HoneyBadger84

EK got back to me this morning, they immediately said they'd send a new one out again and that they have no idea how it would even be shipped in that condition & they're looking in to it.  They also offered a partial refund in the amount of the shipping costs associated with the returns.  So that's cool.
 
They asked me to open it & see if the AIO itself was damaged, as if it were not I could just keep it & take the refund on the shipping return of the first return, but unfortunately:
 


 
I'm amazed it somehow did not leak.  That is a DEEP hit, the frame of the AIO radiator itself is bent quite a bit on that side.

Yeah the Customer Service people via email/communication have been great so far.  Like I said, they even sent out a new unit before the old one arrived to them... but whoever packaged that just... why? lol 
There's no way you wouldn't notice something like that, and it definitely didn't happen during shipping.  It looks like the box got rammed by a forklift or something, to be honest, the hole is pretty sharp & only damaged the "plastic" the box is wrapped in slightly around the area of impact, that to me says it was something that hit it hard & sudden.
Hopefully they continue to be nice & give me the prepaid return label this time.  I really don't want to be out of $80 worth of "free" fans because of them shipping me a damaged item the second time after giving me a defective one the first time. lol

Advice: If you're wanting a GPU now, but you want a 30-series card for sure, buy an eVGA card from an authorized retailer, then you can register for Step-Up after you get it, for a 30-series card.  This will allow you to have a 20-series card "now" to enjoy, and guarantee that "eventually" you're going to get a 30-series card.
That's basically what I did... then I ended up getting a 3080 directly the other day, it will be here Tuesday.
But yeah, if you're able to, get an eVGA 2070 Super or simiilar, and then register for Step-Up, you get the best of both worlds, a nice upgrade now, and a gauranteed 30-series card "eventually" whenever your turn in the Step-Up queue comes around.

In the long run it will result in shorter AIO pump lifespan at worst, and noise level being higher at best, as the Gamers Nexus video said.
 
All in all, you can only fit it the one way... so that's how it has to go, right? lol
 
Unless you're getting insane pump noise or seeing a notable increase (which you should see a DECREASE from going to push/pull) in CPU temps at idle and load, which would indicate the pump has air in it, I wouldn't really worry about it too much.

Download AIDA64 free edition & run just the CPU stress test (uncheck FPU, stressing CPU & FPU will make your temps go up a lot, so do that second) for at least an hour.  If it passes that, you can consider your CPU pretty stable, then if you are okay with your temps being toasty, run CPU FPU & Memory all at the same time, this will likely make the CPU thermally throttle with your temps, but it's still a good test to run.
 
If the CPU is hitting 100C but not crashing (and instead throttling like it should) then technically speaking that's fine.  I would definitely try Push/Pull as it can improve temperatures in a lot of instances, JayzTwoCents did a radiator fan configuration test that showed Pull actually performs better than Push sometimes, and Push/Pull universally improves temps by 2-7C, so it may actually make enough of a difference to stop the throttling you're experiencing at 100C.

O_O at "below 80" and "great" being in the same sentence. 
 
My MOS readout doesn't go that much above 40-45C under full load while OCed so to me those temps are high, though they are "within spec" since most MOSFETs/VRMs on boards can handle 100-110C before actually throttling/tripping OTP & shutting down, though it's entirely possible if you're going from 60 to 80C on the CPU within 2 minutes, that the MOS temp is going higher & maybe also spiking higher faster, thereby causing issues.  Just a possibility I thought I'd point out since some motherboards have very shoddy VRM/MOSFET cooling, especially when lacking airflow.
 
80 is better than the 98C you were showing previously though, that's a very good improvement.
 
Let us know how things go with the RAM set to just XMP/stock voltage.
 

This, yes.  Set the RAM back to XMP & stock voltage, don't overclock it at all, and test the CPU for stability.  Even if you have previously benchmarked the RAM settings stable, at a higher CPU clock & voltage, the memory controller may be under more stress & the RAM settings are causing the instability, or vice versa, the CPU not having enough memory controller voltage could cause the RAM to be unstable... it's all up in the air, which is why you should do one part at time.
 
I am not super familiar with the 10xxxK series' voltage ranges, but 1.32V sounds like it's decent for 5.1GHz, so I would test just that with the RAM on XMP with no OC & voltage set to what it says to set it to on the package/sticks (1.35V I would assume).
 
Also, 80C should not be causing a crash, so something else is causing it (pretty sure your thermal limit & mine are the same, which is about 100C on the cores).
 
How are your VRM temps with the sides of the case off, better or about the same?   Is it possible to put a fan blowing down on to the VRM area of the motherboard from the top?  I know it sounds sub-ideal to have a top fan as intake, but if the rear of your case is set to exhaust, it should work out just fine.  Any airflow directly over the VRM (be it pulling air off them or pushing are over them) is better than no direct airflow at all, and can make a HUGE difference in VRM/MOS temps.  That may be why you're crashing, your VRM may be overheating & instead of throttling, it's causing a crash first.

Is your RAM set to XMP?  What RAM are you using?  Those timings are VERY loose for 3200MHz unless you have a kit that can't hit 3200MHz at better timings... as a for instance, the 3200MHz version of my RAM has CL14-14-14 timings @ 1.35V.  CL18 @ 3200MHz is cray cray imo.
Also, I highly recommend doing any RAM overclocking LAST, after you've found your happy stable spot on your CPU overclock testing.  It is the hardest part of any overclock if you really wanna go hard on getting your RAM fast & tight timings.
Is there any airflow at all over your VRMs?  Some airflow is better than none.  Where and what direction is your CPU radiator mounted in?
A picture of your system's insides would be very helpful.
 
If you want an idea of how crazy I am on "giving stuff good airflow", witness:

 
That 200mm monster moves quite a bit of air over the VRMs while being relatively silent at 800RPM, and the GPU radiator intake fan is "above" them, to pull any hot air out quicker.
 
Edit: I would not push your CPU any further until you get your temperatures figured out, VRMs running at that temp readout is on the borderline of unsafe.

Is your RAM set to XMP?  What RAM are you using?  Those timings are VERY loose for 3200MHz unless you have a kit that can't hit 3200MHz at better timings... as a for instance, the 3200MHz version of my RAM has CL14-14-14 timings @ 1.35V.  CL18 @ 3200MHz is cray cray imo.
Also, I highly recommend doing any RAM overclocking LAST, after you've found your happy stable spot on your CPU overclock testing.  It is the hardest part of any overclock if you really wanna go hard on getting your RAM fast & tight timings.
Is there any airflow at all over your VRMs?  Some airflow is better than none.  Where and what direction is your CPU radiator mounted in?
A picture of your system's insides would be very helpful.
 
If you want an idea of how crazy I am on "giving stuff good airflow", witness:

 
That 200mm monster moves quite a bit of air over the VRMs while being relatively silent at 800RPM, and the GPU radiator intake fan is "above" them, to pull any hot air out quicker.
 
Edit: I would not push your CPU any further until you get your temperatures figured out, VRMs running at that temp readout is on the borderline of unsafe.

So I ended up buying this thing:
https://smile.amazon.com/gp/product/B07X48X4KM/
 
Got it with a $10 off coupon ($42 with taxes included), it arrived today. 
Assembly was easy (and I'm not one that's good at putting non-computer stuff together), the parts don't feel very high quality, but it is very sturdy & not wobbly/unstable at all once the monitor is on it.  I barely had to adjust the "pressure" at all, it says it's specced for 8-11lbs when it's shipped but my monitor is ~16lbs and it had no issues holding it up on the default setting.  Easy to adjust as well.  The tilt mechanism was a little finnicky but I got that sorted pretty quickly.
 
Figured I'd post an update in case any is looking for a stand/mount & happens by this thread.

So I ended up buying this thing:
https://smile.amazon.com/gp/product/B07X48X4KM/
 
Got it with a $10 off coupon ($42 with taxes included), it arrived today. 
Assembly was easy (and I'm not one that's good at putting non-computer stuff together), the parts don't feel very high quality, but it is very sturdy & not wobbly/unstable at all once the monitor is on it.  I barely had to adjust the "pressure" at all, it says it's specced for 8-11lbs when it's shipped but my monitor is ~16lbs and it had no issues holding it up on the default setting.  Easy to adjust as well.  The tilt mechanism was a little finnicky but I got that sorted pretty quickly.
 
Figured I'd post an update in case any is looking for a stand/mount & happens by this thread.

It's a software issue. Gigabyte and GSkill's software conflict with each other. I have the same issue. One day the RAM just disappeared from RGB Fusion and nothing brings them back. Tbh, GSkill's control software is better and has more options, for the RAM at least. 
 
GSkill's software is still in beta technically I think, and I'm sure that's why it conflicts with rgb fusion 2.0

Stress test for aida64 is merely a good starting point. Do not trust it as an overall stress test.
 
What I mean by that is: if it fails an hour of AIDA64 something is clearly unstable, but if it passes, that doesn't mean you're totally stable.
 
For reference, I've had settings pass an hour of combined CPU FPU Memory stress testing on AIDA64 and fail elsewhere. I've also had a RAM OC pass an hour of Memtest only to fail AIDA64's in 10 minutes. So a good mix of tests, along with just plain gaming on it and seeing if it freezes is highly recommended when testing an overclock for stability. I personally still use Prime95 26.6 as well even though some people swear against it these days as you see "unrealistic Temps" in it. 

32GB if you want more future proofing and you're actually doing work/editing that needs a lot of ram.
 
More is always more better ?? unless you need more faster. Lol

I would recommend selling the old kit when you get the new one. Unless you do tasks that actually need more than 16GB of RAM, the hassle you're going to go through trying to get settings that will make the two sets work better together will probably not be worth it. 
 
Mixing kits at best will make the system default to the lower speed kit's settings. At worst it won't boot until you clear CMOS and set timings manually that both sticks can run at the same speed at. 

Having went from an R9 295x2 to a 1080 Ti & then done a jump from a i7 3930k to an i9 9900K, I can tell you that GPU will show its age for sure in newer games. Like the 295x2, it lacks the raw throughput to handle modern titles at acceptable fps at anything over 1920x1080 especially. 
 
Having said that, I would say upgrading your platform (cpu/board/ram) first is the better route. Then later on get a decent mid-range gpu. 

?K dude not gonna argue whether or not it's valid to this conversation if you're going call it "rhetoric" lol
 
If your performance is going down that indicates instability. If you're testing for RAM settings you should be hitting it with AIDA64's Memory stress test for at least an hour, MemTest, and Prime95 large FFTs for at least an hour to check for stability. Chances are if the settings are not stable it will fail AIDA64 or prime95 in under 10 minutes. If it's near stability but still unstable it could go upwards of 30 minutes before failing. 
 
When you get a better board that will probably help a lot. Are you increasing your VCCSA and VCCIO voltages? That can help with instability as well. Setting them both at 1.2V is a pretty standard starting point. On SA don't go above 1.35-1.4, and IO about the same. 

Raw speed doesn't necessarily mean better performance. Latency is the big key, the lower the timings, the better your latency. 3200MHz RAM with very tight timings like CL12 or 1N/1T will actually perform the same or better than RAM running at 4000MHz CL18 in cpu intensive applications. 
 
And what I said was, for maximum RAM overclocking potential, you would want all 4 DIMMs populated on the AORUS Master. It is still very good at RAM OCing with only 2 DIMMs in. 
 
People telling you that "higher speed plainly equals more bandwidth" either haven't tested extensively enough to realize that statement is false, or they're just believing what they randomly read on the internet somewhere. It's just more complicated than that. 
 
Linus even has videos on this along with other YouTube tech channels. Raw speed isn't everything if you're sacrificing low timings as a result. 
 
Here's an article that talks about it further: https://www.crucial.com/usa/en/memory-performance-speed-latency

All you need to do is enable DOCP or XMP, whichever one is listed in bios for RAM settings, and that will run the RAM at its programmed specs. If you had gotten 3000 or 3200 there is a higher chance your board may not have "just worked" with those speeds, meaning manual tweaking would've been required. 

... The processor does 4.7 boost on all cores stock. Just saying. ???
 
Getting to 5GHz isn't that hard as long as you have adequate cooling. Some chips need more voltage than others. 

Pretty sure the AC series dating back at least 3 years have been cpu hogs up to at least 8 threads. It's how they programmed the games. Your CPU is definitely the problem. 

Z399 & other X599 parts are surely in the pipeline, I wouldn't dump a high end X299 system right now, especially for the overpriced ridiculousness that is the W3175X.  I would wait with all the big parts that are on the horizon in the next 6 months.

They aren't giving any market share away by choice, Intel has been having supply issues due to yields from their own Fabs as well as contracted ones, they've had stock issues with almost every line of processors they have recently, all the way up to the high end, they certainly wouldn't prioritize low end low yield parts, that makes no business or monetary sense.
 
Hence, they're hard to find.

Stress test for aida64 is merely a good starting point. Do not trust it as an overall stress test.
 
What I mean by that is: if it fails an hour of AIDA64 something is clearly unstable, but if it passes, that doesn't mean you're totally stable.
 
For reference, I've had settings pass an hour of combined CPU FPU Memory stress testing on AIDA64 and fail elsewhere. I've also had a RAM OC pass an hour of Memtest only to fail AIDA64's in 10 minutes. So a good mix of tests, along with just plain gaming on it and seeing if it freezes is highly recommended when testing an overclock for stability. I personally still use Prime95 26.6 as well even though some people swear against it these days as you see "unrealistic Temps" in it. 

I would contact MSI support directly and see what they advise if someone on here can't help you. Are you running any other RGB/lighting software? A lot of that software has conflicts with each other. Very odd that the MSI app won't even install though. 
 
Not familiar with MSI hardware or software enough to offer much more.