Need some serious help here, guys.

Alright everyone so tonight I got busy with trying to figure out my driver issue first. It was irritating me. A friend suggested (like some of you) that I try to get windows update to work to try out some drivers.. so I did. It wouldn't work... repeatedly. I followed a guide online (found here: http://windowsforum.com/threads/windows-update-keeps-checking-for-updates-forever.73958/page-2) to get it to work. It finally did. I installed a bunch. My chipset stopped giving me issues with the driver (finally).

I figured my issue might have been solved so I decided to check the FPS. Long story short still terrible. My CPU usage wasn't going high enough and.. best I can tell... was bottlenecking my GPUs. I ran a CPU stress test (built into kombustor). Here are my FAIL results:

wtf2.jpg


I needed some confirmation though (a GPU only test) that the CPU was holding my SLI setup back. I was still getting issues with Kombustor telling me I didn't have OpenGL installed so I deleted my video drivers and put them on again. That solved that issue.

Then I ran a GPU test. Settings were all pre set for me. Here are the results:

kombustor1.jpg


The results (seemed) pretty good so I ran intel burn test. I don't have a screenshots of those results but long story short my CPU usage was high but the clock remained low.

Further confirmation still necessary. Ran a CPU intensive (no gpu) physx test:

kombustor2.jpg


One last test (by your request). Furmark (shown below). The first card was getting pretty warm so I bumped the fan speeds for both to 80%:

furmark1.jpg


After these tests I tried to OC my CPU via AI Suite to 4.0 GHz via a preset. No luck. Won't OC.

Thanks everyone for the help. Gather from the results what you will. I'm tired and irritated, lol. I found another power supply tonight (800W continuous). I'm going to test that hooked up to my system with a single kingpin card when I wake up. I also have to send in my old motherboard to Asus. Hopefully this isn't a mobo issue so I don't have to deal with them again.

EDIT:


I hooked both my video cards up to a voltage tester tonight. They came out normal. Doesn't mean there still isn't something wrong with this PSU though. The PSU I would suspect much sooner than the video card. Right now, by the way (just before adding this reply) I did get a voltage warning from AI Suite (something about Warning: 5.0000V). Not sure if this is a false warning or not. I have been idling (with intermittent restarts) all day, though with no problems.

Did you take a gpu-z screenshot 10 minutes into the test?
Non of the programs you screenshoted show with how many lanes is the gpu running at
 
Last edited:
Had a similar problem with one of my 290X's only to find it just wasn't making full contact with the PCI-E lane and wasn't fully seated (doh! was very embarrassing when I discovered it). Although, I wasn't taking a performance hit as big as you are reporting but it's worth checking GPU-Z.

Though it seems very unlikely, I would try each card individually in a different machine just to rule out the possibility of the cards being defunct.

Good luck! You've got more patience than me. I would have just given up by this point and bought new kit.
 
Last edited:
Tested again tonight with another PSU I recently found. XPower1080W. It's an off brand power supply but I'm sure it works. I used it in my old rig and swapped it out when I put new components in. Nothing is wrong with it for sure.

Same results. My usage seems to stop at 37%. Clock speeds stop at 1.18GHz (but mostly sits at 1.15GHz). I should mention that I have a 6970 connected currently. This power supply didn't have enough connections for all those 8 pin connectors on the 780tis. No point anyways in putting them in based on these results.

This test was ran for 2 mins or so. There was no point in dragging it out:

newpsutest.jpg



Now I'm even more confused than I was before... Two mobos, two PSUs, three different GFX cards. o.O
 
I know we've been focusing solely on your CPU and GPUs, but what are your Bus Frequencies set at?
 
looks like some crazy ass laptop power saving feature.
thats the kind of stuff that happens to a del laptop if you dont have the psu plugged in..

i REALLY doubt it is. but have you checked windows power settings?
put it all up to performance.

and you dont have speed step and turbo boost disabled or anything do you?
 
looks like some crazy ass laptop power saving feature.
thats the kind of stuff that happens to a del laptop if you dont have the psu plugged in..

i REALLY doubt it is. but have you checked windows power settings?
put it all up to performance.

and you dont have speed step and turbo boost disabled or anything do you?

Windows power settings I set myself they're OK.

Step step and turbo boost are enabled.

I reverted back to the December 2013 BIOS. Tested with default BIOS. Same issue. Did a manual OC. Same issue. Pulled all but two sticks of RAM out (I need at least two) and did a Memtest. Here's a snapshot early on in the test.

memtest.jpg


Now here is a similar snapshot of SOMEONE ELSE'S memtest from google:

uORZy0X.jpg


Now someone please tell me why he has 6 active CPUs found and started (with the same processor) and I only have 12 Found 6 Started and 1 Active? My memtest came out OK with 0 issues though. Processor is still clocked at 1.15GHz.

I know this sounds kind of bold of me to ask... but is there any way someone from OC3D (like Tom) can pull some strings with Asus to get me a NEW Rampage IV instead of some refurbished one that's going to come to my house broken (like this one?). I also thought Tom was gonna post in the thread after I changed the image tags? :(

At this point I'm all out of ideas.
 
Last edited:
Hi Exaisle, seems you are having some mega problems and I hope you eventually get them sorted. I have not got enough knowledge to help you with positive advice on how to fix the problem but I am good at spotting the difference.

I note the state of your CPU shows up different to what his is (is that because he is on test Number 8 and you are on number 7?) I do not think the percentage of how far the test has run would make a difference but it might be worth checking the CPUs on identical test sequences.

Good luck and I will be following this to see how you get on.

Tom is a busy man and no doubt will check the images when he gets a chance as I am sure everyone on this forum is racking their brains on how to help.
 
Thanks for the reply and the reassurance my friend. No doubt that 100% of what you said was true. Tom seems to be a really nice guy. I can tell by his videos and gifts he gives. He's OC3D santa clause :)
 
Why are we concerned about the CPU usage that much? Wouldn't running Furmark a good way to see if the GPU performs with limiting CPU needed as much as possible?
 
Did i read right that you're using furmark to bench these? ( I haven't got time read the entire thread )

Furmark is a *known* card killer, some cards have built-in power/clock-profile to deliberately and uncontrollably clock-down and you can't do *anything* to stop it, it's built in on some kind of low-level shit.

Furmark artificially stresses cards in a way they will never be used in the real world, I've lost a 580 to it, and I know of others who have had their cards damaged by it. It doesn't even have to be on long or get hot. It just fucks em up.

Do not, use Furmark to test/bench or burn-in your cards. It can, will and has probably already damaged something.

If you're not, ignore my post, I'm just passing through and skim reading.
 
Why are we concerned about the CPU usage that much? Wouldn't running Furmark a good way to see if the GPU performs with limiting CPU needed as much as possible?


ibt should get the cpu to 100% usage no issues at all.
for it to stay at idle speeds is just wrong.
(it seems to stay at idle speeds and voltages and only up to 32% usage)
if the cpu is stuck at idle speeds and voltages and cant use more than 32% of those idle speed there is no way it could compute and power the gpu(s)

personally i think its a issue.
the only things *software* that i have seen that can have this kind of affect on a cpu is a badly configured "throttle stop"
but there has been no mention of that ruinning. and after a restart "unless set up in a way so it boots up and applys the same settings" i dont really see how it could be an issue at all.

il just step back from this one as i am all out of ideas to be honest. and really dont know what to blame here.
But i think that the cpu isnt going out of idle speeds..
why that may happen is not something i can speculate on.
(im just assuming that from the pictures)

also seems HT is dissabled.. but i dont think that would cause the other issues..
and i dont think your board needs a bios update to use that cpu properly.. so like i said i'm out of ideas.
(personally id be going through the bios with a fine tooth comb.)
I would also reset the bios the old gygabyte stupid way..
"power down and unplug the psu"
jumper the cmos (or use the button if thats what it has) "remove the battery for extra measuere (not really needed but just to be extra sure)
try to power on without power connected
leave the system off for a good 10mins
replace the cmos jumper "if needed" put he battery back in.
turn the pc back on.
just let the pc boot to windows with default bios settings "dont change anything as it will use defaults"
Sit in windows for 2 mins.
tell windows to shut down
turn the pc back on
Now set the bios up.

(that was the only way to get the bios to remember settings on a old gygabite board after it decided you didnt know what you were doing and it had to restart due to wrong settings)

i know its conveluted and seems like 1/2 the steps are pointless... but it honestly is how it was.
if you didnt do that the bios would look like it remembered the settings. but it didnt actually apply them it just looked like it had.
 
Last edited:
Sorry for the late reply but I cut my thumb on a straight razor and had to get 4 stitches on my thumb today. lol. I tested some tonight. I did a manual OC last night to 4 GHz (without going through the auto one through AI Suite in windows that changes your settings in the BIOS) and came up with some weird results.

Here's the first screenshot. Notice my clock (4 gigs) and then look down one more line at the DRAM speed. 1373. Should be 1866 to match the speeds of my RAM. It's probably because I'm not in XMP mode but still odd. You will also notice I'm testing with the older BIOS from December that I mentioned downgrading to last night:

1.jpg


Next picture shows my core ratios (39) and the power saving mode being OFF. Nothing is odd about this picture but still figured I'd show:

2.jpg


This next picture is to show that both speedstep and turbo mode are disabled. Speedstep was disabled which I thought (along with the manual OC) would prevent my computer from being able to drop down clocks to save power in any state (even idle):

3.jpg


This next picture shows my speed as being 3303 MHz. I find this odd because it should read 4K MHz to match the clock. I'm not sure this number in the BIOS dynamically changes when you OC but just figured I'd point that out anyways. By the way the RAM only shows as 8GB because I'm only running two sticks. Since this is quad channel memory I need at least two plugged in. These two sticks are the same two that passed my memtest last night. I never put in the rest because I wanted to completely remove the potential of an issue with any of the other sticks (which I believe there is none).

4.jpg


I know one of you asked whether or not hyper threading was enabled. I just wanted to show this screen to you so you'd see that it is, in fact, enabled. All processor cores, as confirmed by the image, are also all enabled:

5.jpg


I thought it might be important to note after I was ready to hit F10 and save the BIOS changes, even though nothing had changed at ALL at this point as far as BIOS settings go, that the BIOS tried to tell me that I was indeed changing the core ratios values from what they were (39) to 39. Kind of odd. I tried to go back into the BIOS again and disable hyper threading (just to see if it would do it again) and it did. No matter what I change here it does the same thing. Not sure if it was a problem or not. I did re-enable hyper threading after this though:

7.jpg


The next image shows the LiveComm.exe process as suspended. I looked this up and apparently it has something to do with the Windows 8 app store? I figured it could possibly be important to show. Also note that in this snapshot my processor is at 0% usage. I thought that was kind of strange even for a near idle state:

9.jpg


The final image, though zoomed in MUCH further than I'd like, shows something pretty important. I went into the power options (sort of shown on the left) and made it so my processor must run at least at 50%. At idle it wasn't going to 50%. It was going to 37%. I ran an intel burn test and the speed never moved but 1-2 MHz off of 1234.0 in either direction (up or down). Task manager reads this as 37% usage @ 1.18 GHz clock. All cores are maxed out at 100%.

I know this was kind of long but I'm trying to keep you all 100% up to date on my findings. The results of these tests for me are inconclusive. I'd be much quicker to blame the motherboard before I blamed the CPU yet I can't be too sure. It's important to note that my boot times since the OC have been BLAZING fast. If I don't hit the delete key fast enough to get into the BIOS it literally zooms past to start windows and I can log in in about a minute. It seems like the OC is affecting my boot times drastically but slows down hard once windows loads. I just wish I knew what the HELL was going on here.

I don't know whether to go through with another RMA to Asus... try formatting again (which I think we all know what'll happen if I do), or just buy a 4930k.

--------------------------------------------------------------------------------------------


You might also find it amusing to know that I called Asus today for two reasons. One: I didn't know which address to send the other board to and Two: I wanted to see if a high level tech could help me find the answers to my problem (which I doubted). I thought it might be worth a shot since my dad helped the engineers at Asus (the actual engineers of the boards in Taiwan) and the RAM engineers at Crucial figure out a problem with the 2011 boards and that particular series of RAM we had trouble with. It got us faster RAM out of the deal so heck.. worth a shot giving them a call. At least if I got nowhere in the call I might have a better idea of a potential solution by the time it was over.

Making this assumption was my first mistake.

Got the address from the lady on the phone and, naturally, she wanted me to explain to her my issue before she escalated my case. Fair enough... I played along. She made notes (from what she said) about my case on the computer, updating my file. 20 mins later and 3 times waiting on hold later she tries to transfer me to the tech and I get disconnected (from her end). Call back and give the new lady on the phone my case # and wait for 5 mins on hold. She can't find records of my call. So then SHE asks me to explain my issues to her. 15 minutes later of me explaining to her everything I did she asks me for my motherboard type and CPU. 3 mins later she comes back and tells me the 3930k doesn't work with the Rampage IV since it's not on the list of supported processors (LOL). I tell her she's full of crap (not in those words) and she looks again...

Sure enough I was right (doh). She then tries to tell me it's 100% a CPU problem (she had no clue what it was). I tell her she's full of it. I get put on hold again. 3 mins on hold she comes back and asks me some more questions. I told her I OCd the CPU through ASUS' software from Windows. She then stops asking me questions and tries to tell me I've voided my warranty for OCing because my issues were probably caused by the OC in the first place. She told me I probably damaged vital components on the board. I reminded her that these issues were present at stock settings from the factory from the very beginning. Then when I tell her "why do you guys include OC capability in your software then tell me it's my fault for using it to test for problems?". It's kind of like buying a car with a no questions asked warranty. During the sale the salesperson tells you that this car has new technology built in for super durable and powerful brakes. Your brakes fall off when you use them to avoid an accident right after pulling out of the dealership lot.. then the motor company telling you it's your fault for pressing them too firmly so your warranty is no longer valid for as long as you own the car.

Then when I get extremely angry (I was calm until this point) and tell her I probably know more about computers than everyone working at the technician center and that Asus has terrible customer service she then proceeds to tell me she'll forward my information to her "colleagues" but she couldn't promise anything.

I've been ready to explode ever since that phone call. Sorry for that rant.

BRAIN = FRIED.
 
Last edited:
6 pages later and from what I see, I'm inclined to believe the others in that it is software related.

Plus.. I really dislike ASUS AI software. I wouldnt use it ever. Especially for overclocking. Any avid overclocker here or anywhere will tell you uninstall it and only OC from the BIOS. The software reports faulty temp readings sometimes such is mobo tempts being -160C.. good job on accuracy there.

Have you followed the overclocking guide on this website? there is a nice back to basics OC guide to help get you through the beginnings, and who knows maybe you have overlooked something that was easily missed.
 
6 pages later and from what I see, I'm inclined to believe the others in that it is software related.

Plus.. I really dislike ASUS AI software. I wouldnt use it ever. Especially for overclocking. Any avid overclocker here or anywhere will tell you uninstall it and only OC from the BIOS. The software reports faulty temp readings sometimes such is mobo tempts being -160C.. good job on accuracy there.

Have you followed the overclocking guide on this website? there is a nice back to basics OC guide to help get you through the beginnings, and who knows maybe you have overlooked something that was easily missed.

You'll get no argument from me on this :lol: I've always Clocked from the Bios since way back in the early 486 days, I've NEVER used prepriatary software to do it. As good as some software has got I.E CCC or EasyTune I still wouldn't do it.
 
i would honestly turn turbo back on..
and change base clock back to 100

i dont think your ram will change frequency when you change the base clock when you have clock gen full reset enabled (although im not that familiar with that)

also i would not want to use, try to use and if possible i wouldnt even have any cpu / bios value editing software installed on my pc.

i still dont get why it is not jumping from idle speeds/voltages. it should at the least jump up to non over clocked core values from idle ones.
Like i said earlier the only time i have seen symptoms like this was with throttle stop running and the wrong settings selected.

edit looking some more..

change the cpu core ratio to by core
and dissable cpu strap (make sure your back to 100 blck when you do that)

i would also make sure that intel speed step is enabled. having it dissabled on some systems can cause turbo mode to not work. "EIST"

all the above should get it reasonably close to being manually set and take a good protion of doubt off of the bios settings.
you could also set your dram manually instead of auto. id use the stock values for the ram not the xmp

and i think thats how i would have it set up to start it.

Basically.
im saying try over clocking just the turbo clocks for now.
enable speed step so the bios can control changes with cpu load
have the cores over clocked in turbo mode via by core "i never had any luck with using 1 ratio for all cores. but manually setting all cores the same works"
set the ram to default loose timings so you wont run in to any issues with that.
dissable cpu strap so everything is definatly working on the same base frequency

and uninstall ai suite thing. any software you may have that can try to adjust cpu speeds should either be uninstalled or not running.
 
Last edited:
I also see that your baseclock is set to 103 MHz. Set it back to 100MHz then check the PCIe lane clock. If its not 100MHz set it to that clock, then set the memory multiplier to 18.6 then you will get 1866 MHz of memory.

Then start windows and make a screenshot of cpu-z and gpu-z running. Then post it back. And we will see.

If you have enought time run an OCCT bench for 30Minutes with linpack, if it fails you will get graphs and logs as well. Maybe you can find the aswers in those logs.

ASUS has never had a good customer service....

:)


Good luck and waiting for the results.
 
HOT DIGGITY DAMNIT!!!! We have confirmation!

THIS IS A SOFTWARE ISSUE!!!!!!!!! You guys were right after all. For the love of GOD!

It just so happens that as I was about to check on your suggestions (messing with BIOS settings) I found the answer to all my problems. Tonight I tried to boot and realized my screen was flickering black. Looks like a bad sign right? This is after I installed windows 8.1 to test and uninstalled the NVidia drivers. I had to boot into safe mode just to get graphics on screen.

Since some of you guys suggested I look for software issues I said "why not try in safe mode?". I booted into safe mode with networking and hit CTRL+ALT+DEL and started task manager. Went to the performance tab and noticed my clock speed was 3.30 GHz on IDLE!!!!!!!!! Now comes the weird part.. I ran an intel burn test and came up with this:

cpuusage.jpg


Notice usage is 100% (like it should be) and the clock is still 3.30GHz. CPU-Z (although not fully functioning in safe mode) reports a clock speed of 3298.05MHz. This explains the software issue. What's causing it.. I'm not sure. I'll uninstall AISuite and see what happens. The only thing I do know is that my clock speed SHOULD be 4 GHz based on my BIOS settings. I'll see what I can find.
 
Last edited:
Back
Top