05-23-2019 01:53 AM - last edited on 03-06-2024 08:15 PM by ROGBot
05-23-2019 03:42 AM
Melkor wrote:
Hello,
So like in the title I have a lot of problems to run G.Skill 3000 CL 14 on this board, unfortunately they are not on Asus QVL for MXIH, only on G.Skill.
Before posting I did a lot of research, many tips from ROG forum and other webs.
Here's what I've been trying:
- manual mode, XMP I and XMP II
- all bioses from release to the newest, which I am using now, 1005 if I'm correct.
- DRAM voltage from 1.35 to 1.50
- VCCIO and SA voltages from standard default to even 1.4 (on auto these are at VCCIO 1.3, SA 1.25 for 3000 Mhz...)
- increasing core voltage
I lost about 2 weeks of my life to get this fixed and nothing, my issues are not about booting into WIndows, it boots just fine, stress tests pass, on Karhu RAM Test there are no errors, running almost 3 hours with about 15000% coverage.
Problems starting during gaming, in example Battlefield 5 or Assassins Creed Odyssey. Games run without problems for about 20-40 mins, then just exit to desktop. Nothing is mentioned in reliability monitor or event viewer. On default 2133 everything is good, at 2666 with timing 14-14-14-34 also.
Yesterday I realised hey what about trying other timings 15-15-15-35 at 3000 (DRAM voltage at 1.35, IO and SA at 1.10), it's working no problems.
What do You suggest, any other ideas or just let it be this way?
Maybe someone can post secondary/tertiary timings for this kit just to compare? F4-3000C14D-16GTZR
05-23-2019 04:39 AM
Melkor wrote:
At the moment I'm at BIOS 1005, tried 0506, 0602 and newer, still same issue.
I'm sorry, forgot to mention, issue persist with both tweak modes.
05-23-2019 07:17 AM
Melkor wrote:
Yes RAM is installed in A2 B2 slots, just tried 3600 Mhz 16-16-16-36 2T at voltages You suggested, boots fine but error after 5 mins in Karhu test. Will try CL 17.
05-23-2019 09:29 AM
05-26-2019 12:56 PM
05-27-2019 11:14 AM
Melkor wrote:
Hey guys, so I have found out what was my issue. It was not voltage or timings. I've noticed when soft crash of game happened, my RAM temperature every time was around 53*C. My PC case is Phanteks Enthoo Evolv X with stock fans, 2x 140mm front and 1x 140mm back. My GPU is RTX 2080 Ti founders edition. For a test I used frame cap in AC Odyssey to 60 fps, less work for GPU = less heat inside. My Ram temp was then at around 45*C and I could play for hours with my RAM settings 14-14-14-34. If i removed cap as fast as temp went to 53*C on RAM then the game crashed. Looks like my memory is very temperature sensitive.
It seems stock fan setup was not enough during gaming and it explains why there was no issues during Ramtest or Prime - they only stress RAM/CPU and it's not enough to rise RAM temp to even 40*C when GPU is idling. I hope changing stock fans to Noctua and adding 2 more at the top will improve airflow and lower temps enough because I don't want to cap FPS at 60 on 165Hz monitor 😉