PDA

View Full Version : ASUS ROG Maximus X Hero (Wi-Fi AC) Every Game Crashes.



Khronos
11-29-2017, 12:41 AM
Brand new built computer runs great but when I play games it crashes to desktop after like 10-60 minutes or so, changes each time. Drivers are up to date, I did turn off Asus's core enhancer or whatever it was called because it was pumping way more voltage than i needed... like upwards to 1.4+

Bios 0802
All Drivers from rog.asus.com

Specs: i7 8700k Stock till I figure this out (EKWB Supremacy MX Block)
ASUS ROG Maximus X Hero (Wi-Fi AC)
32GB (4x8GB) G.SKILL TridentZ RGB Model F4-3000C15D-16GTZR XMP CAS 15 at 3000Mhz
Aorus 1080ti Extreme (EKWB full cover block w/ Backplate)
EVGA SuperNova G2 850W Gold PSU
2x 360mm Radiators w/ Vardar Fans
EK-XRES 100 DDC MX 3.1 PWM with an upgraded 250mm Reservoir.
BitFenix ALCHEMY 2.0 PSU CABLE KIT for EVGA SuperNOVA T2/P2/G2/G3
250GB Samsung 850 EVO SSD(OS)
500GB WD Blue M.2 Sata Drive (games)
2x 2TB WD Green 3.5in Mechanical Drives for video and spill over games

*I do want to add that even with this much radiator space my temps seem really high even after re-seating the CPU and better thermal paste... my GPU is cooling nicely never higher than 41c and stays around 2034Mhz core clock*

Hopeywolf
11-29-2017, 01:28 AM
Is there any error logs or driver crashes in event viewer when you check? Or does it just go straight to desktop without any indication at all?

Khronos
11-29-2017, 01:33 AM
Straight to desktop normally although just now it blue screened on me for the first time while playing starcitizen and gave me a stopcode of :PFN_LIST_Corrupt
I'm currently researching what this means atm.

*update* seems to mean the page frame number (PFN) list is corrupted. Again looking into this further at the moment*

Khronos
11-29-2017, 10:49 PM
**Update 2**

Re-installed windows from scratch and installed all the drivers. Ran what seems like a full day of benchmarks between 3DMARK, PCMARK 8, Prime 95 and did not crash a single time except the last time I ran Prime core 0 crashed but the rest of the workers continued. After that I loaded up a game (DOOM max settings in 3440x1440 and sat still for a good 20m before running around for a bit and killing demons but crashed soon after. what i did notice was my CPU not (Stock settings minus ASUS Multi-Core Enhancer turned OFF) hitting 4.7Ghz in any of the tests and stayed around 4.5Ghz max and was not limited by temps or voltage since I left it on AUTO. Now, for the first time I did in fact get a crash message which I will Post the 2 pictures below.

https://imgur.com/2CRHD5B
https://imgur.com/Ulkdvp9

*one of the Stress Test Benchmarks*

https://www.3dmark.com/fsst/644584

It seems like its a potential Memory problem? My web browser (chrome) crashed while i was writing this the first time and would not let me refresh the page and had to do a full reset on the computer for it to load again. I was honestly hoping after a full battery of benchmarks with no crashes it would keep on going.... This is getting Frustrating and i'm sad I cant even use this computer for what I built it for...

daddyd302
11-29-2017, 11:23 PM
Have you tried other nvidia drivers? Maybe dial back on your gpu clocks? You mioght need to underclock it even to see if it's stable or not. This doesn't sound like it's motherboard, cpu or ram related. It seems the issues is with your video card/drivers. The only other thing I can think of is your 1080ti is not getting enough juice from the psu? Are you using one cable or 2 to hook your video card?

This is what i can find out about your crash data.
https://answers.microsoft.com/en-us/windows/forum/windows_7-gaming/c-visual-runtime-error-r6025-how-to-fix-this/aa073255-f2bd-4c36-b451-6348d65051b7?auth=1

Khronos
11-30-2017, 12:05 AM
I'm personally not even overclocking the GPU that's all Nvidias GPU BOOST based on temps, I guess ill manually set it to base clocks and revert to a previous Nvidia Driver (It was updated using GeForce Experience) as you suggest. As far as my 1080ti i have 2 separate 8 pin's coming directly from the PSU in VGA1 and VGA 2 ports though, one would think a 850w PSU would be enough for 1 card. After i post this i will read the link you added.

*Side Note* Can you even turn off GPU Boost?

daddyd302
11-30-2017, 02:06 AM
I"m not sure you can turn off boost. If you have MSI afterburner you can try to adjust the clock speed there. The latest drivers aren't always the best. The latest might work for me, but not for you and vice versa.

I mentioned the cable because I used to use 1 cable with 2 plugs for my 1080 and Overwatch would sometimes crash. Sometimes I can play for hours and nothing happens, other times it'll crash after 20 minutes. I'm using 2 cables since finding it that it was a bad idea to run with 1 cable and on crashes anymore.

The only way to know for sure if it's the gpu or the psu is to swap it out with a backup, but most people don't have backup laying around.

Menthol
11-30-2017, 02:50 AM
Is your power supply a single 12v rail, if multiple rails make sure your GPU is on it's own rail.
What other software do you have installed, AISuite? Game first? Do you have an OSD running?
Try disabling all excess software including GF Experiance to see if it's a conflict of some kind. If your using GPU Tweak, disable it also until you get it diagnosed and corrected.

Raja@ASUS
11-30-2017, 04:43 AM
PFN list errors are often memory instability related. Revert the system to stock memory frequency and check if it still occurs.

Khronos
11-30-2017, 10:40 PM
last night when i was trying to game again seeing if the previous attempts were successful in fixing my problem i got 3 new BSOD of different things which lead to some tests i did in the content below.
CRITICAL_STRUCTURE_CORRUPTION
MEMORY_MANAGEMENT
and the last one i got was
SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (what failed nvlddmkm.sys)

@Daddyd302 - yeah before in my previous builds i used to use a single cable for a 6 + 6 pin or a 8 +6 but after i had read and saw benchmarks of better performance using 2 separate cables but now that i use a modular PSU i have the option to use as many as id like (i.e 2 different 8 pins coming directly from the PSU). I do have the original cables that came with the PSU and so i will try with those when im not working my 12 hour shifts. i bought the bitfenix white cables for my build for aesthetics but they are the complete replacement and not extenders, just wanna point that out.
(https://www.newegg.com/Product/Product.aspx?Item=N82E16812493209)
i did buy a lower wattage run for pump testing purposes so i will try that although the 650w kinda makes a tad nervous. I also do have my old 980ti classified from EVGA but since i was watercooling with hard tubing and the size difference of the classified card makes it hard to fit it in with my look. (i know i will have to disassemble my loop i just dont want to, you know >.<

@Menthos - if im not mistaken itsa single 12v rail in my EVGA 850w G2 gold rated PSU.i did have AI Suite installed, not sure about gamefirst and i do have OSD running from MSI afterburner to keep an eye on temps although none of the Asus software is running besides the Aura RGB software but that is recent before when all this started nothing was even installed besides base drivers from the Asus Website. i do use GeForce experience for my FPS gauge though. Roger on the disabling of the software i will try it tomorrow after i get some sleep (working 7pm to 7am).

@ Raja@ASUS - i have reverted to stock memory settings (w. XMP settings) and it still does the same thing, i also ran the most extensive Windows Memory Diagnostic Tool test you can with the settings of Extreme and 5 passes and i forgot what the last setting was but i think it had to do with the cache on each stick, i did turn that on to test. that test took like 8+ hours it was crazy i passed out and woke up waiting for my results and nothing came up ( in the Windows Memory Diagnostic thing it says it will restart and show your results but never did get a result once i logged in with my password, however after 4 passes out of 5 that i was awake to see it said no errors detected in the memory. I also ran a Check and possible Repair Windows System File using sfc /scannow command in the command promp and it came back with:

Verification 100% complete.
Windows Resource Protection did not find any integrity violations.

and so now im kinda stumped on where to go now besides the advice given above, Also i want to say thank you for everyone who has pitched in so far! I appreciate all the help, thank you.

Raja@ASUS
12-01-2017, 05:45 AM
I'd try hci memtest. If asus multicore enhance is enabled, disable it. You may want to try one memory module at a time also.

Khronos
12-01-2017, 11:28 PM
so today i changed all the PSU cables to the original and it crashed 2 times both times i was testing it in starcitizen first crash was 14m after launch and the second 7m after launching the game both actually had a crash report but i could not see it since it sent it directly to the game company. At the moment i changed the PSU and put the ram at stock settings without XMP profile, i was a bit worried being this other PSU does not have 2 different VGA connnectors and instead only has 1 cable with 2x 6+2 pins. I also disabled all monitoring software during the previous testing. once i test this PSU and if it fails i will disassemble my loop and try my 980ti that is not watercooled. If that does not work my last action might be to RMA everything and hope for the best... Im worried for the availability of the i7 8700k since its been sold out everywhere and i had to wait like a month just to get mine. i did run a DXDiag which i will upload here just in case it can help solve this problem however it said it found nothing wrong.

https://docs.google.com/document/d/1PvNEsofdZ2qnO35SrLLDZTOT7mK7fSGsx65FYN6zY_E/edit?usp=sharing

*edit* Im running HCI Memtest now with all RAM installed and if that does not pass i will run it with 1 stick at a time in each slot to make sure the RAM slot on the Motherboard is not bad. will take a bit since i have 4 DIMMS and 4 Slots. 16 tests in total.

Khronos
12-02-2017, 01:34 AM
I'd try hci memtest. If asus multicore enhance is enabled, disable it. You may want to try one memory module at a time also.

How long do i let it test? i know the program said to wait for it to hit 100% or 400% for the most rare errors and unless im reading the small coverage % at the bottom of the program ive ran 6 different RAM test at the same time and split up the total available RAM per the softwares request and they are all at 30,718% coverage with 0 errors across all 6. i guess now i will test each one individually.

*update*

after changing PSU's ive sat in Star Citizen 1h 30m so far that ive seen no crashes so far (leaving soon to a party so we shall see how it goes) GPU usage remains at 42% which seems a bit low for starcitizen at max settings Temp: 35C Core auto boosted to 2038 and the power usage is at 75/100%

*update 2* came back from the party and the game was still running. could it have been my PSU? if it is thats strange since it was running my old computer just fine with the 980ti,

*update 3* I find it strange that everything seems to be working. Could a faulty power supply cause memory errors and BSOD? seems unlikely but i guess im wrong,

mdzcpa
12-02-2017, 03:47 PM
I was thinking PSU or PSU cable issues from reading your first post. So, yes, looks like your PSU is not up to the task. A marginal PSU will cause all sorts of instabilities.

Khronos
12-02-2017, 08:02 PM
I was thinking PSU or PSU cable issues from reading your first post. So, yes, looks like your PSU is not up to the task. A marginal PSU will cause all sorts of instabilities.

Im lucky that my PSU comes with a 10 year Warranty and EVGA does a good job with that, so far it seems like i will have to ship it to them and from my initial findings UPS wanted to charge me like $80 to get it to California. ill go into USPS and look at their flat rate shipping boxes, the PSU weighs 6 lbs according to the website but im sure that includes all the cables and box itself.

P.S. ill keep everyone posted once i get the new PSU, although i might just upgrade to the new model the Supernova 850w G3 which is just smaller in size but everything else is the same (spec wise)

SherardG
12-04-2017, 06:59 PM
Hi Khronos,

I just noticed that your memory sticks (please clarify if I'm wrong!) 2 kits of 8x2 GB each? Basically, you bought 2 boxes of 2x8GB RAM kits?

I did some extensive research here, and even jumped over to the G.Skill boards and spoke to their reps there as well as via email and they confirmed that 2 different kits even of the same model # & chips do not play well with each other with XMP and you should use custom timings to run them at their rated speed; even then, only 40% of the time would it work. Could you try running them without XMP enabled just at the default clock rate before sending in your PSU for RMA?

Khronos
12-13-2017, 04:17 AM
I ended up finding out it was the RAM sticks that were the issue, once I removed one of the sticks it ended up being stable with the xmp profile. I'm waiting on the new RMA Ram (sent the 2 sticks back) so I can test again. I'm glad though that my system is in fact working just fine as of now. My new psu arrives on the 19th and I can go back to my nice modular psu and white bitfenix psu cables.

I had already done the RMA for the PSU and the RAM before I read your post SherardG but you are correct in saying it was 2 different kits of 8Gb sticks. Only reason I bought another kit was because it was on sale. Might just return the new RMA sticks and avoid any issues. I just wanted the 4 dimm slots to be filled for RGB purposes and not because I needed 32Gb of RAM to play games.

Ps I've been working a lot recently and haven't been able to update. 7days straight 12 hours each day tires people out.

no1yak
12-13-2017, 08:50 AM
Hi Khronos,

I just noticed that your memory sticks (please clarify if I'm wrong!) 2 kits of 8x2 GB each? Basically, you bought 2 boxes of 2x8GB RAM kits?

I did some extensive research here, and even jumped over to the G.Skill boards and spoke to their reps there as well as via email and they confirmed that 2 different kits even of the same model # & chips do not play well with each other with XMP and you should use custom timings to run them at their rated speed; even then, only 40% of the time would it work. Could you try running them without XMP enabled just at the default clock rate before sending in your PSU for RMA?

Not always the case, my 2 sets of 8x2 Gskill 3200 ran fine in my Hero IX and now they are in the Hero X. Both MB's run this memory under XMP with no problems at all.

Yes I know it's a gamble but, it can work(if your lucky). If I had a problem l then I would have split the memory to my Hero IX and sold the 2400 Corsair. Unfortunately It doesn't help the OP.

Khronos
12-13-2017, 09:26 AM
I've done this a few times with previous DDR3 memory in my previous computers and it's always worked XMP included. Now, I do know it's always different which is why, now that I know my computer works for sure I will test with the new kit and see of I have the same problems I had before. If I do I will return the RAM and and stuck with 16Gb which is 100% fine for playing games and some streaming.

SherardG
12-13-2017, 08:11 PM
Not always the case, my 2 sets of 8x2 Gskill 3200 ran fine in my Hero IX and now they are in the Hero X. Both MB's run this memory under XMP with no problems at all.

Yes I know it's a gamble but, it can work(if your lucky). If I had a problem l then I would have split the memory to my Hero IX and sold the 2400 Corsair. Unfortunately It doesn't help the OP.

I would agree with you, due to the fact that I'm running 2 kits of 8x2 G.Skill at 3200 as well. Although a lot of people from this board and even the G.Skill TechRep strongly advised against doing so;

Bottom line is, it doesn't matter if the chips are the same, if they were manufactured on the same date and are of the exact same batch, the only guarantee that they will work is if you buy kits that are actually tested and rated for the speed and timings you plan to run them on.

In my case, I had no choice since I ordered and paid for 2 kits of 8x2 GB without knowing about this issue, but one kit was delayed which allowed me time to do research, ask people, and perform my homework as to what I can do (as like OP, I used to do this back in the day, when the standard was just DDR) - where I come from, RMA is unheard of, and it's the same with cash refunds (you can only replace with items in stock of the same value; yeap, they do not protect consumers here) - what I have done was to force the shop that I pre-paid the incoming kits for to provide me with kit serial #'s of which I checked on G.Skill website until I found a pair that was manufactured on the same exact date as the sticks I was running - then I emailed the tech reps to verify the chips used on the card (they were Hynix, same as mine; I had to contact G.Skill directly because if the shop did the test themselves, they'd have to open up the DRAMs, slot them in, then check , basically making the item "good as sold" and I would be royally screwed over if they happened to be Micron or Samsung lol.



I ended up finding out it was the RAM sticks that were the issue, once I removed one of the sticks it ended up being stable with the xmp profile. I'm waiting on the new RMA Ram (sent the 2 sticks back) so I can test again. I'm glad though that my system is in fact working just fine as of now. My new psu arrives on the 19th and I can go back to my nice modular psu and white bitfenix psu cables.

I had already done the RMA for the PSU and the RAM before I read your post SherardG but you are correct in saying it was 2 different kits of 8Gb sticks. Only reason I bought another kit was because it was on sale. Might just return the new RMA sticks and avoid any issues. I just wanted the 4 dimm slots to be filled for RGB purposes and not because I needed 32Gb of RAM to play games.

Ps I've been working a lot recently and haven't been able to update. 7days straight 12 hours each day tires people out.

Glad to know you got your issue sorted out! And yes, I do clock 90 hours a week but I definitely spend Saturdays and Sundays for sleeping in and family time :D


I've done this a few times with previous DDR3 memory in my previous computers and it's always worked XMP included. Now, I do know it's always different which is why, now that I know my computer works for sure I will test with the new kit and see of I have the same problems I had before. If I do I will return the RAM and and stuck with 16Gb which is 100% fine for playing games and some streaming.

I would like to share to you the options that G.Skill provided me, as it might help you;
1. It's best if you get 8x4GB sticks rated and tested for the speed you would like them to run on
2. If this is not possible, you can try your new sticks on an XMP profile
3. Alternatively, you can try and ask your purchase point to provide you with sticks manufactured on the same date, or do as I did with the serials and confirm with G.Skill tech support before shipping; if they can provide you with at least a series of SN's for chips that they may ship you, you could use this page [https://www.gskill.com/en/check] to verify the manufacturing date, most likely they are made with the same DRAM IC's as they are done by batch.

Otherwise, you can pray that the new DRAM sticks they send you would run seamlessly with each other.

Best of luck! ;)

Khronos
12-14-2017, 03:32 AM
Moving to a New platform always has its issues Specially with DDR 4 and Ryzen. I just hope everything works because I don't want to drop another $400+ on more ram.