http://www.3dmark.com/3dm/18792092 7689 , Still quite happy with the result the videocard is even clocked to 2050Mhz so not that high.
I'm getting the same ~6400ish GPU score with my 980ti @1500/7600. You might want to try with lower VRAM clocks. At 8000 I'll score lower than 7600 on my card as it's unstable (yet not unstable enough to crash). http://www.3dmark.com/spy/1476711
Ive noticed this with the core speed of my 1080T, until the voltage control is working in Afterburner il leave it at +100
@pegasus1 I think it makes sense. I suppose the GPU sends corrupted data to VRAM that then tries to autocorrect as more CRC checks are failing. Performance can only suffer.
Just ran Time Spy on my new Gigabyte Aorus GTX1080Ti and got a graphics score of: 9763 No overclocking, power limit or core voltage adjustments, only a custom fan on Auto.
Testing old drivers versus new drivers , # 1 test with good old trusty 378.92 = GTX 1080Ti @ 2126/1539Mhz http://www.3dmark.com/spy/2119126 Will test latest drivers released today 384.94 tomorrow night to compare
LOL, I can run all other 3dmark stuff's, but I get error's upon loading up Timespy, which is why its not on my playlist yet. Sometimes it works, then sometimes it will quit very unexpectedly. Good program though.
Might be related to your dual cpu set up , send Futuremark a bug report to see if they fix it , i would love to see your score on that beastly machine of yours
Actually I got it finally to work this time. Pinpointed the main problem or problems. Apparently whats causing this issue is the fps meter programs such as fraps,Action Recorder, bandicam,etc. They all doing the same thing, it does NOT work with these programs, moreover, With precision XOC, it does NOT work with it, even if I put the memory and gpu speed at +0, All dont work but one program did: The one and only Msi Afterburner. Works perfectly with this program but can't get the pwr fan(fan on card itself) to work, need xoc for that but doesn't matter, got the bench finally to work and that's all that matters in the end. My score at default settings, using one card only, over 9k, but thats with using the shadowplay recorder at MAX settings, so score will be higher without it, haven't tried without it, will soon, Will give update later. I did sli mode and as expected, the sli usage aint 80/90+% on both gpu's, more like upper 70's, the settings and res are just to low for it, so I only did one gpu, plus there isn't many bits selection to use within nvidia inspector due to the nature of this program being DX12. I did do a custom run/video record(2 1080ti) @ 4k(2160p) MAX for kicks, will be posting that on my playlist in comparision with my old 2 980ti hybrids. Sadly, in order to view the final net score is only through default bench run from what I understand, not custom runs. UPDATE: Single 1080ti without video recorder turned on, final score: Over 9.7k @ 2038/12.5k Good cpu utilization but used no where near 100%, not even close but decent amount, about half usage per core, ab only shows up to 32 though, so I don't know if this program can use more then 32 cores/threads,etc. Since 32 cores and under is what locked for 32bit applications, since this being 64bit, it could have used more, hard to say, graph showed up to 32, even if it did used more, more then likely only used less then half cpu usage per thread/core during cpu bench, absolutely used no where of its potential of these cpu's, but decent amount per core.
