3200 Mhz RAM, causing freezing at stock.

[JasonOrme]

Hi,
I hope you guys can help me out, you have been a godsend in the past.


I purchased:
Corsair CMK16GX4M2B3200C16 Vengeance LPX 16 GB (2 x 8 GB) DDR4 3200 MHz
Back in May along with:
ASUS ROG STRIX B450-F GAMING AM4/B450/DDR4/S-ATA 600/ATX Socket Motherboard
A new CPU
AMD Ryzen 7 2700X Processor with Wraith Prism RGB LED Cooler
And a new graphics card:
RADEON RX VEGA 64 NITRO+ 8GB HBM2 PCI-EXPRESS GRAPHICS CARD

Since fitting everything I have been getting random lockups/freezing, oddly enough only a handful while gaming and at least once a day while idle. Most nights I go to bed I will wake up to my computer with a black screen and not responding.
Sometimes if I have been gaming (Division 2, Guild Wars 2) for hours, when I'm finished and the game is closed, it will happen 10 minutes, or an hour or so later.

Googling the issue it all seemed to boil down to my graphics card and I have been tinkering for over a month with AMD's Adrenaline, underclocking, overclocking, installing new drivers, old drivers, fresh installations of windows and other things based on advice and nothing appeared to fix it.

I thought I could just, put up with it, since it very very rarely happened while gaming. But it's starting to grate on me now, as I walk away from my computer for 10 minutes after a gaming session and I'm forced to turn the power off on the back of the computer.

Now with all the fancy lights still going on inside the case after the computer freezes I neglected to notice the orange coloured light on the motherboard indicated a RAM issue.

I re-seated the RAM and disabled the 3200 speed and returned it to stock settings to see if this would fix the problem, but sadly it didn't.

Have you guys any advice? Surely it must be a RAM issue if the motherboards stating so?.

 

[JasonOrme]

I tried them using the DOCP settings, it still crashes.
Someone suggested upping the ram voltage slightly.. by default it's 1.35v, so I have upped it to 1.38v.
The PC has been up for a day without crashing so far.

 

[JasonOrme]

I spoke too soon. It just crashed.

I've dropped the ram speed down from 3200 to 3000.
So will see how this goes.

 

[JasonOrme]

Thanks.

Sticks are definitely in the correct slots according to the manual.

I'm trying those settings now.

The LLC levels for the CPU & SOC gave me "low, medium, high and extreme" options, so I went with high.

 

[JasonOrme]

Same thing has happened, was playing a game for 4 hours or so, closed the game, went downstairs and returned 10 minutes later to black screen and orange blinking RAM light on motherboard.

 

[JasonOrme]

PSU is a Corsair GS800 (800w).

USB devices:
Keyboard
Keyboard EXT
Mouse
USB HUB
Headset
Graphics Tablet

Checked power settings and they are already set to what you suggested.

Thanks for the help, but yeah.. this one really has me confused.
I'm reading a lot of people not being able to hit 3200mhz without crashes and having to dial it back to 2933mhz.

 

[JasonOrme]

Set it to 2933mhz still occuring. Absolutely baffled.

 

[JasonOrme]

Unsure.

Well, I'm 2 days in and it hasn't froze yet.
Lowered the ram speed to 2933. I increased the RAM voltage to 1.39
I changed load-line calibration for both CPU and SOC from high to extreme.
And SOC voltage' a positive offset of +0.175v.

But I also changed a few things on Radeon Adrenaline.
I changed the minimal GPU state to 3.
The min/max GPU Memory state to 3.
Also changed the cards power limit to 50%.

For 2 days I have been playing the Division 2 on and off, running stress tests etc and so far so good.
CPU and GPU temps are good (I think). Under heavy load the CPU maxed at 73c and GPU at 80c.

Obviously I have changed quite a lot of things, so will give it some time to be sure it stable and then try changing some things to isolate the problem.

 

[JasonOrme]

So I managed to go 6 days without any issues, but while pottering about the bedroom and the PC being idle my screens went black, keyboard / mouse lights went off and computer was un-responsive. Fans were still going and LED's were working.
But this time no orange/yellow RAM light on the motherboard.
So maybe it's a mix of GPU & RAM issues, and the RAM issue is now possibly fixed?.

 

[JasonOrme]

Ran the built in memory diagnostic, no errors found.

 

[JasonOrme]

RAM has been running at 2933 for some time now. I updated my bios two weeks ago but as you said there is already a new update, so I have updated it and will see how it goes.

 

[JasonOrme]

New bios update was a big mistake as you can see in this thread.
https://rog.asus.com/forum/showthre...2&highlight=Search...ASUS+STRIX+B450-F+GAMING

Lots of issues, can't use mouse in bios and so I can't edit CPU fan curve, CPU fan is driving me mad revving up and down every 2 seconds.

I did however do a fresh windows install on a new SSD yesterday, no freezing yet, but too early to tell.

 

[JasonOrme]

I'm very much regretting it now, hoping ASUS will hurry up and roll out an update so I can fix the CPU fan curve.

 

[JasonOrme]

They released a bios update that fixed that small issue.
My CPU and GPU are currently overclocked and my RAM is running at full clock speed.
I have yet to have a crash and the only real change was I did a fresh Windows installation on a new SSD.
Maybe with a new motherboard, gpu, etc I should have done a fresh installation all along, or maybe my old SSD has some issues?.

Other changes (incase anyone stumbles upon this thread with the same issue).
I increased the RAM voltage to 1.39
I changed load-line calibration for both CPU and SOC to extreme.
And SOC voltage' a positive offset of +0.175v.

I don't know if any of those things have anyting to do with it, but I have kept those settings on the bios even after the fresh windows installation.

I have run 3DMark ultrafirestorm stress test and passed with 99%, Heaven benchmark has been run for hours and is fine. Also been running The Division 2 on ultra, 4k for a few hours a night and no issues.

Fingers crossed this is the end of the issue, thanks for the help guys.

 

[JasonOrme]

Sorry to revive my thread nearly a year later.
My last post was indeed the end of my troubles... so much so that I decided to purchase another 16gb of RAM and now I'm experiencing the problem again.

I can get 3 sticks running at full 3200, any slot, any stick.
But as soon as 4 sticks are in my system ends up in a booting loop. This can only be fixed by dropping the frequency down to 2933.

I had already increased the RAM voltage to 1.39
I changed load-line calibration for both CPU and SOC to extreme.
And SOC voltage' a positive offset of +0.175v.
These settings fixed my problem before.

Do these settings need changing now that I have more ram?.
Any help will be most appreciated.

 

[JasonOrme]

If you do this, remove the offset from the SoC voltage. Stepping down to 2933, you should be fine on the Auto settings. Also mem voltage will be fine at 1.35v.
It also slows the Infinity Fabric from 1600 to 1467MHz, but I've done a lot of benching, and the difference in real terms is negligible.

Rather that than a PC which crashes all the time...

This currently seems to be the best option. At first I thought my new ram was faulty but turns out memtest would freeze randomly with the frequency at 3200. At 2933 it passed after 9 hours with no errors.

make sure Gear Down Mode/GDM is set to On/Enabled. Have you done anything with the primary and secondary timings?

I've never enabled gear down mode and it appears to be on AUTO. I haven't changed any secondary timing, but I did test the ram at 3200 by entering the primary timings manually instead of auto, but no luck.
Have you tried this DoneADougalOnSofa?

 

[JasonOrme]

So now since fitting my ram and keeping it at 2933 I've been encountering some strange behaviour.

Programs I install wont open. Some say they have corrupted and need to be reinstalled.
One exe even opened up a different program.
Destiny 2 crashed with a corrupted binary error.
Installs seem to freeze.

And I've had a few different bsod.
Usbhub.sys
Ql2300i.sys
Pagefile.sys
Ntiskrul.exe

I formatted my drives and did a fresh installation of windows 10 but got bsod either during or after installation.

Could this be caused by faulty ram?.

At first I thought it was a problem with my SSD but Samsung magician says the ssd is healthy with no errors.

Currently running memtest86.
Ran it for 9 hours with all 4 sticks and it crashed before It could finish.
So I'm currently doing one stick at a time in the same ram slot.

Very confused.

 

[JasonOrme]

Looks like my Win 10 Boot USB is damaged.
I created a new Win 10 Boot on a different USB stick and so far no BSOD with this installation.
I also memtested all 4 sticks of RAM and all were fine.

I'm currently just using 1 stick of 8GB at the moment.

One thing I did notice was my first 2 sticks from a year ago are identical to the new 2, apart from my old ones say ver 4.32 and the new ones say ver 5.32.

 

[JasonOrme]

I ended up dropping the frequency to 2966mhz and putting the dram voltage to 1.37v.
It's been stable now for about a month and a half.

 

[JasonOrme]

I ended up in a clear cmos booting hell for awhile too.

I found that putting the 2 sticks in first, and then changing the frequency and voltage, booting into windows and then shutting down the computer and then putting the other 2 in worked.
If I stick them all in at once It just wouldn't boot and I would end up having to clear CMOS.