dx12 and spreads out over all 16 cores

cooler can barely keep this intel cool....it hit 85c (prime95avx@5ghz all cores) .....same loop could keep the threadripper at a cool 62 max.....must be the tiny surface the intel has vs threadrippers massive chip

so there you have it intel wins, between 7 and 15% better.....

its been a week now....and it might not seem like a big difference in avg framerates, but intel feels a lot smoother than the amd did...i sold the threadripper and motherboard...went back to intel for now

only disapointment is the gigabyte board, i couldnt get the ultimate version, and "settled" for the master....which seems cheap compared to my previous boards