- Oct 10, 2019
- 886
- 175
- 16,390
I have two memory related questions with regards to an issue I've been running into lately. I apologize for the long post, but I have spent a lot of time with this problem already.
I suspect the IMC in my 9900K CPU has degraded to some degree.
I have run medium/"high" frequency RAM with my i9 9900K since new (First a set of 3333MHz Corsair LPX RAM and more recently a set of 3600MHz G.Skill Trident Z), but lately I've run into BSOD issues with no stop code (like 0x00**** for example), but a KMODE Exception Not Handled error.
When looking up the KMODE... error, it primarely comes back as a driver crashing or possible RAM errors. To begin with, I mistakingly identified my netvork driver as being the issue, since all MiniDumps pointed to applications which used my LAN. But several clean Win 10 installs later, the BSOD persisted.
When running MemTest86 on the G.Skill RAM, with XMP on, it fails quite badly - but only a couple of months ago, they came back error free, even at XMP on setting and 8 consecutive MemTest86 passes - so the problem has developed within approximately 2 moths.
When running MemTest86 with XMP OFF now, the RAM passes 8 consecutive passes of MemTest.
However, I decided to buy a set of Kingston HyperX Fury 3200MHz RAM, which I currently use, because I suspected the RAM might have become faulty somehow.
But I still got the BSOD, although less often at 3200MHz.
Since that, I was beginning to suspect the IMC could no longer handle the XMP settings, and I gradually lowered the Memory speed to a point, where I'm now at the CPU's IMC native 2666MHz - and it seems to have been stable for the last week - which is a good improvement (with the 3600MHz G.Skill RAM, the BSOD occured on a daily basis).
The CPU is 1 year old, and has since day one beeen running 6 months with 3333MHz RAM (no BSODs) and another 6 months with the 3600MHz RAM - again without any BSODs, until about a month ago. I'm a bit surprised with this development, but I guess it isn't unheard of.
I was on UEFI V1.70 when the issues ocurred initially, I’m now on the most recent UEFI version.
Windows fully updated. Fast boot and safe boot disabled, as per some instructions I found via Google on how to handle KMODE... error.
So question 1) Am I right, that the IMC has possibly degraded, and can no longer cope with high speed RAM (which I know is technically an OC)?
Question 2) Since my 3600MHz G.Skil RAM passes 8 consecutive passes of MemTest86 with XMP off, is it safe to assume the RAM is in error free condition? I don't need two sets of RAM, so I'd like to sell the 3600MHz kit, but I don't want to sell someone a bad set of RAM.
Do you agree, that the RAM is most likely ok, and the problem lies elsewhere (with the CPU)?
By the way, I have had 2 different motherboards as well, during this time, and I've been using my current motherboard for the last 6 months.. -I don't really suspect any issues with the motherboard - but I just though I'd mention it.
No overclocks or anythinglike that, except for using the XMP profiles for the RAM.
Thanks a lot in advance, for any input
EDIT: By the way, I also ran manually set IO and SA voltages, as I found the two values to be unneccissarily high, although not abnormal at Auto. I have since the problem began, removed the manually set IO and SA voltage, and set them back to "Auto"
I suspect the IMC in my 9900K CPU has degraded to some degree.
I have run medium/"high" frequency RAM with my i9 9900K since new (First a set of 3333MHz Corsair LPX RAM and more recently a set of 3600MHz G.Skill Trident Z), but lately I've run into BSOD issues with no stop code (like 0x00**** for example), but a KMODE Exception Not Handled error.
When looking up the KMODE... error, it primarely comes back as a driver crashing or possible RAM errors. To begin with, I mistakingly identified my netvork driver as being the issue, since all MiniDumps pointed to applications which used my LAN. But several clean Win 10 installs later, the BSOD persisted.
When running MemTest86 on the G.Skill RAM, with XMP on, it fails quite badly - but only a couple of months ago, they came back error free, even at XMP on setting and 8 consecutive MemTest86 passes - so the problem has developed within approximately 2 moths.
When running MemTest86 with XMP OFF now, the RAM passes 8 consecutive passes of MemTest.
However, I decided to buy a set of Kingston HyperX Fury 3200MHz RAM, which I currently use, because I suspected the RAM might have become faulty somehow.
But I still got the BSOD, although less often at 3200MHz.
Since that, I was beginning to suspect the IMC could no longer handle the XMP settings, and I gradually lowered the Memory speed to a point, where I'm now at the CPU's IMC native 2666MHz - and it seems to have been stable for the last week - which is a good improvement (with the 3600MHz G.Skill RAM, the BSOD occured on a daily basis).
The CPU is 1 year old, and has since day one beeen running 6 months with 3333MHz RAM (no BSODs) and another 6 months with the 3600MHz RAM - again without any BSODs, until about a month ago. I'm a bit surprised with this development, but I guess it isn't unheard of.
I was on UEFI V1.70 when the issues ocurred initially, I’m now on the most recent UEFI version.
Windows fully updated. Fast boot and safe boot disabled, as per some instructions I found via Google on how to handle KMODE... error.
So question 1) Am I right, that the IMC has possibly degraded, and can no longer cope with high speed RAM (which I know is technically an OC)?
Question 2) Since my 3600MHz G.Skil RAM passes 8 consecutive passes of MemTest86 with XMP off, is it safe to assume the RAM is in error free condition? I don't need two sets of RAM, so I'd like to sell the 3600MHz kit, but I don't want to sell someone a bad set of RAM.
Do you agree, that the RAM is most likely ok, and the problem lies elsewhere (with the CPU)?
By the way, I have had 2 different motherboards as well, during this time, and I've been using my current motherboard for the last 6 months.. -I don't really suspect any issues with the motherboard - but I just though I'd mention it.
No overclocks or anythinglike that, except for using the XMP profiles for the RAM.
Thanks a lot in advance, for any input
EDIT: By the way, I also ran manually set IO and SA voltages, as I found the two values to be unneccissarily high, although not abnormal at Auto. I have since the problem began, removed the manually set IO and SA voltage, and set them back to "Auto"
Last edited:
Facebook
Twitter
Instagram