This post is more than 5 years old
8 Wizard
•
17K Posts
0
14941
Aurora-R6 : Hard-Lockup and crash while gaming (SOLVED)
On 3-9-2016 I had a hard-lockup while playing Fallout-4 (Steam version). I was unable to access Task-Manager. The (two week old) Aurora-R6 rebooted itself shortly there-after.
It was pretty scary since it first looked like the SSD had fried. On consecutive reboots (and even attempted cold boots) it just kept trying to boot from the network instead of the SSD. Turns out the BIOS decided to move the Windows Boot Manager to the lowest Boot Priority.
Event Viewer did not record anything important before or during the lockup.
Its been my experience that this usually points to a hardware problem (or running out of power).
I mistakenly thought the Aurora-R6 had been stable since installing it a few weeks ago. However, after checking the Event Viewer ... I found it had happened 4 other times before since 2-18-2017. Apparently, it's been happening while I'm not around (like when waking itself to install updates or do other maintenance).
These are the only "Critical" errors found.
Quantity:5 (in less than a month)
Level:Critial , Source:Kernel-Power (un-clean shut-down) Event-ID:41
I set Aurora-R6 to not reboot on crash
Set crash-logs to "Small Memory Dump" (mini dump) ... Just to see if it will write one
Set Virtual Memory to Custom on C: 4000/8000
My Aurora-R6 passed 2 scans with ePSA Diagnostics.
BIOS is currently v1.0.1 (what it shipped with)
Ran Alienware Update (I always run it manual)
- It seems a new BIOS was recently posted
- Said BIOS v1.0.3 was ready to install. Crossed fingers and let it.
- It rebooted to DOS and did it successfully
BIOS v1.0.3 NOW (3-11-2017)
Check BIOS settings ... still the same.
It's happened 2 times since the above time on 3-9-2017.
- Once just an hour into Fallout-4 and once while playing FarCry-3.
- I run this same Steam version of Fallout-4 on my Aurora-R1 (also with GTX-1070) for many hours and sessions with no problems what so ever. Same with FarCry-3 on UPlay.
So, BIOS 1.0.3 didn't seem to help with the lockups.
Hopefully it at least fixed the shifting Boot Priorities.
- Would be nice if we could get the issue details (or a least a proper summary) about Dell BIOSes when they are published.
Only difference is now, it doesn't reboot itself now (since I told it not to).
- Sometimes the screen is frozen and sound looping.
- Sometimes the screen eventually goes to sleep due to lack of video signal.
I still can't access Task Manager to get a clean shut-down. I have to hold-down power-button to force it off.
Still, no Event Viewer entries about it (other than the Critical Kernel Power entry that is created on reboot after the un-clean shut-down)
No BSOD
No crash logs or memory dumps created.
It's connected to a good APC 1500-LCD UPS.
(Click my name to see it's detailed config)
Edited after initial posting (left out some details)
| Dell Rockstar Registered Microsoft Partner and Apple Developer |
Carbon Based Lifeform
872 Posts
0
March 31st, 2017 12:00
the first Driver release for my shiny new Titan X Pascal crashed randomly. i was so mad at NVIDIA because there was no alt Driver available. now i pick one and if it works, i keep it and don't bother to update.
unfortunately, i've seen alot of hardware failures lately. my Asus Z170i PG board kept crushing. bad BIOS, faulty WLAN module and USB 3.1 causing BSOD. at some point i had enough and returned it. the mainboard i ordered before was DOA. bent CPU socket pin. then there was that broken EVGA Card. oh yeah. btw. what happened to the guy from few days ago? where is he?
your two GPUs Setup is fine. infact, you could use both GPUs if you wanted to. normaly you should be able to enable / disable the IGPU in the BIOS Settings. but I don't know how Dell handles it. there is a tiny drawback with the Intel GPU. it uses a part of the System Memory. something between 128 MB and 1 GB.
i hope the new driver fixed the issue. we'll see.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
March 31st, 2017 12:00
1. Same here
2. Same here. I have a good theory. I've written about it on my other blog. The second thing I'm seeing is components are getting increasingly delicate and more condensed, while it seems the manufacturers and technicians are getting rougher handling the components (especially their cables, connectors, and sockets) .
3. Right, I think so too. I just meant it might facilitate "bad video drivers" to create the serious Hardware Failure Hard-Lockup symptom in this machine, while not in others with a single-GPU.
4. Me too .
Carbon Based Lifeform
872 Posts
0
April 10th, 2017 16:00
was about to bump this thread. i wanted to know how it goes.
so it was the GPU Driver after all?
i'm glad you sorted it out.
Edit:
i'll just drop it here since i can't PM you anymore.
Are there any Samsung PM961 Drivers? - [Solved] - Storage
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
April 11th, 2017 12:00
Thanks.
However, for me ... it's like we were saying with video drivers ... once we find a working one. I tend to keep the dependable, reliable, working one for a while. Especially with something like a disk driver. If my OS or dev projects code started getting corrupted due to buggy r/w, that would be especially bad for me.
IMO, anything r/w over 3000/1600 is good enough for me. That's like 5-6 times faster than the fastest SATA-3/600 SSD anyway, right? I don't think 10% faster makes a difference.
However, if there is an official Samsung pm961 firmware update ... yeah, I want that. However, if there is, I'm wondering why Dell is not validating and posting it for everyone. I'm sure the Dell business customers want it too.
=================================
Thanks to Dell, I am now running my 16gb of ram memory in Dual-Channel Mode finally.
=================================
I saw that new Aurora-R6 BIOS v1.0.4 says:
Enhancements:
- Enhanced system boot time.
... but haven't been brave enough to try it yet. I'm just enjoying having a 100% stable system for the last 30 days.
I sent you a PM.
Carbon Based Lifeform
872 Posts
0
April 11th, 2017 13:00
1. ah okay then. i didn't know you need the machine for work. in that case i wouldn't mess with it either. my AW PC is only for gaming. thats why i don't mind to test new stuff. last weekend, as you know, i installed new FW and Drivers for my M2 SSD. but it was necessary.
well maybe you'll need the Drivers one day.
2. i can't understand why people complain about Dell Support. i called them last week and ordered some parts. usually they don't sell them. atleast not in Germany, but they made an exception for me. the key is beeing nice i guess.
3. yes. i noticed.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
April 11th, 2017 20:00
If PCIe/NVMe, what are you getting in Crystal Disk-Mark now?
Carbon Based Lifeform
872 Posts
0
April 11th, 2017 22:00
now that you mention it.... i should check when i get home.
Edit:
Done. nothing changed. everything almost the same.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
May 11th, 2017 10:00
Now this is more like it.
During this whole period, Aurora-R6 was subjected to heavy usage. This includes long multi-hour Fallout-4 and FarCry-3 gaming sessions, general productivity applications, and even some stress testing.
and OCCT Power-Supply Test (whole machine)
Alienware Aurora (R6)
Intel i7-7700k 4.2ghz
Liquid Cooling - Intel z270 Chipset - VRM HeatSink
16gb DDR4 Hyper-X Fury 2667mhz Ram (Dual-Channel)
850w Power Supply (now N1WJD Delta-model)
Nvidia GTX-1070 OEM (8gb)
Samsung PM-961 512gb SSD (M.2 PCIe NVMe)
Toshiba 1tb 7200rpm HDD & DVD Burner
Killer Networks e2400 Gigabit Ethernet RJ-45
Intel 3165 802.11ac Wi-Fi & Bluetooth v4.2
-
Dell 27" u2717D UltraSharp Infinity-Edge IPS 16:9 QHD Monitor
Logitech G110 Keyboard and G500 Mouse
Plantronics GameCom 780 USB Headset
Dell AE415 2.1 Speakers
APC Back-UPS XS-1500 (BX1500g)
-
Windows-10 Pro (64-bit)
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
May 17th, 2017 14:00
I installed Creators Update the other day.
It worked fine. Creators version now correctly shows as v.1703 (instead of v.1607 which was last year's Anniversary Update).
I did have a little problem with Thermal Controller, so I would suggest you exit it (the System-Tray icon) before allowing the upgrade to start. The Thermal Controller will auto-start again after a reboot (and work fine).
According to the Reliability Monitor report, I noticed all drivers (and I think Modern/Universal Apps) seemed to get inventoried, and re-installed (migrated) during upgrade.
The Aurora-R6 is still working fine in use ... even Alienware Command Center v4.6.20.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
May 17th, 2017 17:00
The fan bearing went-out on my Huntkey-made 850w Power-Supply.
Dell sent me a Delta-made 850w Power-Supply the other day to replace it. The Aurora-R6 is back to 100% working order.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
July 1st, 2017 19:00
Installed BIOS v1.0.4 and after it finished, tested computer and Windows.
Downloaded each in Windows. Ran from Windows but computer reboots and installs from UEFI BIOS (which is good and correct).
A few more strange reboots later and then it's installed completely.
Then I did BIOS v1.0.5 the same way.
v1.0.5 seems to mess with a few minor BIOS settings a bit. I switched them back. Probably a good idea to take pics or otherwise record BIOS settings first.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
November 9th, 2017 12:00
The other day, my wife was using Edge browser on the Aurora-R6 and asked "is the Internet working for you"? Mine was fine but her's was real slow. Reboot didn't help. I checked ethernet cables, desktop-ethernet-switch, router, etc. Packet-loss is a good description. She started playing World-of-Warcraft again online and it strangely appeared fine.
mattww3 mentioned a similar problem with his new Aurora-R7:
The killer drivers had to be installed directly from Killer website. Both Dell and MS had older drivers at the time. Had at least 60%+ WiFI packet loss until BIOS/Chipset + killer update.
So, I updated the on-board "Killer Networks e2400 Gigabit Ethernet RJ-45" wired-ethernet driver from Device Manager. Driver changed from older Microsoft to much newer Rivet. Windows update is on common Consumer branch, so it's lame we have to do this. Seems to have fixed it. Generally, I hate Rivet/Killer but I had no choice with this purchase (since it's build-into the motherboard). I also stay away from their "heavy and invasive" driver suite ... I just run the core driver.
This is the first problem I've had with it. I think it might have started with recent Windows-10 update. If it crops-up again, I might just so with original plan which was to disable it and install a working PCIe gigabit ethernet card.