Memory Problems

[freeagent]

I do try very hard not to get down about things

Just think, stock is just fine too.. it’s better than a Dell

 

[tabascosauz]

just try my procedure above should correct the speeds changing at least

gear Down mode, lowers ram voltage when the pc is idle or asleep it overrides xmp settings and manual settings. soset xmp then disable gear down mode save and exit after above procedure let me know if it worked i had similar issue when i played around with it

This is straight up not how Geardown works at all.

Power Down Mode is the name you're looking for.

 

[Blaeza]

Just think, stock is just fine too.. it’s better than a Dell

I know about Dells... My first PC lol
16GB DDR3, did have a GT710 but that changed to a GT1030 GDDR5, then to a AMD Sapphire RX470 Nitro+ 4GB and a i5 2400. All those years ago (2020)

 

[A Computer Guy]

Well, I actually have XMP WORKING now and I did disable geardown mode and it's fine with youtube?!? I've not adjusted vsoc or anything else for that matter.

Well that's interesting. Just to clarify so you're running ok at DDR4-3200 (XMP) or at your OC of DDR4-3800? Can you take a new screenshot of your zen timings? I'm curious to see if settings landed a bit differently than what you previously showed us.

 

[Blaeza]

Well that's interesting. Just to clarify so you're running ok at DDR4-3200 (XMP) or at your OC of DDR4-3800? Can you take a new screenshot of your zen timings? I'm curious to see if settings landed a bit differently than what you previously showed us.

Here you go! 3200XMP

 

[A Computer Guy]

Here you go! 3200XMP
View attachment 314780

Just some follow-up questions, I would understand completely if you didn't want to mess around anymore now that things are working again.
1) if you only flip GDM back to enabled does the error still occur?
2) if you reduce tRFC back downward to 580 (or 305ns) does the problem still occur?

 

[Blaeza]

Just some follow-up questions, I would understand completely if you didn't want to mess around anymore now that things are working again.
1) if you only flip GDM back to enabled does the error still occur?
2) if you reduce tRFC back downward to 580 (or 305ns) does the problem still occur?

I like you, you are brave... NOOO!!!!!
Joking aside I'm happy to try what you say, but not now. I'm actually not sure how I would? Sorry I'm a bit stupid at times.

 

[A Computer Guy]

I like you, you are brave... NOOO!!!!!
Joking aside I'm happy to try what you say, but not now. I'm actually not sure how I would? Sorry I'm a bit stupid at times.

I just find it difficult to believe GDM Enabled was the issue but since your are working ok now it's easy to test for confirmation and revert back if needed. If GDM enabled does not cause the original problem to recur, then whatever did cause the original problem can potentially blindside you again in the future, perhaps some other aspect of the original overclock.

 

[Blaeza]

I just find it difficult to believe GDM Enabled was the issue but since your are working ok now it's easy to test for confirmation and revert back if needed. If GDM enabled does not cause the original problem to recur, then whatever did cause the original problem can potentially blindside you again in the future, perhaps some other aspect of the original overclock.

Righty dokey skip, I'll try it now. I like your thinking. Back in 10.

EDIT Enabled again, off to youtube and some Metroid Dread.

 

[tabascosauz]

Righty dokey skip, I'll try it now. I like your thinking. Back in 10.

EDIT Enabled again, off to youtube and some Metroid Dread.

For general stability and VDIMM requirements, 1T >>> GDM ≈ 2T. At the higher end of VDIMM (ie. 1.45V+) and at higher freq (ie. 4000+) the differences between GDM and 2T can start to widen, but at 3200 all you need to know is that they both make stability much easier than pure 1T (1T GDM off).

I think most people doing a lot of mem OC just get into the habit of disabling Power Down on a desktop so as to rule out any potential problems it causes.

 

[Blaeza]

@A Computer Guy
I've got mechanics syndrome. Put on YouTube, passed out and YouTube was still going! It's annoying as hell when something won't go wrong...

For general stability and VDIMM requirements, 1T >>> GDM ≈ 2T. At the higher end of VDIMM (ie. 1.45V+) and at higher freq (ie. 4000+) the differences between GDM and 2T can start to widen, but at 3200 all you need to know is that they both make stability much easier than pure 1T (1T GDM off).

I think most people doing a lot of mem OC just get into the habit of disabling Power Down on a desktop so as to rule out any potential problems it causes.

So my ram as it is is fine, if I went to the more extreme end then it would be more likely to fail? As you probably read, I tried to recreate the issue but it's not having it. Any thoughts on what else it could be?

 

[tabascosauz]

@A Computer Guy
I've got mechanics syndrome. Put on YouTube, passed out and YouTube was still going! It's annoying as hell when something won't go wrong...

So my ram as it is is fine, if I went to the more extreme end then it would be more likely to fail? As you probably read, I tried to recreate the issue but it's not having it. Any thoughts on what else it could be?

No, the takeaway is that Gear Down is completely and utterly barking up the wrong tree lol, don't worry about it now that it's on either GDM or 2T.

I think I would pin it on Windows being Windows, honestly, being as old an install as it is. HW accel issues in Chrome is more an Nvidia thing and your BSOD didn't really point to it either.

What matters is that TM5 goes through fine. I would run the shorter 1usmusv3 config if you have some time, but it looks pretty settled to me. The Fabric and RAM speeds are also generally too low to provoke CPU memory controller instability.

You still have logs for the WHEA 18s?

 

[Blaeza]

No, the takeaway is that Gear Down is completely and utterly barking up the wrong tree lol, don't worry about it now that it's on either GDM or 2T.

I think I would pin it on Windows being Windows, honestly, being as old an install as it is. HW accel issues in Chrome is more an Nvidia thing and your BSOD didn't really point to it either.

What matters is that TM5 goes through fine. I would run the shorter 1usmusv3 config if you have some time, but it looks pretty settled to me. The Fabric and RAM speeds are also generally too low to provoke CPU memory controller instability.

You still have logs for the WHEA 18s?

So I have been mislead! Hmmm... Well if youtube goes wrong again, I shall be in a quandary but I'm on TPU, where everyone is helpful.

I've no idea how to do a clean install, but some reading will fix that and I may do that sooner rather than later.

1usmus is part of another program, isn't it? Ryzen something or other, err...

All the whea 18 errors were on Win 10 in march, so no unfortunately. It was a critical error, I remember that much, just instant off.

As this is my thread, I have ANOTHER issue and that is I have a Macrium reflect? image of a Win 10 back-up I was advised to make when all the ram OCing was being done.
How would I go about deleting that? Yes, I really am this bad at something I really enjoy (when it works).

 

[tabascosauz]

So I have been mislead! Hmmm... Well if youtube goes wrong again, I shall be in a quandary but I'm on TPU, where everyone is helpful.

I've no idea how to do a clean install, but some reading will fix that and I may do that sooner rather than later.

1usmus is part of another program, isn't it? Ryzen something or other, err...

All the whea 18 errors were on Win 10 in march, so no unfortunately. It was a critical error, I remember that much, just instant off.

As this is my thread, I have ANOTHER issue and that is I have a Macrium reflect? image of a Win 10 back-up I was advised to make when all the ram OCing was being done.
How would I go about deleting that? Yes, I really am this bad at something I really enjoy (when it works).

Nope, not yuriy's other programs, just another config for TM5. It's in the folder. Just rename it to MT.cfg, replace the existing MT.cfg, and delete cfg.link then run it again.

 

[Blaeza]

Nope, not yuriy's other programs, just another config for TM5. It's in the folder. Just rename it to MT.cfg, replace the existing MT.cfg, and delete cfg.link then run it again.

I shall try that very shortly. Thanks!

EDIT Not shortly, but some time after I ran that 1usmusV3 and all clear. I clicked okay and it disappeared. But I remember it said 41 minutes or so.

 

[basco]

maybe not related but in your system pic above i see a 4pin atx12v cpu cabel?
do ya have the 8pin atx12v cpu plugged in or just 4pin?

 

[Blaeza]

maybe not related but in your system pic above i see a 4pin atx12v cpu cabel?
do ya have the 8pin atx12v cpu plugged in or just 4pin?

That was my first PC from 2020, that was only a 4 pin.
My current system is 8 pin and has an 8 pin plugged in. Sorry for not replying sooner, it didn't notify me.

 

[AMF]

wellat any rate it worked 4 me and 4 him glad its working good now