PC Powers OFF during 3dMark : HELP Please!

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530
I have just completed my first solo pc build. Now I am in the testing phase and having a big problem. I started with the Prime95 blend test and the pc performs well. It maintains temperature and does not crash. Now I am running 3dMark and during certain tests the pc just powers off and reboots. The event viewer only contains a log for an unexpected shutdown.

Ice storm - all tests run to completion with FPS in the thousands.

Cloud gate - pc powers off immediately on the first test.

Fire strike - first test runs with over 100fps. Pc powers off and reboots near the start of the second test.

Here are the components in my PC, running Windows 8.1:

http://pcpartpicker.com/user/zoop/saved/3NzE

Intel Core i7-4770K 3.5GHz Quad-Core
CPU Cooler Corsair H110 94.0 CFM Liquid
Thermal Compound Arctic Cooling MX4 4g
Motherboard ASRock Z87 OC Formula/ac EATX LGA1150
Memory Corsair Dominator Platinum 32GB (4 x 8GB) DDR3-2133
Storage Samsung 840 Pro Series 256GB 2.5" SSD
Western Digital Caviar Blue 1TB 3.5" 7200RPM
Video Card EVGA GeForce GTX 780 Ti 3GB
EVGA GeForce GTX 780 Ti 3GB
Case Rosewill BLACKHAWK-ULTRA ATX Full Tower
Power Supply SeaSonic 1000W ATX12V / EPS12V
Optical Drive Pioneer BDR-2209 Blu-Ray/DVD/CD Writer
Operating System Microsoft Windows 8.1 Pro - 64-bit (OEM) (64-bit)
Custom NZXT Fan Controller

Just prior to the power going I notice some sort of wheezing or clicking sound I believe to be from the GPU. I observed the same behavior with SLI enabled and disabled.

Please help me out here fellows.. I am totally lost and unsure what to do next. This pc represents a considerable investment and I really want to figure this out from both a personal pride standpoint as well as a monetary one.

Thoughts? I've read elsewhere that the pc powering off must be psu related. I am running a seagate 1000w platinum. Is that enough for my build? I thought I was giving myself some breathing room by choosing a 1000w but maybe no?

Thanks everyone... I really appreciate and advice and help I can get on this one. I'll run and suggested tests, I just want to know what is going on here. This pc should be a beast: not die on a cross platform stress test :(.


Thanks !
 
Solution
you dont happen to know anyone with a power supply that can run your rig to test it out do you? btw dont think i said but this is one beast of a rig. bet BF4 looks sexy on High resolution/triple monitor setup.

Bad_Kitty13

Admirable
i would say it is something to do with psu. the way i would trouble shoot it is to run the test on a single card removing the second, then switch the cards out and run it on the second, if they both do fine then run them sli again. if it fails=psu...if one of the cards fail = gpu. and dont panic its a new build so you should still be in your rma window. although an inconvience at least you know your covered. also try running gpu-z on a second screen to see if you are getting abnormal voltage spikes.
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


I have not tried FurMark yet, I will download and give it a run.
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


I tested using only a single card installed in the case at a time. Per your suggestion the second card was removed entirely from the case. The test runs just a few seconds longer, but the end result is the same: a system crash / reboot without any warning or messages. The behavior is identical regardless of which Graphics card is installed.

When the test starts I do hear a high pitched sound coming from the GPU.

Does this seem like a PSU issue? Is there any way to test? Should I move forward with the RMA on the PSU now?

I will see about running GPU-Z. At the moment I only have a single monitor in the house but I'll see if I can dig something up. FWIW the cards were not hot (hardly warm) to the touch when I removed them immediately after powering down the PC.

I'm a complete newbie on this stuff and have no clue. Thanks for all the advice thus far. I intend to utilize every bit of it.
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


Since I do not have a second monitor to use at the moment I configured GPU-Z to write to the log every .1 seconds. When I look at the log following the crash I see the following:

VDDC goes from .9 to 1.2 before the system crashes and the log ends

TEMP goes from 31c to 51c before the system crashes and the log ends.

Is there anything else I should record that is relevant. I'm not sure what constitutes a voltage spike so I'm not sure how to read this data at all. If helpful I am willing to post the contents of the log. This log was captured using a single GPU (780ti)
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


Thanks, i'll try that. I have the latest drivers, but not BETA. I will give the BETA drivers a test and report back. T

Thanks again for sticking with me on this; I'm seriously at a loss without this community.
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


I have installed FurMark. What should I be running? I have the windowed mode burn in test running for the past 15 minutes without issues on a single GPU. How should I use this utility to max effect?

 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


I have come up empty on my search for alternate drivers. When I use the tool on NVIDIA it gives me only one choice (latest) for the GTX780ti on Windows 8.1. I'm about to do another search for the BETA's.

I did double checked and I am running the latest release version of the NVIDIA drivers. Provided I don't have any issues running p95 and Fumark together how concerned should I be?

Thanks again !
 

DelroyMonjo

Distinguished
Don't entirely discount that it COULD BE the PSU. I had an Antec 750W HCG which I was using on aASrock Z67 mb, i5-2500K and a GTX670 and kept getting dumped from games with a variety of errors. I tried other video cards, driver, updates, switched/changed RAM, etc. Built another rig with an ASrock Z87 Extreme Pro mb, i5-4670K and the same GTX670 GPU....same problem! Games were crashing for no reason and even the web browser would crash occasionally. Finally, one day I was fooling around with the case open and wiggled the 24 pin power cable from the PSU to the mb...system crashed! Tried this several times, same results. Replaced PSU with an Antec 650W and has worked fine since then.
So, maybe if you have another PSU........
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


I have finished a FuMark benchmark for 15 minutes (burn in test) and here are the results in SLI:

33xaqzb.jpg


I am going to try running some games and see what happens. I have not had any issues outside of 3dMark. Perhaps I should look in to RMAing the PSU just in case? Is there any way to stress test a PSU?

 

Bad_Kitty13

Admirable


The thing that made me think it was the PSU in the first place was that the whole pc powers off when it starts drawing to much power on the gpus, if it was just drivers wouldnt it just crash the display and the motherboard/cpu/odd/etc still have power going to it? i know seasonic is a top of the line PSU but sometimes even the best of the best have assembly errors. also it not crashing in other stress test says drivers are good and that it is in issue somewher with power draw?
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


Both suggestions seem good to me. I am going to check every connection this evening. I have submitted an RMA on the PSU, but I might try to find someone local with a 1000w or higher modular unit so I don't have to destroy my cabling job before returning a product that may/may not be defective. I wish there were a definitive way to tell what the PSU is doing.

I concur with your hypothesis. It does seem like only when the power demands get extreme the PC shuts down. That said, I just finished playing BF4 for over an hour on the highest possible settings (single play / campaign mode) and did not encounter a single issue.

I think I am going to run a mem test just for kicks to rule out the RAM as a possible issue as well.
 

Bad_Kitty13

Admirable
you dont happen to know anyone with a power supply that can run your rig to test it out do you? btw dont think i said but this is one beast of a rig. bet BF4 looks sexy on High resolution/triple monitor setup.
 
Solution

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


Big Update!

I noticed a setting in my BIOS that would turn off the power if the MB was getting too hot. When I turned this setting / protection off I was able to run all 3dMark tests without issue.

I'm not sure how to check the MB temps during the test. Do you know of an application that will create a log file including MB temps while I run 3dMark.

It seems unusual to me that the MB would hit a critical temp so quickly (within 1 second) of running 3dMark. A red herring? Any thoughts or suggestions? Should I just leave the setting off during benchmarking but turn it back on during normal usage since nothing else has shut down on me spontaneously yet? Is this indicative of an issue with the way I installed the MB?

Is it possible to have too many power connectors running to the MB? I have two power connectors side by side on my MB that are basically the same description in the manual. I plugged in both; is that correct?

1. ATX 12v Power Connector (12v1)
2. ATX 12v Power Connector (12v3)

I do not build PC's... and this manual does not include documentation around how to connect the power cables so I just plugged everything in to the correct spots. The two above are going to the CPU power slots I believe.

Any help or suggestions, as always, are very much appreciated. I really hope I'm not frying my MB. I just don't see how it could possibly heat up so quickly as to cause a shutdown.
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530


Thanks badK, I am looking for a friend with a PSU capable. The highest most of my buddies have gone is an 800W. That said, please see my post above about the MB / BIOS setting when you have a chance.

I just finished the 3dMark benchmarks and scored in the 99th percentile in all except the Firestrike Extreme test where I lag behind in the 95 percentile :p. More than good enough for me benchmark wise... now I just need to figure out if this MB is running hot or not. It seems odd that only the 3dMark application makes it go past the threshold and does so almost instantly. I'm hoping it's just a goofy quirk but it will drive me crazy until I know for sure.

Thanks again for you input, advice and feedback. If I ever have a chance to return the favor rest assured I will.

Oh, here are those results for anyone interested :).

http://postimg.org/image/52o8xa41j/

It's a larger image so I am posting the link and not embedding it.
 

Bad_Kitty13

Admirable


no problem and im looking into it now
 

zoopnfunk

Distinguished
Mar 18, 2008
35
0
18,530
I was able to get some MAX/MIN temps for the MB while running 3dMark. Are these values anything that should concern me?

1. SYSTIN - 33c
2. CPUTIN - 35c
3. AUXTIN - 41c
4. TMPIN3 - 55c

I don't know what these mean, but those are the readings I was able to get. I also noticed that the CPUID HWMonitor does not always load all of my HW. Prior to changing the BIOS setting for auto shutdown it would only display SATA devices essentially. I had to start it up twice this time in order to get all components to identify. Indicative of a MB problem or just glitch software?
 

Bad_Kitty13

Admirable
just looked at it. if it is the option in the h/w menu of your bios it is over temperature protection. it is used as a safety so your cpu doesnt get to hot. try to use this to monitor your temps on your cpu.

http://www.cpuid.com/softwares/hwmonitor-pro/versions-history.html

however i would say most likely that since it is shutting down, your heatsink from the puck to the cpu isnt making good contact or is not seated properly maybe the way you applied your paste. look into this try reseating and dont run with your protection disabled at least i wouldnt recommend it.