Abnormally low benchmark scores and keyboard lag on start up

[Paladone]

Am I the only person who noticed that the OP is running the benchmark at 1080p? I thought the 1080 was already pretty much CPU bound in just about every situation with a single GPU at 1080p, forget a second one. The 6900k makes it that much worse because you have more lower clocked cores. The OP should be using a 4k monitor for something like that.

I plan on getting a 4k monitor when I can afford it I am not sure which one, but I was thinking maybe an Asus MG24UQ 24".

 

[DeathtoGnomes]

Ahh another blunder by manufacturer software installations, here I thought only ASUS had shit support software like GPuTweak 1/2, and whatever that other bundled overclocking shitware that had 16 some-odd thousand versions of special uninstall programs.....

So. Lesson here, dont install "that kind" of software until the drivers are all installed and running.

 

[Paladone]

It seems to be fixed. The issue was that the power saver mode was on balanced. Although I changed it to high performance, I forgot to change it back after reinstalling windows. I scored 5507 in heaven and 6609 in valley. I feel like such an idiot lol.

 

[trog100]

Hey guys,

So I only finished putting my PC together a few days ago, but have been having some issues with it since.

The first thing is that after logging in, for around 20 seconds or so there is a considerable delay in response time between my keyboard and the PC; I'll hit the windows key and the start menu will pop up half a second later and the same goes for regular typing. There however is no lag when opening programs or with the cursor.

The second issue is that the benchmark scores I've been getting are abnormally low, I attached a couple of images:

For Time Spy, I got 13561 graphics score (test 1=85.10 FPS, test 2= 80.48 FPS) and 9872 in CPU

________________________________________________________________________________________

My Graphics cards, which are EVGA GTX 1080 SC ACX 3.0 cards, both on water (as is the CPU). Using EVGA precision X, they seem to be stable at 1830/1330MHz with 120% power consumption. They idle at around 30 degrees celsius and during heaven averaged at around 50-55, peaking at 65. In valley however they generally stay below 40. In time spy, I wasn't able to check because EVGA precision X prevented me from running the benchmark for whatever reason. During Heaven, the VRAM usage was around 1.5GB, and around 2.5GB during valley if I remember correctly.

If it matters, I'm using an EVGA Pro SLI bridge with 2 slot spacing and the cards in lanes 1 and 4. I am not sure why, but the card in lane 4 was automatically made the primary card. (Or at least I think it is; the monitors are plugged into the card in lane 1, which shows up as card (2) in the Nvidia control panel).

SLI is enabled with the setting "Maximise 3D performance" and the PhysX processing is set to auto. I have 3 monitors, all are 1080p.

________________________________________________________________________________________

My apologies for the essay, I am just trying to provide as much information as possible as I haven't a clue what is causing the issues.

Any suggestions are welcome,
Thanks

EDIT: I forgot to add, my heaven score isn't too far off what it should be; maybe around a thousand point? The valley score however is way below what it should be and closer to what you'd get with a single 1070.

EDIT 2: I disabled SLI and ran valley and heaven again. With a single 1080 I score 2864 in heaven, and 5373 in valley (Effectively the same in valley with as with sli on, but around 60% as high as in heaven with SLI on). I don't understand what would cause that.

I removed the CPU part as the issue is resolved; it turned out to be EVGA precision XOC causing the startup lag.

From what I've seen online, people with single 1080's are getting around 3000-3200, while I'm only scoring 2850. I cannot find any for dual 1080's, but 4600 seems to be even less than what you'd score with two 980 TI's.

i get the same thing in valley when i run 1080 sli.. my 1080 score is only sightly more than my 1440 score.. there is something wrong with how it reports.. i have no idea why..

i am running a pair of palit 980 tI cards and i use the plait software.. no lag problems just the low valley score at 1080.. i dont have anything to compare with but at 1440 at get around 5400 with around 5700 at 1080.. clearly the difference should be much larger..

heaven seems to report okay valley dosnt..

trog

ps.. i run in balanced mode never seen a problem elsewhere with it though.. only with valley.. which begs the question.. why valley..

ps.. 2.. with valley at 1080 i see about 17% cpu usage so it aint cpu bound.. no way.. least i dont see how it can be.. though i will admit sli scales much better at higher resolutions than it does lower ones.. my time spy score is around 9000.. everyday settings..

 

[Paladone]

i get the same thing in valley when i run 1080 sli.. my 1080 score is only sightly more than my 1440 score.. there is something wrong with how it reports.. i have no idea why..

i am running a pair of palit 980 tI cards and i use the plait software.. no lag problems just the low valley score at 1080.. i dont have anything to compare with but at 1440 at get around 5400 with around 5700 at 1080.. clearly the difference should be much larger..

heaven seems to report okay valley dosnt..

trog

ps.. i run in balanced mode never seen a problem elsewhere with it though.. only with valley.. which begs the question.. why valley..

ps.. 2.. with valley at 1080 i see about 17% cpu usage so it aint cpu bound.. no way.. he he

That's odd. I don't know just valley for you, but putting it in high performance mode boosted all my scores significantly.

 

[trog100]

That's odd. I don't know just valley for you, but putting it in high performance mode boosted all my scores significantly.

to be honest i have always left balanced mode on.. most of my benchmark scores seem about right.. i have a good idea of what they should be.. valley is the only oddity i have noticed.. i assume windows has the brains to know what what to do and when.. but i could well be wrong.. he he

i do know cpu clock speed makes next no difference with valley or heaven.. also the vram usage is very low.. its set to not use much.. the very low minimums are caused by it stopping dead for about a second to load its next set of scene textures from ram.. they paint a false picture..

it also reports the boost speeds wrongly..

trog

 

[FR@NK]

high performance mode boosted all my scores significantly.

Do you have the turbo boost driver installed?

 

[Paladone]

Do you have the turbo boost driver installed?

I did, but I uninstalled it.

 

[Kanan]

It seems to be fixed. The issue was that the power saver mode was on balanced. Although I changed it to high performance, I forgot to change it back after reinstalling windows. I scored 5507 in heaven and 6609 in valley. I feel like such an idiot lol.

Performance just sets the CPU to max. speed all the time, that's the only difference. So there was something not working correctly with your CPU before, normally "balanced" and "performance" should work ~equally, because "balanced" is supposed to clock your CPU to max. state if you start a benchmark or other demanding app. Maybe you should install that Intel Turbo Boost 3.0 app again.

 

[kn00tcn]

Performance just sets the CPU to max. speed all the time, that's the only difference. So there was something not working correctly with your CPU before, normally "balanced" and "performance" should work ~equally, because "balanced" is supposed to clock your CPU to max. state if you start a benchmark or other demanding app. Maybe you should install that Intel Turbo Boost 3.0 app again.

you can even control the min/max states as a percentage in the windows power settings, resulting in no need for the high performance profile, just customize the balanced one to your liking

regardless, things need to be logged: the cpu mhz, gpu mhz, cpu usage, gpu usage... there's no reason to turn into some voodoo magic install/uninstall of programs/drivers, stay scientific