BSOD/freeze with Ryzen 1700x, GTX 460 1GB, 600w PSU

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
Furmark freezes computer at 70 degrees C. Also experiencing freezes while gaming at random places. Issues only occur on heavy GPU load. In idle or light GPU load, no issues. "Event ID 14 from source nvlddmkm cannot be found" found in Systems logs. I suspect a new PSU is needed. A 6-year-old Cooler master silent pro 600w is in use (5 year warranty).
http://www.coolermaster.com/service/support/model/RS-600-AMBA-D3/

Things I've tried:
- Setting power management profile to Ryzen balanced or performance.
- Setting power management on the PCI-E to "Off"
- Update chipset drivers
- Update bios to F7
- Update graphic drivers nvidia
- Reseating all components aside from CPU cooler and CPU.
- Double checking power cables and cleaning out dust.
- Memcheck86 (4 passes for ~5 hours) and windows 10 built-in Memsched. No issues.

My old rig runs furmark just fine and does not freeze with the same GPU. It should consume ~354W.
https://pcpartpicker.com/list/q4f7xY

My new build consumes ~379W and it seems like if I unplug anything but bare essentials it sometimes passes Furmark beyond 70 degrees and stabilises at 76 degrees C.
https://pcpartpicker.com/list/72JxzM
 

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
So the PSU arrived yesterday. No difference, still freezes. Also tried with downclocked the factory overclock of the gigabyte geforce gtx 460. No dice.

So I guess I'll have to take the cooler of the GPU then.

I also noticed a BSOD with a "IRQL NOT LESS OR EQUAL" should have a minidump somewhere. After reading some threads I'm starting to think that nvidia don't have stable drivers for Windows 10 for GTX 4xx. It runs fine on my Windows 7.

https://forums.geforce.com/default/topic/862108/geforce-500-400-series/windows-10-has-killed-my-gtx-460-/1/

TLDR; Quite an impressive factory overclock at 1800 Mhz mem (1350) and 715 (over stock 675) Mhz.
 

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
At this point, thermal paste and even more rigorous dust cleaning. I don't have the equipment to test capacitors. But I do have a bag capacitors. But I'd rather spend money on a new GPU if that's what's wrong before doing soldering work on this chip. Thanks for your input.
 

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
Holey moley!! [strike]It sure looks like thermal grease did the trick! I did not expect that.[/strike] It seems there was an improvement of about 2-3 degrees C using Furmark (100% load) and roughly 5 degrees during gameplay (99% load). No stuttering and, so far[strike], not failing a stress test (5 in total)[/strike]. After some more testing it eventually failed again, however this time at a lower temperature.

I also did the following changes:
- Installed MSI Afterburner and running at GTX 460 stock rather than factory overclock. GV-N460OC-1GI.
- Re-installed nvidia drivers ticking the "clean"-option and not selecting the audio driver
- Update BIOS to F8
- Installed the Realtek audio driver (which was not installed before). It's suggested to conflict with nvidia drivers in some cases.
 

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
I'd say this could be something. I've been checking out the VGA BIOS update for the card and noticed there has been updates to the card in terms of memory support. From what I can interpret there are either related to DDR3 support for Samsung modules or DDR3 support for Hynix depending on what make/revision of the card you have. However none of these updates relate to my particular card. Which seems to be a GV-N460OC-1GI-rev-20 (by Nvidia referred to GTX460 1GB SE Rev 2) with BIOS version F11. And from that particular BIOS-version there is no upgrade path.

"Please note:
You can only update to a VBIOS version of the same series.
If your VBIOS version is:
F1, it can only be updated with VBIOS versions F2-F9.
F10, it can only be updated with VBIOS versions F11-F19.
F20, it can only be updated with VBIOS versions F21-F29."

From using Gigabytes own software I got the following:
https://www.gigabyte.com/Graphics-Card/GV-N460OC-1GI-rev-20#support-dl
However when I look up the card it differs in looks and is more similar to the following:
https://www.gigabyte.com/Graphics-Card/GV-N460OC-1GI#support-dl

The listed BIOS update is for F70/F80/F90 or F3.


From what I've read somewhere, but I can't remember where, the card has 1GB on-board video memory but can also allocate 2 additional GB from RAM somehow. This is all fairly confusing to me. I might end up buying a new card, but will make contact with Gigabyte first.
 

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
So sent Gigabyte a question in the topic of DDR4 support. I also did some digging around in the eventlogs and I find the following.


Level,Date and Time,Source,Event ID,Task Category
Error,2017-09-08 11:30:57,nvlddmkm,14,None,"The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video1
CMDre 00000004 00000ffc ffffffff 00000007 000000ff

the message resource is present but the message is not found in the string/message table
"


After some digging around I summarize the addresses mentioned after CMDRe they seem to always occur in the following order. Regardless of when and how the freeze occur. I don't know what CMDRe means but the supposed addresses appear in the following order:


CMDre 00000000 00000ffc ffffffff 00000007 000000ff
CMDre 00000001 00000ffc ffffffff 00000007 000000ff
CMDre 00000002 00000ffc ffffffff 00000007 000000ff
CMDre 00000003 00000ffc ffffffff 00000007 000000ff
CMDre 00000004 00000ffc ffffffff 00000007 000000ff
CMDre 00000005 00000ffc ffffffff 00000007 000000ff
CMDre 00000006 00000ffc ffffffff 00000007 000000ff
CMDre 00000007 00000ffc ffffffff 00000007 000000ff
CMDre 00000008 00000ffc ffffffff 00000007 000000ff
CMDre 00000000 00000ffc ffffffff 00000007 000000ff
CMDre 00000001 00000ffc ffffffff 00000007 000000ff
CMDre 00000002 00000ffc ffffffff 00000007 000000ff
CMDre 00000003 00000ffc ffffffff 00000007 000000ff
CMDre 00000004 00000ffc ffffffff 00000007 000000ff
CMDre 00000005 00000ffc ffffffff 00000007 000000ff
CMDre 00000006 00000ffc ffffffff 00000007 000000ff
CMDre 00000007 00000ffc ffffffff 00000007 000000ff
CMDre 00000008 00000ffc ffffffff 00000007 000000ff
CMDre 00000000 00000ffc ffffffff 00000007 000000ff
CMDre 00000001 00000ffc ffffffff 00000007 000000ff
CMDre 00000002 00000ffc ffffffff 00000007 000000ff
CMDre 00000003 00000ffc ffffffff 00000007 000000ff
CMDre 00000004 00000ffc ffffffff 00000007 000000ff
CMDre 00000005 00000ffc ffffffff 00000007 000000ff
CMDre 00000006 00000ffc ffffffff 00000007 000000ff
CMDre 00000007 00000ffc ffffffff 00000007 000000ff
CMDre 00000008 00000ffc ffffffff 00000007 000000ff
CMDre 00000000 00000ffc ffffffff 00000007 000000ff
CMDre 00000001 00000ffc ffffffff 00000007 000000ff
CMDre 00000002 00000ffc ffffffff 00000007 000000ff
CMDre 00000003 00000ffc ffffffff 00000007 000000ff
CMDre 00000004 00000ffc ffffffff 00000007 000000ff


During the second run. Starts at the very same "00000000 00000ffc ffffffff 00000007 000000ff"


CMDre 00000000 00000ffc ffffffff 00000007 000000ff
CMDre 00000001 00000ffc ffffffff 00000007 000000ff
CMDre 00000002 00000ffc ffffffff 00000007 000000ff
CMDre 00000003 00000ffc ffffffff 00000007 000000ff
CMDre 00000004 00000ffc ffffffff 00000007 000000ff
CMDre 00000005 00000ffc ffffffff 00000007 000000ff
CMDre 00000006 00000ffc ffffffff 00000007 000000ff
CMDre 00000007 00000ffc ffffffff 00000007 000000ff
CMDre 00000008 00000ffc ffffffff 00000007 000000ff
CMDre 00000000 00000ffc ffffffff 00000007 000000ff
CMDre 00000001 00000ffc ffffffff 00000007 000000ff
CMDre 00000002 00000ffc ffffffff 00000007 000000ff
CMDre 00000003 00000ffc ffffffff 00000007 000000ff
CMDre 00000004 00000ffc ffffffff 00000007 000000ff
CMDre 00000005 00000ffc ffffffff 00000007 000000ff
CMDre 00000006 00000ffc ffffffff 00000007 000000ff
CMDre 00000007 00000ffc ffffffff 00000007 000000ff
CMDre 00000008 00000ffc ffffffff 00000007 000000ff
CMDre 00000000 00000ffc ffffffff 00000007 000000ff
CMDre 00000001 00000ffc ffffffff 00000007 000000ff
CMDre 00000002 00000ffc ffffffff 00000007 000000ff


 

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
Additional things I've tried, going through suggestions elsewhere:
+ Replaced SATA drives
+ Disabled Cool&Quite in BIOS
+ Disabled Realtek HD Audio in BIOS
+ Trying to disable motherboard surge protection. No such BIOS option for Gigabyte Gaming 3 AB350.
+ Disable Windows Update automatic installation of drivers and applications
+ Checked thermals situation again. All well below 60 except for VRM mosfets.
+ Switched electric power strip
+ F7 and F8 BIOS
+ Check all sticks and used slots individually with Memcheck86 (rather than in dual channel mode). All ok.
+ Checked 800 MB of VRAM of GTX 460 with 10 passes with memtestg80. All ok.

No difference. These are planned changes.
+ New GPU
+ BIOS F9b
+ Change PCI-e slot

According to support things point towards GPU or motherboard. It could also be a driver issue or the DDR4 issue mentioned above. Old hardware is probably not tested against new CPU and motherboard very often, especially custom factory overclocks. Some people have been reporting issues with surge protection on similar motherboards.
 

jmrnilsson

Honorable
Mar 20, 2014
57
1
10,665
The new GPU arrived yesterday. Not to sure about anything yet, but it sure looks like the geforce 400-series Fermi drivers were a bit flaky on AM4. Especially under heavy load. Aside from that I also updated the BIOS to F9d. But it looks like GTX 460 simply was too old and unsupported.