Question Windows failing to boot on restart, only booting when forced through BIOS

[nickmnader]

Hi all,

I am upgrading/rebuilding my PC with a new MOBO, CPU, GPU, and RAM.

New Part list:
Mobo: ASUS Z790 Gaming Wifi7
CPU: i9-14900k
GPU: ASUS GeForce 5070 OC Tuf gaming
RAM: Corsair DDR5 6000mhz 2x16gb

Old parts:
NZXT Kraken 360
M.2 SSD
SSD
HDD
Etc

Behavior:
I did a fresh windows install on the M.2 as my boot drive.

I've gotten it to boot through going into the BIOS boot menu and selecting the M.2.

One booted I was about to install all my programs, run tests, update windows etc.

Main problems:
1. I incurred issues with the CPU temps especially when stress testing that concerned me, spiking to 100c. The CPU default clock is set to 5.9ghz at 1.34v. I've tested it at both the BIOS default settings and Intel Tuning optimized settings. I've tested it as low as 5.5ghz with same results. My radiator fans are running full speed and seem fine. Although the BIOS doesn't recognize them since they're going through the NZXT fan hub and into a USB header on the mobo instead of the CPU FAN header. It would also seem that the fans aren't kicking into higher speeds without NZXT Cam running, which could be problematic in the future.

2. When I restart windows for updates or just through the software restart, it goes black and display never turns back on but hardware continues to run.

3. It boots into BIOS every time cause of forced shutdown on previous boot.

4. It has occasionally changed behavior and booted straight into windows as I try things like removing 1 of the RAM sticks, or my other 2 drives but is never consistent (e.g. removing drives and 1 RAM doesn't solve boot issues.

I've tried repeating and reapplying thermal paste 3 times. I've reset BIOS to default settings. I've bread boxed it with minimal components.

The next step is another fresh windows 10 install from boot media.

Edit: I also noticed that the CPU socket has a few pins towards the top that stand out, is this normal or a defect that could be causing the issues?

Errors from Event Viewer during restart attempts:
"The server could not bind to the transport \Device\NetBT_Tcpip_{C6D72FB2-0F6F-4544-A361-E80C30781EE3} because another computer on the network has the same name. The server could not start."

"The IPv4 TCP/IP interface with index 17 failed to bind to its provider."

Any suggestions?

 

[Lutfij]

You could try and flash the BIOS to the latest version but this, is a point of concern;
Edit: I also noticed that the CPU socket has a few pins towards the top that stand out, is this normal or a defect that could be causing the issues?
If the board came like that(out of the box) then please RMA the board and don't attempt to fix the bent pins.

 

[Darkbreeze]

I'd bet good money those pins are the cause. Why? No idea, but ANY bent pin can cause widely varied and unpredictable behaviors, or unexpected but completely repeatable behaviors. And you have several bent pins there. Whether you did it or it came that way is anybody's guess, but what I can tell you is that the pin bed on 12th-14th Gen is FAR more delicate than on past generations as the socket size has remained the same while about 500 additional pins have been added in the same amount of space so they are much smaller and far more easily bent. You need to be VERY careful during removal of the socket protector and even more so until (Which you should do immediately after removing the socket protector) the CPU is installed. I'd say, at a guess, it's like 50% more likely than with previous LGA sockets that you might inadvertently damage the pin bed on LGA 1700 than it was on LGA 1200 or previous Intel sockets.

However, very recently, within the last month in fact, I have received two motherboards, one from Amazon and one from Newegg, that were both supposed to be new and both of them had bent pins out of the box. I would return that board for replacement or refund before doing anything else.

 

[johnbl]

Hi all,

I am upgrading/rebuilding my PC with a new MOBO, CPU, GPU, and RAM.

New Part list:
Mobo: ASUS Z790 Gaming Wifi7
CPU: i9-14900k
GPU: ASUS GeForce 5070 OC Tuf gaming
RAM: Corsair DDR5 6000mhz 2x16gb

Old parts:
NZXT Kraken 360
M.2 SSD
SSD
HDD
Etc

Behavior:
I did a fresh windows install on the M.2 as my boot drive.

I've gotten it to boot through going into the BIOS boot menu and selecting the M.2.

One booted I was about to install all my programs, run tests, update windows etc.

Main problems:
1. I incurred issues with the CPU temps especially when stress testing that concerned me, spiking to 100c. The CPU default clock is set to 5.9ghz at 1.34v. I've tested it at both the BIOS default settings and Intel Tuning optimized settings. I've tested it as low as 5.5ghz with same results. My radiator fans are running full speed and seem fine. Although the BIOS doesn't recognize them since they're going through the NZXT fan hub and into a USB header on the mobo instead of the CPU FAN header. It would also seem that the fans aren't kicking into higher speeds without NZXT Cam running, which could be problematic in the future.

2. When I restart windows for updates or just through the software restart, it goes black and display never turns back on but hardware continues to run.

3. It boots into BIOS every time cause of forced shutdown on previous boot.

4. It has occasionally changed behavior and booted straight into windows as I try things like removing 1 of the RAM sticks, or my other 2 drives but is never consistent (e.g. removing drives and 1 RAM doesn't solve boot issues.

I've tried repeating and reapplying thermal paste 3 times. I've reset BIOS to default settings. I've bread boxed it with minimal components.

The next step is another fresh windows 10 install from boot media.

Edit: I also noticed that the CPU socket has a few pins towards the top that stand out, is this normal or a defect that could be causing the issues?

Errors from Event Viewer during restart attempts:
"The server could not bind to the transport \Device\NetBT_Tcpip_{C6D72FB2-0F6F-4544-A361-E80C30781EE3} because another computer on the network has the same name. The server could not start."

"The IPv4 TCP/IP interface with index 17 failed to bind to its provider."

Any suggestions?

NetBT_Tcpip i think is netbios over tcpip
before registering a name it will ask the network it any machine is using that name. I have seen old cases where a machine that has more than one network connections will register the name on one adapter then see the announcement from the second adaptor on the same machine then get this type of failure. I think there is a windows method to prevent this but it can be bypassed and who know what would happen during a plug and play setup of a second device. It used to be that the slowest device would not be used. you could work around it in various ways but just for testing you would use the wireless device and unplug any hardwired network cable and see if the problem goes away.

for the problem during a restart, windows should not complete the restart if it thinks it is going to loose some data. this happens if windows tries to flush a buffer to disk but does not get a response. for this case you would google "how to make windows use verbose logon or logoff messages."
then make the registry or app changes and the next time windows should tell you what it is waiting for.

I have seen this problem with some 3rd party network attached storage, bugs in ssd firmware and some firmware in older drives.
(cases that require you to turn off drive lazy writes to get proper shuddowns)

in these cases the screen goes black and it looks like it shut down and did not restart. You hit the power button and it forces a dirty shutdown, the system reboots and runs chkdsk on the drive and puts a warning in the event log. in reality, windows was waiting for a signal before completing the shutdown and the user forced a dirty shutdown.