26 minutes ago, Paul J said:

I may be wrong here, (someone PLEASE correct me if I am) but your siggy shows an Intel i5-12400 - I don't think it's overclockable (it's a locked chip) and and Intel shows max turbo as 4.4-gig. I might suggest manually setting it, but not XMP. I've always found that will cause "issues".
Shoving the memory timing up will be ok. If you really want to clock it - go and get an i7-x. They can be pushed.

I thought XMP changed the memory only not the CPU. See https://www.crucial.com/support/articles-faq-memory/what-is-xmp#:~:text=Extreme Memory Profile (XMP) was,specific needs of the memory.

May be I am wrong?

 

you have two glaring issues (in my humble opinion going on)

1 .. The Memory is not on the motherboard QVL

2. Filling up ALL the slots usually increases complications (Memory Controller issues, 4 different qualities of silicon to mange, xmp issues, latency ect)

 

Chipset newness is also a thing. In other words .. the bios for your mainboard is probably going to have a few revisions and releases for a bit which may or may not help matters (since the ram is not on the QVL)

 

What i would recommend in light of this is to manually bring the speed down to  the INTEL Alder lake recommended DDR4 speed  to DDR4 3200 and regain performance by tightening the timings as much as possible.

Your other option is to find out and ensure that the RAM being used is getting the manufacturer recommended voltage when the xmp profile is enabled. But even then you still have the memory controller bogged down trying to manage 4 slots + xmp above Intel's recommendation which may be the straw breaking the camels back.

So yeah .. manual  DDR4 3200 and start with cas 16 ? (if it was two slots used i would have recommended cas 14 but again .. all 4 slots used)

Edited January 22, 20224 yr by Maxis

Yeah - you're right, pgde: it's been a long time (sigh)  and I immediately associated it with o'clocking the proc. It's what you get for quickly skimming through a post. . . Sorry..
However - all XMP does is set the memory to the on-stick rated speed and timing, so that it should be running at its correct speed, as well as setting up the right voltage for it.  That doesn't check the memory for any defects, though - and that might be the case here (after reading 109's post again.)  but it might also be the motherboard too.
MemTest is best used to check the memory, and run two or three times, swapping the sticks around in between tests as well. https://www.techpowerup.com/memtest64/

pj

 

 

Edited January 22, 20224 yr by Paul J

15 hours ago, 109Sqn said:

• Previous Page
• Next Page

 

Thanks for all the responses so far. I'll try to respond to individual points soon. But the first thing I'm going to try is updating the AMD Adrenaline drivers as I've been running version 20.8.3 (from October 2020 if I remember) ever since: 

CTD with Radeon 5700 XT, becoming more frequent - Microsoft Flight Simulator (2020) - The AVSIM Community in January 2021. If that immediately stops the microstutters, I'll stick with it whilst addressing the other problems. If not, I'll revert to 20.8.3 to concentrate on the RAM issues first.

Maybe not the most logical way to tackle things but it might help decide if the microstutters and RAM are separate or connected problems.

 

 

---

• Previous Page
• Next Page

 

 

 

My suggestion would be to forget the sim for now and concentre on other stability stress tests. 

I notice you have all Dimm slots populated which is not always ideal, two tends to be more stable but shouldn't be any kind of an issue if you aren't overclocking. My Apex board for example is a dedicated OC board and has just two slots.

I would set everything in the BIOs to optimised defaults.

Make sure you have the latest BIOS.

All other motherboard drivers, chipset drivers, updated. Check on the Asus website for the latest driver updates, in fact update everything that you installed relating to the board.

Update graphics card drivers too.

First, if you look in your BIOS you will see "Memtest" so run Memtest and check for any RAM issues. 

https://rog.asus.com/articles/crosshair-motherboards/memtest-user-guide-for-rog-motherboards/

Set XMP. Run something like Cinebench R23 and monitor the temp with Hardware Monitor. If temps are good and you aren't getting any overheating and thus CPU throttling, make a note of the Cinebench score and compare to others who have the same CPU, to make sure you are in the ball park.

Run other stress tests too, like Asus RealBench.

If all looks good, try other software. If still looks good I would suspect a sim issue. 

If you get crashes with other software other than just the sim, you can tweak the RAM timings down. I would do this first rather than dropping frequency. Avoid sub timings. 

Sometimes dropping the RAM Command rate down can help. 

 

Which XMP profile are you using? And does that profile match the stated speed and timings for your RAM?

There's usually more than one profile.

Sometimes voltage can need tweaking. 

 

 

 

 

 

 

Edited January 23, 20224 yr by martin-w

15 hours ago, pgde said:

I thought XMP changed the memory only not the CPU. See https://www.crucial.com/support/articles-faq-memory/what-is-xmp#:~:text=Extreme Memory Profile (XMP) was,specific needs of the memory.

May be I am wrong?

 

No, you are not wrong. XMP is a form of memory overlocking. It doesn't matter if you have a non "K" CPU or not, you can still use XMP. 

Right, I haven't had a whole lot of time to work on this over the weekend, but here is what I've experienced so far...

 

 

 

On 1/22/2022 at 8:10 PM, 109Sqn said:

•  the first thing I'm going to try is updating the AMD Adrenaline drivers as I've been running version 20.8.3...
•  

If that immediately stops the microstutters, I'll stick with it whilst addressing the other problems. If not, I'll revert to 20.8.3 to concentrate on the RAM issues first.

That just added to my problems☹️ I uninstalled 20.8.3, then ran AMD's cleanup app to be sure I was on the basic Windows driver. But when I went to install the latest Adrenalin driver [21.10.2 Recommended (WHQL)] (not 21.11.2 Optional), I got an error that vc_runtimeMinimum_x64.msi couldn't be found on my system. I've not had any Visual ++ errors before now. I tried to resolve that with a reboot, using the Windows Create Install Media fix,  repairing any C++ files where that was an option in Windows Settings>Apps, following advice from How to download vc_runtimeMinimum_x64.msi (windowsreport.com) as well as Windows help pages, to no avail - the error still appeared. So even a full C++ install package didn't work. However, Device Manager shows no problem with the Graphics driver.🤔 But messed up C++ is worrying me (even though I've never really had a clue what it does!).

Anyway, running MSFS with the newest driver appears to offer no improvement - I still have microstutters.

 

On 1/22/2022 at 8:19 AM, rka said:

Your RAM should be able to do 3200 with a sensible CL (try 16) if its rated for 3600. I wouldn't fiddle around with HT. Just leave it on. MSFS is contemporary software and not from 1995 🙂

I've left HT on.

 

On 1/22/2022 at 8:55 AM, iwebber said:

Check your RAM voltage, over the years I've noticed XMP doesn't always set the voltage correctly (if at all), it ups the frequency and tightens the timings but leaves the voltage 🙄.

I don't know exactly what memory kit you have but check the specs for the highest voltage supported then set that as the memory voltage in the BIOS.

The BIOS suggests the voltage is correct - 1.35v. I'm running Corsair Vengeance LPX Black DDR4-RAM 3600 MHz 2x 16 GB - CMK32GX4M2D3600C18  That's 2 lots of 2x16GB. I've been sure to keep them matched in the pairs they arrived in

 

On 1/22/2022 at 11:23 AM, bobcat999 said:

I wonder if you have had a RAM module go bad.  It can happen.  Have you tried running a RAM checker? Some motherboards even have one built into the BIOS now.   

Voltage is normally set by the board automatically on XMP.  I have run XMP1 (3600) on my old PC for 3 and a half years now and no problems in MSFS.

So if your RAM can't take 3600, try 3400 or 3200 set manually with an appropriate voltage looked up on the internet.

PS.  Not sure about W11 helping you RAM problem, but I would definitely consider it when running Alder Lake chips. 

Re: RAM checker, see my next post.

Re: XMP. At first, I thought there was only one XMP mode on this mobo because they confusingly have the XMP enable/disable button on the home page, but actual XMP settings deeper within the BIOS.🙄 More on that in the next post.

There's a serious temptation to go ahead and upgrade to Win11 because it's an Alder Lake chip, but I'd worry that rather than solving my problems, I'd simply be importing them into the newer OS, possibly making things worse. It feels the better solution is to try and fix everything I can in Win10 then upgrade. But eventually I might feel forced to try that route instead.

 

Really appreciating all the suggestions so far, even if I haven't responded to you all individually.👍

On 1/23/2022 at 11:15 AM, martin-w said:

My suggestion would be to forget the sim for now and concentre on other stability stress tests. 

Agreed, makes sense.

On 1/23/2022 at 11:15 AM, martin-w said:

I notice you have all Dimm slots populated which is not always ideal, two tends to be more stable but shouldn't be any kind of an issue if you aren't overclocking

I've never understood why most mobos allow more sticks than ideal, but that's a long discussion for another time If I an get all these issues fixed, I may just content myself with 2 sticks on this PC as MSFS rarely never close my maximum RAM.

On 1/23/2022 at 11:15 AM, martin-w said:

I would set everything in the BIOs to optimised defaults.

Make sure you have the latest BIOS.

All other motherboard drivers, chipset drivers, updated. Check on the Asus website for the latest driver updates, in fact update everything that you installed relating to the board.

Update graphics card drivers too.

Check, check, check and check, though as you can see in my previous post I have had problems even with the GPU driver update. It definitely shows as the newest version, despite the C++ file missing during installation. 

On 1/23/2022 at 11:15 AM, martin-w said:

First, if you look in your BIOS you will see "Memtest" so run Memtest and check for any RAM issues. 

Set XMP. Run something like Cinebench R23 and monitor the temp with Hardware Monitor. If temps are good and you aren't getting any overheating and thus CPU throttling, make a note of the Cinebench score and compare to others who have the same CPU, to make sure you are in the ball park.

Run other stress tests too, like Asus RealBench.

Didn't spot Memtest in the BIOS but I'll have a look for it. I've been trying the MemTest suggested by Paul J and I'll report on that in my next post.

I'm trying with XMPII first, before 'downgrading' to XMP I.

I'll use your test suggestions next...

On 1/22/2022 at 10:13 PM, Paul J said:

but it might also be the motherboard too.

I blooming hope not as it's a brand new board to replace a dead MSI.🤯

On 1/22/2022 at 10:13 PM, Paul J said:

MemTest is best used to check the memory, and run two or three times, swapping the sticks around in between tests as well. https://www.techpowerup.com/memtest64/

This is causing all kinds of frustration! I get errors with all 4 sticks, whether combined or individually. I've so far tried them in the recommended D2 slot for a single stick but I suppose I should do at least one test for each slot to rule out the slot itself as a fault.

An example of what I see:

I've tried Indefinitely and limited to certain numbers of loops or up to 10 hours. Faults every time and sometimes ending up with a BSOD, with errors such as 

REFERENCE_BY_ POINTER
UNEXPECTED_STORE_EXCEPTION
??????????????? (no time to read the error message before restart, but I think it referenced Memory)

I have also run sfc /scannow in Command Prompt. Too many errors to quote the whole thing, but for example:

Quote

Package_for_KB3025096~31bf3856ad364e35~amd64~~6.4.1.0 [HRESULT = 0x800f0805 - CBS_E_INVALID_PACKAGE]
2022-01-23 00:42:53, Info                  CBS    Failed to internally open package. [HRESULT = 0x800f0805 - CBS_E_INVALID_PACKAGE]
2022-01-23 00:42:53, Info                  CBS    Failed to create open package. [HRESULT = 0x800f0805 - CBS_E_INVALID_PACKAGE]
2022-01-23 00:42:53, Info                  CBS    Failed to OpenPackage using worker session [HRESULT = 0x800f0805]
2022-01-23 00:42:53, Info                  CBS    Session: 30937074_674373510 initialized by client WindowsUpdateAgent, external staging directory: (null), external registry directory: (null)
2022-01-23 00:42:53, Info                  CBS    InternalOpenPackage failed for Package_for_KB3025096~31bf3856ad364e35~x86~~6.4.1.0 [HRESULT = 0x800f0805 - CBS_E_INVALID_PACKAGE]
2022-01-23 00:42:53, Info                  CBS    Failed to internally open package. [HRESULT = 0x800f0805 - CBS_E_INVALID_PACKAGE]
2022-01-23 00:42:53, Info                  CBS    Failed to create open package. [HRESULT = 0x800f0805 - CBS_E_INVALID_PACKAGE]
2022-01-23 00:42:53, Info                  CBS    Failed to OpenPackage using worker session [HRESULT = 0x800f0805]

I seem to have gone from a very stable Windows installation to a spaghetti pile of problems sine being forced to change CPU and mobo.😩 And there I was thinking I was going to see improvements.

12 minutes ago, 109Sqn said:

I seem to have gone from a very stable Windows installation to a spaghetti pile of problems sine being forced to change CPU and mobo.😩 And there I was thinking I was going to see improvements.

 

RMA the board and get Asus to send you a new one. If no improvement then at least you've eliminated one variable. Asus aren't known for their efficient support, but if you have a chat with them they may be able to help. 

Pretty sure Corsair RAM has a lengthy warranty too. If in doubt RMA it. 

Highly unlikely anything wrong with the CPU. CPU's are very very rarely faulty from new. 

Do try Memtest. Always worth trying more than on software when checking for issues. 

Edited January 24, 20224 yr by martin-w

Maxis is right on concerning the memory not being on the motherboard QVL It is entirely possible this is the cause of all of these memory errors. At the same time - they may be the cause of the issues that scannow is coming up with.
However -  before RMA'ing the mobo - do you have a spare drive kicking around that you can do a clean OS install on? Then run the MemTests again. This way you can verify that the OS is ok, and that the issues are not caused by problems with the OS, or with MSFS or drivers  or anything else.

(I am assuming here that you had replaced the mobo and the cpu, but used the original "C:\"  OS/disk from the previous CPU/mobo? i.e. you didn't do a fresh install of the OS on a clean drive - right?)

Edited January 24, 20224 yr by Paul J

10 minutes ago, Paul J said:

I am assuming here that you had replaced the mobo and the cpu, but used the original "C:\"  OS/disk from the previous CPU/mobo? i.e. you didn't do a fresh install of the OS on a clean drive - right?

You assume correctly. I'd just bought a new M.2 SSD as the one with my MSFS addons was full, so the one it replaced is now empty. 

So... use Media Creation Tool to create a USB install for the blank SSD. And I'll need my Install key, won't I?  I'll just be prompted to activate (which I won't need to do for the purpose of testing).

Got that all right, yes?

Yes - how for you can get before you're prompted to enter the key - I'm sorry 109 - I just don't remember.  Someone else will have to jump in. I reckon it will be quicker than all the work involved in pulling the new mobo though. At the same time - you should also disconnect any other drives that you have installed, too. Make it as clean as you can. minimum memory, memtest, add a stick, memtest again - switch slots - memtest again - it's an awful long process.

1 minute ago, Paul J said:

Yes - how for you can get before you're prompted to enter the key - I'm sorry 109 - I just don't remember.  Someone else will have to jump in. I reckon it will be quicker than all the work involved in pulling the new mobo though. At the same time - you should also disconnect any other drives that you have installed, too. Make it as clean as you can. minimum memory, memtest, add a stick, memtest again - switch slots - memtest again - it's an awful long process.

I think I'll be prompted during the actual install IIRC. Thankfully i found my DVD if I need the key.

It's already been a long process, but I agree it would be quicker and less hassle to try a clean install and test first. I'd probably be happy if it was only a Windows issue, even if it meant reinstalling MSFS and everything else. I've got the same Crucial RAM but 3200MHz in another PC so even if the 3600 fails, I can then test with that. And if that worked, I'd probably just swap them over. That PC is mainly for Photo/Audio/Video editing, so would probably be happy with 3600, and is running on a Ryzen 5 3600 just as this one was before the MSI mobo failed.

Yup: Once (assuming it passes memtest in this config) you can then dig up some timings from the overclocker websites and then manually tighten them, Memtesting and monitoring each stage. Let us know how it goes.

Edited January 24, 20224 yr by Paul J

3 hours ago, Paul J said:

Let us know how it goes.

It's just gone from bad to worse.

I created a bootable USB, disconnected all drives and rebooted the PC with the USB drive plugged in to a rear socket. Got into the BIOS fine and there was the USB drive listed as the only bootable source. Saved the setup (not that I'd changed anything anyway as I'd set the BIOS defaults beforehand) and rebooted. The Black screen with the TUF Gaming logo appeared and then added a message below that it was preparing automatic repair. "Ok, maybe that's in preparation for reading from the USB drive", I thought... But it hung on that screen for several minutes before I decided to power down (which only took a few seconds so I believe things were still only in the early POST or BIOS stages) 

Wondering if the USB files were incomplete or corrupt, I created them on a different USB drive. Same result. So I reinstalled the Boot SSD. Still unable to get to Windows - again I had the preparing automatic repair message. Each time I rebooted, the same happened - I'd either get two of what would be a revolving circle of dots had it ever completed, or no dots at all. 

I eventually put all the SSDs back in but I still can't get into Windows or get either USB to get to the fresh install stage - I just get the TUF Gaming screen, with or without the repair preparation message.

 

Now things are really screwed...

 

 

EDIT: Update to the above

Looks like I found the issue with installing Windows to the empty drive. Looking further into the BIOS, I found a setting for CSM (think that's what it said) and one setting there was to allow legacy drives to boot the system. That seemed to be a possible change I needed to make, and so it was. I'm currently installing Windows - just reached the Region prompt...

Edited January 25, 20224 yr by 109Sqn