suchamoneypit

Those are pretty new GPUs, I probably wouldn't tear those things open just to refresh thermal pads and paste. Probably a good chance you cause more problems by doing a full tear down.

AFAIK there is no correlation with the monitor. At worst, you should just be losing the display output, not crashing the entire system. You probably want to focus on troubleshooting the ryzen master issue. Checkout this reddit post. Edit registry at your own risk, I've never had this problem myself, but there are people saying this stopped their crashes. Someone as recent as 15 days ago saying this solved their crashes.

Sounds like a power delivery issue to me. As soon as high power draw happens, you're shutting down. First I'd simply try another outlet, although if there were issues you'd probably be tripping a circuit breaker. I'd try and get a second PSU to test out. You can likely buy one from a local store with a 30 day return policy. Something else that might make it suddenly shut off is severe thermal throttling. You should monitor CPU/GPU temps to make sure something isn't running super spicy.

I got a server, for future upgrades, I'm limited on PCI-E slots. I'm wondering what kind of penalties I might see if any If I run the GPU on x2 PCI-E 3.0 lanes. For gaming yeah, I see obvious concerns. However this GPU is used exclusively for (rare) video transcoding for a PLEX/Emby/Jellyfin server. Is transcoding video going to saturate those lanes enough to cause major issues? The CPU with be a 5950x soon once I get it installed so Im also thinking in the rare cases I need to transcode, the CPU can just suck it up and get it done. But I do already have GPU. Anyone know how much bandwidth transcoding video will use?

yes it looks fine to me

Are you able to get the GPU connected without the riser cable? Definitely something to look at. You're having display issues and this is something extra between your GPU and display output. Or maybe start by making sure the riser cable is firmly installed. Reseat it. Remove it and firmly install. I don't think you should have any concern about the power cables to the GPU if they are firmly connected. If you really can't get it working, using those other parts to do this is your best bet. Confirm it's working out of case, you can easily test parts. I know the pain with a SFF build, it sucks to undo everything, but If you get at this point it's the only way to figure out what is the problem.

You can install drivers one version back and test with that. Do you still "stutter" with V-sync enabled? Is this only happening in a single game or all?

This is extremely vague. How did you confirm RAM is the problem? Did you run diagnostics with both sticks installed? What diagnostic tool did you use? What are your system specs? No one can lookup any info specific to your system like what error lights are with zero info on parts. Blue screens give you error codes. What are those codes? They usually hint at what is the problem. This randomly started happening without any recent hardware changes?

Scores in some synthetic benchmarks: Steam Deck Cinebench multithreaded (CPU power) - 4436 Cinebench single thread (CPU power) - 977 3Dmark Firestrike (GPU power) - 4856 Ryzen 3500u Cinebench multithreaded (CPU power) - 3544 Cinebench single thread (CPU power) - 876 3Dmark Firestrike (GPU power) - 2559

Your CPU power cable, and GPU power cable are connected and fully seated on the motherboard? Especially the fully seated part? no LEDs inidcators on the motherboard? Like any red ones? I don't mean to doubt you but I've been told by friends with 100% confidence that wasn't it, only to find a partially unseated CPU power cable. trying clearing CMOS and booting up. Don't wipe your storage drives, they don't even come into play until after you get to BIOS, and you don't seem to be getting to BIOS. I'm on x570, it doesn't take 10 minutes to post, more like 10 seconds. I have never bothered with RAM compatibility lists and never had any issues. I actually had great success with the gigabyte BIOS update utility in the past, but it also seems like your motherboard ships with ryzen 5000 compatibility so you shouldn't be required to do this. Also, I'm running the same PSU but with higher power GPU so power should not be an issue here. Have you tried a different monitor, display cable, and/or different display outputs on GPU?

MSI recommends 400w for that GPU, both your systems are 300-320W it seems. You either have a dead GPU or you don't got enough power. Do you have a system with higher wattage PSU to confirm with? You need to install graphics drivers for this when installed as well.

I'm curious if you look at task manager when doing games or running cinebench, is your CPU actually reaching the proper clockrate? It should be boosting around 4.4ghz. When my BIOS was wrong for my 5800x3d I saw performance issues and my clockrate locked at a lower than normal clockrate.

The new cooler they are mostly talking about gains on Ryzen processors which you don't have. Even then, its on specific ones, and they say its only up to a 3 degree difference on CPUs that aren't de-lidded (that's you). That is an extremely small difference especially considering the price you'll pay. If you want to change the looks do it, but from a performance standpoint its not really a wise spending of money.

could be something from manufacturing or just from it being pressed together at those spots. Plastic really won't leave a burned surface appearance, and if it did, it would come along with melted plastic and warping, neither of which I can see in that video

If you aren't getting to BIOS then the system either isn't posting (getting to BIOS) or your iGPU isn't working. Simply disabling a network card should have no bearing on whether the system boots to windows or not. You'd just not have an internet connection. Do you have another display cable to try? Do you have another display port on your motherboard to try? Do you have a dedicated GPU to try and see if you're getting display output? Do you have another monitor to try? if none of the above works and you already tested PSU, really the only things between you and the BIOS/windows is a faulty CPU or motherboard which is hard to diagnose without spare parts.

your video works for me and I dont see anything in it that looks like burn marks

You say you're on latest bios and chipset drivers, but have you updated your GPU drivers?

Personally I'd keep the Noctua cooler. Its a quality cooler, I don't think you'll see a significant performance differences if at all. You'll essentially be spending $150+ for a visual change. Noctua sells Chromax (black) fans you could swap in for a cheaper visual change. Of course spend how you choose. I used to always use liquid coolers but I grew out of it and love the reliability of a metal heatsink versus a liquid cooler with liquid and a pump. I've never gotten a call to replace a faulty air cooler from a friend, but I've replaced several liquid coolers. The fact that 7 years later you're still running great goes to show that. I've also grown to prefer the aesthetics of metal heatsink versus a plain liquid cooler pump head and pipes. When I was younger and getting into pc building though I had to have a sick liquid cooler so I get the feeling.

Thank you, that did make it show up! That is what I figured it was, I'm able to see my progress. Thanks!

It looks like JBAT1 is your Clear CMOS pins. Did you short these two using something like a screwdriver or paperclip? This should reset your bios to default and you should hopefully at least get to your BIOS. I always started with only shorting JBAT1, and then if I was desperate, both shorting JBAT1 and removing the battery.

Came back here after a several year hiatus, and see these new ranks. Cant seem to find any information on how they are unlocked or how you progress to the next tier.

I do understand for the media itself its certainly pointless, but yes VMs will be run. Perhaps even as far as a high performance gaming pc VM will be run, so having all that data on high speed SSDs is desired. Also, all data like the plex app itself, and all other dockers like game servers will be saved on the SSDs. I can also lazily do things like set aside a 200GB SSD chunk for a VM or something. For media downloads I'll have it hit the SSD cache first, but then it later transfers to the HDD array. I can send chunks of 400GB-1TB downloads at once, and sometimes it takes several days for this data to make its way over to the HDD array so having a buffer large enough to not have to care about that is nice. If I have something like a large slow torrent or download in general, a lot of space can get tied up for a few days while those finish especially if that space gets pre-allocated. Everything can hit the cache at first, then certain data like media gets moved to HDD arrays and stays there. But my VMs and dockers can all stay on high speed SSD. I could likely comfortably get away with 2TB of SSD without problems but the SSDs were on sale and its something I wanted to splurge on for a while so I went all in and got 3 drives so I could have 6TB of SSD to match my 6TB hard drives. a short description of unraid is you creates shares which are like folders of data, and you can choose how data is handled per share. You can have it only use the array (hard drives typically), have it only use Cache (typically SSDs, or have it use cache initially, but then transfer to the array (once the data is not in use). So for my media shares, I can have it hit cache then transfer to array, for my VM and server data, have it be cache only. Normally (when I was low on SSD space) I had media set to only use the array as large downloads tended to fill my SSD to capacity and then cause my VMs and servers to slow down.

My goal was a total of 6TB of cache, to match a single HDD. I got 3 Kingston NV2 2TB drives, they seem to has an endurance of 640TBW. That 6.5PB of endurance is pretty impressive but those are used drives so who knows the lifespan left, and its like quadruple the cost per TB. I don't expect to use the cache heavily at all but want it to be as fast as possible when it is used, which is why I went with triple nvme drives for cache. I got the NV2 drives for $84 a piece new. I'm very unlikely to ever come close to the 640TBW and the 3-6x speed improvement will be nice. ok, that makes sense. thank you. I'll probably stick with a single card as I think with the PCI-E slots saved I might be able to get a 2nd GPU in the system. My Networking is 1 gigabit to everything, but looking at 2.5GB options currently. I'm using unraid, and currently primarily use it as a media sever with occasional game hosting. I want to be able to run VMs well. With a 500GB SSD cache before between VMs and multiple 4k movie downloads my cache would max out, then fail to empty fast enough, causing everything to come to a standstill. I know I'm splurging on certain things.

Thanks, I missed that in the manual, so definitely makes sense why E5 isn't working. As for the speeds, it is important they all run at max speed. All 3 will be together in a shared cache. I can't have 1/3 drives slowing down the entire cache pool, especially since I bought them all together explicitly to be a max speed cache. With what both of you said, it seems like at least an HBA with 2 SAS ports is definitely something I need. I am ordering one. If I get one, I save a PCI-E slot and probably get better speeds.

thanks for all the info and links. That seems way better. So that first card, I take up one slot, and using those 1 SAS to 4 SATA cables, I could run 8 hard drives off this one PCIE slot right? I'm a little confused on the SAS expander part. Is it wired up like this? Both cards are just $40. Whether I buy 2 of the first card or 1 of each, this way I end up with the same 4 ports for drives regardless. At least using just the first card, this would at minimum let me drop down to 1 PCI-E slot in use instead of my current 2 because my motherboard has 6-8 SATA slots itself too. So I can do 8 HDDs on the card and 3 off my motherboard.