Hi
This is my third attempt to improve the settings of my computer with a
5800X processor cooled with an NHD15 and Crucial Technology BL8G32C16U4B. M8FE1 (E51CFE5F) DDR RAM on an x570 chipset -ASUS TUF GAMING X570-PLUS.
If the first time I made settings with several types of memory just to raise the memory synchronously with the cpu in stable conditions, without exceeding 3600MHz, after a period of time I researched the forum here very carefully, on the dedicated thread, and I found a configuration with the cpu running synchronously at FCLK 1933 with the RAM, having negative CO -5 on all cores. I basically had the same configuration that Zen timings shows now, with a few differences:
-before I had VDDG CCD and VDDG IOD at the same value - 0.9475V
-the values for CAD Common values were those from ZEN2: 20-20-20-20 compared to 24-20-24-24
All according to the recommendations here:
Inhaltsverzeichnis: RAM OC Board/UEFI Limitierung IMC (CPU) Limitierung ProcODT (CPU) Limitierung RAM ICs Limitierung RAM Taktpotenzial der Mainboards Mainboard Sammelthreads FAQ Hilfreiche Threads RAM OC Nach oben Allgemeines Anleitungen und Guides RAM OC Anleitung für Zen/+...
www.hardwareluxx.de
But HWinfo always showed me a lot of WHEA 18 and 19 errors until now, even hundreds, most of them when I was running the browser with a lot of open tabs. I very rarely experienced a restart, sometimes even in idle.
After almost 2 years, having a few days off, I went back to see what's new, and I see that there is
Corecycler with Automatic Test Mode to automatically adjust the Curve Optimizer values! Done, now I can make fine settings for the most efficient undervolting!
Initially I started with the old settings, from the first Zentimings window (-5 on all cores), with the Corecycler default version, with Prime 95, but immediately WHEA 18 and 19 errors started. Then a restart followed, after 2 complete test runs.
Then I said I wouldn't waste any more time with the old settings, and I switched to running Corecycler with Automatic Test Mode to automatically adjust the Curve Optimizer values, to balance the settings and I found
that
core 0 is extremely weak compared to the rest of the cores (?!).
Yesterday I tested for a very long time, not necessarily rigorously, but intensively browsing through a browser with many tabs. However, although I only had WHEA 19 errors, the test ran until I closed it, for 14 hours, 30 minutes
Today I resumed the tests, but I made the following adjustments:
-
VDDG CCD increased to 0.96V (0.9592V real) and
VDDG IOD to 0.97V (0.9681V real), actually adjusted as in Zentimings
-
CAD Common values 24-20-24-24
The test ran WITHOUT WHEA errors for 4 hours and 20 minutes. After 3 hours I boldly started the browser and started writing this message. After about 40 minutes I had a restart, without seeing any indication in the event viewer. Because I am aware that this is not how a proper test is done, I still chose to post here what I obtained, and will do the recommended tests next week without interfering with other activities.
I would appreciate any advice on improving stability, especially regarding VDDG CCD and VDDG IOD, about tRFC or other values that you may notice are inappropriate.
Thank you.
www.asrock.com
.
)
