New to Overclock.net
Join Date: Sep 2011
Location: North Jersey (NYC Area)
Here are some preliminary results and some data for my chip. Batch X022C643 MM# 99A5N8 & S-Spec SRG19... There isn't much data that I could find on this chip but it appears to be a newer R0 stepping according to some googling.
So far the chip needs around 1.30v - 1.31v for 5.0ghz stable, depending on the stress test. This is with AVX -2 and a cache of 46. After vdroop (line 2 - Asrock Z390 Phantom iTX) vcore readout in HWinfo is around 1.248v - 1.28v. Temperatures are in the mid to low 80's after 4 hours of benching. My cooling solution is a modest 50$ Thermalright true spirit 140 power with a Noctua 140mm Redux fan. CPU Package power is between 195 - 205w depending on the bench/stress.
So far, my chip is plagued with CPU Cache L0 Errors. I can boot, bench and run non-AVX P95 worklouts without error (application detected) at 50x multiplier as low as 1.23v but to fully get away from the L0 WHEA errors, I need much much more voltage. Thoughts on this? Is it normal? The chip can pass a quick 1 hour RealBench passes without WHEA error (50x 46c -2 AVX) at much lower voltages than required for "P95 L0 WHEA free" testing.
Anyway. I think there is lots of room left to tune this chip, I just need the time. I'll be switching to a custom loop (CaseLabs S3 build), time permitting in the near future.
Any tips/tricks you guys would like to share? Thoughts on how I should be stress testing to validate? P95 Small FTT w/ AVX is brutal (temperatures) on my chip, however, RealBench AVX isn't that bad. I think I could easily go 0 AVX with RealBench testing, pass and maintain acceptable temps, but I don't think P95 w/AVX would be possible for my current cooling solution.
Last edited by Spin Cykle; 07-29-2020 at 04:48 AM.