Help Needed

Triq · 14 Jun 2017 at 22:16

I recently posted a thread over a potential GPU issue which resulted in games to freeze and subsequently on rebooting the screen would be black, although the PC seemed to boot and even could hear the windows start up noise.

So assuming it was the GPU i have purchased the ASUS dual 1060 during that amazing deal. It arrived today and I am happy to say it is now running fine.

However, I tried one last time with the old 290x to see if I could get anything for resale value. To my surprise no signal through to display port, oddly though I changed to the HDMI slot and the screen almost magically turned on. I expect I'll sell it now for possible mining uses and recoup some of the new card cost.

When i tried to put the 1060 in I got a black screen on the boot up and was immediately worried that something larger was wrong. I restarted the PC and got a screen to show up straight away but I had the message that the boot up failed due to an overclock issue. That is something I had already set to defaults again when I was first investigating.

This has left me completely baffled and any advice would be greatly appreciated. Especially if it means I may need some other parts.

I did a memtest before for 4 passes and had no errors on that at all.

Specs

2500k at stock
16gb ram at 1333hz (also set back down)
Z68AP-D3 Gigabye MB
290x/now 1060
550W Superflower PSU

TWiNKLeT0Es · 14 Jun 2017 at 23:39

If your still getting the message every boot then go back into the bios and take pictures of as many settings as you can then just clear the CMOS.

Plec · 15 Jun 2017 at 00:34

EDIT: My below post isn't relevant re-reading your post but have left the reply in full just in case there's the remote chance that it's the same error displaying itself in a slightly different way. But given your earlier issues it appears that you still have an underlining issue you've yet to resolve.

Will re-read again in the morning and look over your other thread.

The above post is a good starting point. Take pics as suggested, sub menus too and scroll down. Once you're sure you've got pics of all your settings - clear the CMOS. Then boot back into the BIOS, load optimised defaults and enter in your original settings. Save and exit.

If you know your BIOS settings just clear and do what you need to.

How are your temps?

Triq said:
>snip<

The nVidia 9 and 10 series have a known quirks when using the DispayPort connector with 'some' computers.

I occasionally have the same issue with my GTX 1070 and it's an unusual fix - it may work for you.

Turn off your computer and then completely power off your monitor, unplug it from the wall and deplete all residual power from the panel by pressing the on switch. Leave it unplugged for 10 minutes. Also disconnect the monitor from the PC.

Plug DP lead back into PC, switch on power to monitor and power up PC - and hopefully it will spring back to life.

If it is this you can eliminate the problem altogether by using the HDMI cable.

Triq · 15 Jun 2017 at 09:19

I have had a check in the BIOS and found something that may have caused it. Everything was set to Auto other than the memory voltage which was sat at 1.56v rather than 1.5v. I don't know if that would have had any issues.

The problem isn't every boot either, it has been fine today on the few boots I have done.

Plec · 15 Jun 2017 at 09:57

Triq said:
I have had a check in the BIOS and found something that may have caused it. Everything was set to Auto other than the memory voltage which was sat at 1.56v rather than 1.5v. I don't know if that would have had any issues.

The memory voltage is fine - even if it is slightly overvolted it's very unlikely that is the cause - unless you have a fundamental problem with your memory anyway.

Although you may have 1600MHz stciks(?) and a lot of these sticks were originally 1.65V at XMP but were usually run as close to 1.5V to clock the 2500K.

I mention the above as it sounds like you removed the clock manually rather than clearing the CMOS? - as the memory would have reset to auto.

If this is the case i would really recommend clearing the CMOS. ANd even if you did i would try another reset.

Are you running the latest BIOS?

Triq · 15 Jun 2017 at 09:59

I'll clear that now and double check the BIOs while I am at it. The sticks are actually meant to be 2133MHz but have removed that and disable the XMP.

Triq · 15 Jun 2017 at 11:27

BIOS all up to date an no issues today. My main concern now is that it seems to have maybe been a bit of an issue that has sorted it self out and meant me buying a part I didn't need!

Plec · 15 Jun 2017 at 11:28

Doubl

Triq said:
BIOS all up to date an no issues today. My main concern now is that it seems to have maybe been a bit of an issue that has sorted it self out and meant me buying a part I didn't need!

Possibly, but too early to tell given the past problems you alluded to. You need to run it for a good few days to be absolutely sure you're stable. Use and abuse it over the coming week to be absolutely sure.

How long have you had the card? You could always DSR it.

Triq · 15 Jun 2017 at 11:50

Card arrived yesterday. I'll probably sell the 290x as it seems like it would get a good price to miners. I'm hoping it isn't CPU related though as that really is an expensive change.

Plec · 15 Jun 2017 at 12:08

Triq said:
Card arrived yesterday. I'll probably sell the 290x as it seems like it would get a good price to miners. I'm hoping it isn't CPU related though as that really is an expensive change.

I would test your system at stock/auto with the new BIOS for few days and see how it performs (I wouldn't worry about what ifs).

If you don't encounter any issues set up your memory and then test again for few more days.

Again, if that remains stable you could look at re-introducing your clock - but keep in mind your clock may have been degrading/becoming unstable all along, so I would attempt a lower clock but with similar voltages to start off with and work from there.

It's not that unusual for an aging clock to become unstable and require adapting to keep things stable. I had to drop my 2500K from 4.7/8 to 4.5GHz last year due to cold boot issues - didn't change any settings other than the multiplier. I then experimented with a slightly lower vCore but the cold boot issues reappeared - even at the lower frequency.