BSODs during Windows 10 installation on a Ryzen machine

dmtlevin

Prominent
Oct 1, 2017
4
0
510
I have problems installing Windows 10 (and almost any other operating system) on my new freshly built Ryzen PC. The specs of the machine are:

  • CPU: AMD Ryzen 7 1700X
    MB: Asus Prime B350M-A (BIOS ver. 0806)
    RAM: Kingston HyperX Fury DDR4 2666MHz 2x8GB (HX426C15FBK2/16)
    GPU: Nvidia GeForce GTX 970
    PSU: Corsair RM550x
    SSD: Samsung 850 EVO 500GB
    HDD: Western Digital Red 2TB
    TP-Link N900 Wireless PCI-E card.
I tried to install latest Windows 10 (1703 10.0.15063) first, but got various BSoDs during installation with following errors:

  • IRQL_NOT_LESS_OR_EQUAL (x1)
    DRIVER_IRQL_NOT_LESS_OR_EQUAL (x2)
    SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (x6)
    KMODE_EXCEPTION_NOT_HANDLED (x2)
    BAD_POOL_HEADER (x1)
    KERNEL_SEQURITY_CHECK_FAILURE (x2)
    CRITICAL_PROCESS_DIED (x1)
    UNXEPECTED_KERNEL_MODE_TRAP (x1)
    0xC000021A (x1)
BSoDs occur at random moments during installation, but always before the actual file copying.

I’ve read similar threads where people suspect hardware problems, so here are my attempts to fix or diagnose the issue:
1. Clean the contacts. Checked both MB slots and RAM pins - no dust or oxidation.
2. Remove one RAM stick. Installation succeeds, system is stable under various loads (two days of gaming, browsing, heavy math computations, benchmarks, stability tests etc.)
3. Try other slots. I tried ALL 12 possible memory placements for two sticks, and it happened, that no matter where I put the memory modules, two 8GB sticks don’t work, while one 8GB stick does work.
4. Try reducing memory frequency. Set the frequency to 2400 and 2133MHz with no effect.
5. Check the memory. Checked each stick separately, then both together with Memtest86 - over 9 passes and no errors.
6. Remove all unnecessary hardware. Assembled a minimal setup with just MB, CPU and graphics card - but nothing changed.
7. Reset CMOS battery. Removed battery for 10 sec, 1 min, 10 min at various stages of my shenanigans - no effect.
8. Install Windows 7. At the end of installation it just suddenly rebooted and corrupted the disk partition, so I should repartition the hard drive using another PC.
9. Install Linux. Linux Mint 18.3 Live USB displayed “Unknown hardware!” error in the terminal and frozen.
10. Try another memory Installed 2x8GB kit G.Skill Ripjaws 4 Blue DDR4 3000MHz (F4-3000C15D-16GRBB) with no effect.
11. Try single 16GB stick. Installed TEAM Elite DDR4 2400MHz (TED416G2400C1601), which is also in both Asus and TEAM QVL lists - no effect.
12. Try another motherboard. Took MSI X370 Gaming Plus, but the problems were still the same, so I returned it back to the store.
13. Install Windows 8.1. Oh, magic! It successfully installed, addressed all 16GB of RAM, utilized all 16 CPU cores and stably worked for two days under various loads, benchmarks, tests etc. AIDA64 report also didn’t show any strange things at the first glance. I was ready to leave it, but unfortunately it does not support other software I need.
14. Install newer Linux. Installed Ubuntu 17.04 with kernel 4.10, but it also failed with error “ACPI Error: Needed [Integer/String/Buffer], found [Region] 0xFFFF...”. However after booting with “acpi=off” parameter, it started and successfully installed the OS, but detected only one CPU core and had problems with Wi-Fi connection (but it’s a usual thing, so I guess, it’s solely Ubuntu’s problems)
15. Bring the PC to the service. Guys from the local store, where all this crapware was bought, briefly checked the PC and said that each component is OK, and this is incompatibility problem, not a warranty case. Guys from another third-party service checked the PC and also said that each component individually is OK, but they also didn’t managed to install Windows 10.
16. RMA the CPU, mobo etc. The store refuses to RMA my hardware, “because the stickers on the boxes were cut”, and it’s not in marketable condition anymore.

Additional information:

  • CPU temperature seems to be fine, as it never raised above 56 degrees under 100% load
    Chipset temperature is also fine at ~30 degrees all the time
    PSU outputs correct voltages, as motherboard and guys from the service told me.
Please help me, I'm not able to start working for three weeks now and I'm quite desperate and thinking to build a new Intel-based machine, which is a ton of money.
 
Solution
Windows doesn't lack drivers, it has generic drivers for various things. Missing drivers are downloaded through Windows Update like chipset drivers, display drivers and audio drivers.

dmtlevin

Prominent
Oct 1, 2017
4
0
510

The store provides 36 months warranty for CPU and mobo, but they tested my machine and said that each component is fine, so no warranty service. They also agreed that they couldn't install win10 on the machine too.

 
They quite often test using software that cannot detect a lot of hardware failures, or using the eye which also cannot see many problems.

Unless it is explicitely written on their website that they won't RMA if a sticker is broken, they must RMA it.
If you happen to find it explicitely written somewhere on their site, you might be able to contact AMD directly for an RMA. When I had problems with my system and suspected the motherboard to be the cause, the store I bought it from told me I should contact Gigabyte for RMA.
 

dmtlevin

Prominent
Oct 1, 2017
4
0
510
Also I was thinking, that if all the BSoDs are related to drivers problems, maybe windows image lacks some chipset drivers. Have just tried to take the latest ones from Asus website and inject them into windows 10 image using DISM and then boot with that. Didn't know what drivers to put - there are around 20 of them, - so I added all. Still no luck though - the CRITICAL_PROCESS_DIED BSoD just after the language selection screen.
 

TRENDING THREADS