Yeah, I totally agree with @TenPatrol that a reset is not out of the ordinary - my bios will always reset with any new memory or cpu changes.
Does your kit have an XMP profile also? I’m using an XMP profile for my DDR5-8000 kit and I don’t recall there being an EXPO profile. I bought mine last year so maybe they’ve updated it with an EXPO profile.
Keep in mind my motherboard is different than yours so ymmv with my memory timings.
My 6400 RAM used an XMP profile with tightened timings. There was no Expo profile available.
I ended up setting it to 6000 MT/s and adjusted secondary timings. That was my baseline and was very stable. With that kit it was MCLK=3000, UCLK=3000, FCLK=2133. Setting FCLK higher did give me a measurable improvement. I tried FCLK=2200 with that kit and the system wouldn’t boot.
This new RAM (v-color Manta) has both XMP and Expo profiles available.
For sure! Like I said in the earlier post, everything was stable using Expo with no other tweaks.
I started messing with timings and had another HAL moment ("I’m sorry Dave, I’m afraid I can’t do that.) Q-flash didn’t even work to recover it, and I ended up having to put the computer on the bench and clear the CMOS. Ugh… 
I’ve been messing around with BIOS settings since, and I think that with 8000 RAM things are on a bit of a razor’s edge. I could downclock it in order to play with tightening up timings, but then I lose the FCLK=UCLK synchronization, which Buildzoid said is highly recommended.
One change I made that seemed to really help. My motherboard has two options that are supposed to improve DDR performance:
I’m not sure what either one actually does. ‘Low Latency’ had no measurable benefit.
But ‘High Bandwidth’ really improved my OCCT and AIDA benchmarks.
I had one pretty funny thing happen: I was running benchmarks and all of a sudden they tanked. I mean really tanked. Turns out Prime95 was still running, and using all 48GB RAM, even though I’d exited the program before running the next benchmark. Oops!
I think you’ll be very interested to see the data I’ve collected. Synthetic benchmarks are down a bit for the 8000 (Expo) vs. 6400 (downclocked to 6000 with tightened timings.) I kind of expected that with the asynchronous UCLK=MCLK/2.
I did some flights with the new setup, mirroring the exact same flight I’d done with the old RAM.
I saw what I’d hoped for - and bought this faster RAM to get - namely (at first glance) around 20% better 1% Lows.
I did 10 flights, and the 1% Lows were consistent and much improved. I need to tread carefully with trying to tighten timings further. Things look pretty darn good right now.
To reiterate: I’m running the sim @ 1080p, trying to stay CPU-bottlenecked.
I did have a few blue screens when loading Windows, and backed off of a couple of BIOS settings for the moment. I had set SPD PowerDown to ‘disabled’ and applied a PBO -20 all core offset. I changed PowerDown to ‘Auto’ and set PBO to ‘Auto.’
I’ll mess around with PBO later. Right now I’m seeing both CPU and GPU around 60°C during flight. Maybe I don’t understand PBO well enough, but I’m not sure I see the benefit of undervolting when I’m nowhere near thermal limits.
MSFS2024 will immediately CTD if the total of physical memory and virtual memory is exceeded.
If virtual memory is low, it will CTD even with 64GB.
What do you mean when you say, “Virtual Memory”?
Virtual memory is the Committed Memory that can be seen in Task Manager or free tools.
MSFS2024 consumes more than 64GB of committed memory in the photo area, and will crash immediately if the set value is exceeded.
I designed a dedicated button on the back of my PC case that connects directly to the clear CMOS jumper.
Although my motherboard already has a clear CMOS button, its placement makes it hard to access when the case is positioned with its back against the wall. With my custom button, I can easily locate and press it by simply reaching behind the case.
If the committed memory of 50.1GB is not automatically expanded and MSFS CTDs the moment it exceeds 50.1GB, it is better to increase the virtual memory.
I’d like to clarify that virtual memory and committed memory are two distinct concepts. Mixing these terms can lead to misunderstandings.
How does one increase the “Virtual Memory?”
I don’t see anything in my screenshot that mentions that.
Are you referring to the Page File?
Depending on your language, the Windows settings screen may be called page file or virtual memory.
Increasing this will also increase the committed memory area.
If MSFS2024 isn’t CTD, the current situation may be fine.
I see. Thanks for the clarification. I thought maybe there was a setting I was unaware of. I understand that some people use different terminology. And of course some things don’t survive translation very well… 
Anyway, I just let Windows manage my Page File. Some people claim that it’s best to manually set the size. That was true 20 years ago. Not so much today.
Best thing is to figure out why the system is using >64GB RAM (which to me is really excessive and certainly can cause problems if that amount is exceeded.)
That’s the page file I spoke of.
I keep my system with the default “System managed size.”
I think that trying to manage it yourself can cause more problems than it’s worth.
That’s great! Makes all the time and trouble invested worth it 
That’s very interesting - I’ll have to check out my Asrock x670e Taichi motherboard to see if it has something similar.
You can see in this screenshot that I’m using 23GB out of my 48GB of RAM.
It doesn’t feel like I took a step back going from 64GB DDR5-6000 to 48GB DDR5-8000. In fact, I achieved the goal of seeing a significant improvement in 1% and 0.1% Lows.
It was a real struggle getting things stable, though. A BIOS update helped a bunch, as did backing off from trying to manually overclock things. Right now the only overclocking is using EXPO, PBO set to Auto, and a couple of minor tweaks (DDR High Bandwidth enabled, and LLC set to Medium.) The system is very stable now, and I think these results speak for themselves:
ETA: I need to say this: For the majority of those with AM5 systems, DDR5-6000 (or 6400, properly configured) is still the best choice overall. Transitioning to DDR5-8000 was a really painful process. It took me DAYS filled with boot failures, CMOS resets, and BSOD’s to finally get things settled down. 8000 MT/s RAM is really pushing my motherboard/CPU to its limits, and I came very close to abandoning this project.
I do not recommend this upgrade for everyone. I say this not with a “Look at me!” attitude, but with the knowledge that only my 20+ of experience building and repairing high-end computers, plus invaluable help and advice from a fellow member (you know who you are) got me through this process with my sanity intact.
Was it worth it? Ultimately, yes. Would I recommend it to everyone? Absolutely not, and I pray I’m not misleading anyone into thinking otherwise.
ETA: May 28
I’m back on DDR5-6000/CL30.
Started having stability problems again with the 8000 MT/s RAM. 
When I say ‘Stability Problems’ I’m talking about random BSOD’s.
The faster RAM has been returned.
Stability is king…
I was looking up background info on DIMMs and now can’t find what I saw a while ago but saw an article that was saying with AMD, there’s something fairly magical about DDR5-6000 and how the AMD chips deal with memory and other speeds can be real bears to get working right.
Might not even be relevant to your situation - I didn’t read the article but just saw the headline.
AMD is unique in that - unlike Intel - the memory controller is on the CPU. That allows the RAM and the CPU to have a direct data link. That link is represented by two parameters, UCLK and MCLK.
Ideally, you want the speed to be a 1:1 ratio.
Let me explain…
In simplified terms:
Ideally, you want MCLK=UCLK. For DDR5-6000 both would be set at 3000.
The UCLK can’t really run higher than 3200 without potential problems, so DDR5-6400 is the practical limit when keeping that 1:1 ratio.
The UCLK:MCLK ratio can be 1:2
In the case of DDR5-8000 the UCLK=2000 and the MCLK=4000 (1/2 DDR speed.)
This introduces latency.
However, there is another clock called FCLK. That is the Infinity Fabric AMD uses to get data from the CPU (UCLK) and passes to the motherboard chipset. Ideally you want FCLK to equal UCLK, but that’s not possible with DDR5-6000 RAM, since UCLK=3000, and 90% of CPU’s will crash if FCLK >2200.
So DDR5-8000 can have:
UCLK = 1/2 MCLK (Not great for CPU <–>RAM latency.)
FCLK = UCLK = 2000 (That’s good for UCLK<–>chipset latency.)
DDR5-6000 can have:
UCLK = MCLK = 3000 (Great for CPU<–>RAM latency)
FCLK = 2/3 UCLK = 2000 (Not ideal for Chipset<–>CPU latency)
CAS latency (abbreviated CL) is a timing on the memory module itself. Lower CAS latency means data is moved around the memory chips themselves a little faster.
There are literally dozens of subtimings that overclockers tweak to improve RAM speed.
Most of them are concerned with benchmarks.
We are concerned with sim performance.
I’ve discovered first hand that trying to maximize RAM speed is a dangerous road.
To wit:
Stability is King.
Actually, that’s not quite accurate anymore. Intel has also had the memory controller built into the CPU for many years now, just like AMD. While AMD was the first to do it, modern Intel processors also have an IMC.
Thank you for clarifying my misconception. So one difference is that Intel can only run in Gear 2, whereas AMD can run in Gear 1.
What about Infinity Fabric? How does data transfer between the Intel IMC and the chipset work?
