PineyCreek

Does CrystalDiskInfo or Samsung Magician show the same number? Moreso Samsung Magician, as that's your drive's main utility for firmware updates and diagnostics.

Are you installing drivers from Gigabyte's motherboard driver page, or are you trying drivers directly from Intel? Might be worth a shot installing the ones from Intel if you haven't tried. https://www.intel.com/content/www/us/en/download/19351/windows-10-and-windows-11-wi-fi-drivers-for-intel-wireless-adapters.html Also, try uninstalling the device (via device manager proper) and then rebooting and see if it returns to the same status after reboot.

I believe in the newer generations it's not abnormal for it to hit 90C in synthetic benchmark testing. At least, my 5800X does it. The key is that though Tdie hits 90C (i.e. the measurement IN the CPU, not close to it) whether or not it thermally throttles. At least based on HWINFO64, mine will run Cinebench, hit 90C on Tdie, but not thermally throttle. It hit its coded limit, but the cooling system is enough to prevent it from exceeding the limit, so it doesn't thermally throttle...at least that's my understanding. I have zero stability issues in gaming or synthetics, and people say the 5800X is a heater. Runs at 70s during gaming, never above. Maybe it's the same scenario? The same sensors page for HWINFO64 should show thermal throttling flags (Yes/No) further down the list of sensors I believe that you could check while running a synthetic benchmark. Either way, the 700s and up don't come with stock coolers for a reason, I would guess. If you have a poor airflow case then either get a CLC for the CPU and vent it out the top maybe (and it still won't be as efficient as getting better cold air coming in) or better do what @TylerD321suggested. Before any of that though see if you can get it to the latest BIOS without issue first if you're buying a new CPU, or at least have the ability to return an unopened CPU easily. No point in wasting money if your decision might change due to a non-zero (but quite low) chance of bricking the motherboard on a failed BIOS update. Also, read the notes to check if there's any bridge BIOS updates you need to hit in between your current and the most recent updates. https://rog.asus.com/motherboards/rog-strix/rog-strix-x370-f-gaming-model/helpdesk_bios/

CPU has some PCI-E lanes then more come from the chipset. The ones directly leading to the CPU are considered faster (lower latency?). On desktops that's usually the top x16 slot and the top M.2 NVMe slot I think...varies per CPU/chipset. ARK page for 12700H says 28 PCI-E lanes, CPU on PCI-E 4.0, Chipset on 3.0. M.2. SSD is on 4.0, x4 lanes. In the spec sheet their secondary storage is listed using the same NVMe options, so highly likely your assumption about 16 (GPU) + 4 + 4 (2 NVMe on 4.0) is correct. Probably safe to assume the dGPU is on x16? I can't find a full chipset diagram either though. iGPU isn't counted in CPU PCI-E lanes. https://support.hp.com/us-en/document/ish_6507941-6507994-16 Note I'm basing this entirely on the specs of the drives listed for both primary and secondary being 4.0 and Intel's ARK page saying the lanes coming from the associated chipset are 3.0. Perhaps the remaining x4 4.0 are going to the chipset as well.

I got a 5800X for $250 USD. It runs a fair bit hotter than the 1700 that was in there previously. Stock cooler is no bueno. 5600X has lower TDP and comes with a cooler. If you get a 5800X, maybe get a nice nice tower that fits like a Noctua or a CLC if you swing that way. Looking at the 5800X vs. the 5800X3D I didn't see an overly compelling reason to pay the extra, really even for gaming. I'm sure it makes a difference, but it wasn't worth the extra money for me personally. Have you checked while playing and experiencing one of those drops in performance if the thermal throttling flag is definite, i.e. like via HWINFO64 sensors?

You can set per app GPU settings in the standard nVidia Control Panel but GPU selection I've only seen on laptops with i and dGPUs in the same system. I would assume the same options would appear on a desktop with the same scenario, but I've not personally seen it.

Think that's an X-Fi PCI from 2005. They're selling for like 50-70 USD on Ebay. Lol @Eigenvektor @ the retro build comment as I look at the SB16 ISA card I have on my shelf.

Oh, the warranty's already gone. I checked and verified it was bought May 2017. The Celsius S36 came with a 5 year warranty. To its credit, it seems to be performing as well as it did when it was first installed. There's signs of permeation, sure, but for example, not enough to where you hear large amounts of sloshing, etc. when you move the case. I've found out that my view of Tmax on synthetic benchmarks is flawed. In the past few generations it's more common to ramp up high or hit max, and your performance is more based on how much cooling you have moreso. For example, during those synthetic benchmarks when the internal CPU temp was always hitting its max limit, it still wasn't thermally throttling (per HWINFO64 flags and also frequency observations).

Well...tried a few CPU-oriented games that barely touch the GPU...seems the 70s is just where Tctl/Tdie likes to sit. I can hit two core 4.8 (normal) or all core 4.4 GHz in the 70s...never goes to 90C on anything I've tested thus far...only on synthetic. Guess this is normal, though I'd still appreciate your thoughts. I'll note the sensor on the motherboard near the CPU behaves moderately close to the way it did when the 1700 did, running about the same at idle and only very slightly more than the 1700 did under load, as a point of comparison.

Ok, here I am post-swap with a few observations that I wanted to bounce off forum members. I went ahead and swapped from 1700 to 5800X on the Asrock X370 Taichi board, everything went fine. Now, before I make the below observations, I know the 5800X is a volcano of a CPU. Also, all temperatures stated/observed are Tctl, or the sensor on the CPU, not the sensor on the mainboard close to the CPU. What I noticed is that the CPU is perfectly stable, but it won't boost to the rated 4.8GHz in CPU benchmarks due to it hitting Tctl of 90C. I'm not too broken up about it because it never goes above high 70s in actual gaming or gaming benchmarks, and the CPU performance overall still feels about 50% better and the clocks are an easy increase of 25%. I also noticed that the S36 lists itself as a DC AIO, yet has a PWM setting (selection ring) and a 4 pin connector. It behaves as a DC pump/AIO as when you change it to PWM, everything maxes out. Now, one would think that maxing out the fans and pump to where everything's easily 5x more audible would make a difference in thermal performance or readings, or perhaps prevent it from hitting Tctl of 90C. It does not. In fact, it makes no discernable difference whatsoever, so now I'm questioning if the age of the pump and/or fluid is catching up to me, hence impeding the thermal transfer between cold plate and radiator, or if this is just the way it is with a 5800X. Right now I have the header set to PWM but the S36 is set to Auto with its selection ring. Previously the header was set to DC and the S36 set to Auto. I see no difference in temperatures, and the pump and fans are quiet. Setting both Though, even on the quieter settings, I still see the boost in performance, I never hit Tctl/90C in gaming, and at current I can be perfectly happy as is. What conclusions should I draw from this, and what if anything might I proactively do? Or perhaps the pump was weak from the start as @Motifator mentioned and now it's just more obvious with a beefier CPU? It could also be that it's improperly seated and/or at an angle. When in PWM mode on both the AIO selector ring and the header, it scales differently (as expected) but again, even at 100%, it can't prevent Tctl/90C from being reached with synthetic benchmarking, but again, I also never hit that temperature in any use case (thus far). Maybe I should find a almost 100% CPU-dependent game to test with.

Completed the CPU swap. Seems to be doing fine. Doing gentle work to allow it to heat soak. Amusing observation in CPU-Z considering we talked about PCI-E earlier. CPU-Z under mainboard will pull its Bus data as PCI-E 4.0 with the CPU installed (under the mainboard tab, not CPU). It's definitely pulling that from the CPU, not the mainboard/chipset. It pulls the GPU stats correctly as PCI-E 3.0.

Sucks, but at least you have an answer.

Made sure to connect the extra 4 (or 8) pin EPS connector on the mainboard?

How did you create the boot media (the bootable USB key)? Also, what's the exact model number of the laptop from the label (likely on the underside)?

From past experience when I had coax Internet, I also had up channels periodically not responding. The ISP had to send someone to swap out filters on the line or something at the utility pole outside. Cheap coax splitters can also cause some issues as well. I'd stick with what they gave you (the new adapter) and monitor it to see if it reoccurs. If it does, there's another datapoint for the service personnel and technical support. You should also be able to look at the labels on each adapter and see what the specifications are. If they're the same specification, then it's a subjective move (or experience) from the technician. Either way, don't argue with results since the up channels are working now...just see if the fix sticks.

Something important to note if you care, the 5500 is PCI-E 3.0, not 4.0.

Can't think of any scenario where a loss or surge of power would selectively damage one facet of a connection. More likely to be an unresponsive upload channel or throttling from the ISP or bandwidth limitations upstream (or a crap connection to begin with). What kind of incoming signal media (coax, DSL, fiber)? What model modem/router?

Well, probably not software then. PSU, connections to GPU, or GPU itself. I'd ask in the card manufacturer's forums if they have them and/or file a ticket with their support. Even if out of warranty, you could get their opinion (YMMV with that).

Could still be a crash related to power I suppose. I understand you already did a Windows reinstall (don't know when that was done mind you) but at any point after that did you try something like DDU to clean it up before reinstalling the latest driver?

Speaking from some minor experience, shipping to Brazil or India or China from NA tends to be amazingly bad in terms of dealing with Customs. Europe should be better than that. Also keep in mind thinks are also likely more tightly-controlled or :: finger-quotes :: optimized to save fuel with all the shortages. Shipping large amounts by country to then be shipped to individuals, sometimes they're done in bulk as well (aka customs/shipping broker waits until they have a pallet's worth going to that particular country to get a better rate, etc.).

Mmm...the only reason I entertain the idea is that the top x16 and M.2. are direct links to the CPU. Probably wishful thinking though.

Mmm...if you had/have the knowledge and/or tools you could test the output of the PCI-E connections from your PSU. I'm not sure how a GPU would react with highly fluctuating power on those connections though. PCI slot power is 75W I think. Trying a different PSU is cheaper than a new GPU anyway. You might ask the OEM (brand) of the GPU on their forums or search those forums to see if others with the same model have had similar experiences. Devices dropping offline and coming back though...you could also check event viewer for device and power events related to the GPU. Might be something there that gives a clue.

Background: I've got an Asrock X370 Taichi motherboard. It has been upgraded (BIOS) along the OEM's recommended BIOS path to the latest (and likely last) BIOS update that supports 5k series CPUs. A 1700 (non-X) is in the CPU socket). Previously I tried purchasing another kit (same make, model, part number, timings) of 2x8GB GSkill FlareX RAM CAS 14 to use in my other two slots. It failed miserably, would never get out of memory training, would never complete boot until I went back to my original two sticks. It did boot up just fine on those two new sticks in the original two slots. I sent the kit back (this was a couple years ago). Conclusion from trying previously: Likely memory controller limitations since so many people reported similar issues with Zen 1. Also a possibility, bad slots. Current scenario: I'm swapping in a 5800X soon. I hear in the latter generations the memory controller on CPU as well as AGESA have improved memory-handling immensely. Is it worth trying again with another of the same kit, or would I be better off considering the relative cheapness of RAM at the moment to just grab a 2 x 16 or 2 x 32GB kit and swap it in?

76C should be normal-ish load temp. Do you see fans not spinning when your GPU's under a heavy load that's consistent? If you override the fan curve to full blast all the time and play does it also crash? Might also check the PCI-E connectors to the card as well.

I apologize if this comes across as a highly idiotic question, but I've honestly never really put serious thought into it and have suddenly encountered something that brought it up...but if you have a motherboard with a chipset built around PCI-E 3.0 for example, and the 16X slot is linked to the CPU, and the mechanical interface (pinout) between generations is the same, if you upgrade CPUs that span PCI-E generations (ex. Zen 1 vs. Zen 3 (at least the ones that actually are 4.0)), you still don't get PCI-E 4.0 suddenly on the CPU-linked x16 slot and/or NVMe M.2 slot, even in terms of speeds, correct? I'm guessing no. I know there are other changes between generations, but since there's mechanical compatibility, that thought just passed my brain. The chipset obviously isn't changing unless the motherboard is swapped. Also, yes, I know that there's very little that's going to push the x16 3.0 bus. Scaling on the 4090 shows maybe a 3% drops between generations.