Zando_

Definitely. It will not run some modern titles at all due to not supporting full DX12 features, the 3GB VRAM chokes hard in some newer titles, and they draw insane wattage (my GTX 780 Classifieds pull 350W). Driver support - outside critical security updates - is also dead, so no optimizations for new games, again making it an even worse option. They're cool retro cards but I'd really advise finding something newer if you want a decent do-it-all PC.

+1 for TrueNAS, though be sure you're using the same capacity drives (not sure what medley of old drives you're using), and be sure that the PCIe SATA card you use plays nice with ZFS.

What fans? Your AIO fans should be set based off coolant temp so they don't do that, if they're set to run off the CPU temp then they will ramp up and down obnoxiously. Unless power is expensive for you then the PC draw isn't really an issue. Though PCs are incredibly good at blasting that wattage out of the case as heat, so I do find with my PC that it will fight my AC on warm days, and that could measurably bump the power bill given the wattage those units draw. As far as PC power itself goes, I only ever noticed a large power bill spike when running ~400W of hardware 24/7 for a full month during a folding@home event a few years back. Those happen during the winter months so it wasn't fighting my AC (in fact it actually helped my heater instead), so I do wonder what the numbers would have been like if it were summer.

Pretty much, unless you're worried about the actual power draw itself as Schnoz noted. That'd be the only real reason to up/sidegrade IMO.

13900K adds 8 more E-cores and... that's about it. It's clocked slightly higher, and there's a bit more cache but that's likely there due to the higher core count. You'd likely see 0 benefit in games, in allcore workloads it would be quicker due to the 8 more E-cores (slower than P-cores but still, 8 cores is 8 cores). Every review of them I've seen cools the chip with a 280mm or 360mm AIO just fine. You can add a contact frame if you want, AFAIK they're like $20, not a big deal.

Likely not. 7800X3D will be significantly slower in allcore render loads, about the same (within literally a couple fps) in most games, and then actually reasonably faster in games that leverage the X3D cache properly. A 7950X3D would be a better fit for what you're doing, but also more of a side-grade than an upgrade.

You can set up a fan curve in the BIOS, no need for a dedicated fan controller.

Don't ever give people physical access to your data if you're worried about it. Get a second external HDD for sharing.

Furmark is a power virus (workload designed purely for maximum power draw), it can kill cards with inadequate cooling/VRMs to handle the wattage draw, so OEMs made the newer cards able to recognize this sort of load and automatically throttle the card. Use Unigine Superposition on the 1080p Extreme preset, it will pull similar power but while actually rendering a game scene, so the card won't throttle itself down. Much better for troubleshooting purposes. That's definitely wack, minecraft should run on a moldy potato left in the sun. GPU drivers are up to date, and you're using the display outputs on the GPU itself, not the motherboard? Install HWiNFO64 to keep an eye on CPU/GPU clocks and temps as well, see if anything looks strange there.

BIOS up to date? DDR5 has had some teething issues AFAIK, important to update the BIOS as it will usually include fixes for that sort of thing.

I had Xfinity (Comcast), renting their stupid box was $10/mo, but it folded in the cost of removing the 1TB data cap ($15/mo) so it was actually cheaper than just getting my own kit.

I have had that once so far, I cannot rip my Solo (the star wars one) Blu-Ray, it just returns an error. I got it for $5 from Dollar General tho so no biggie.

The MakeMKV forums has a flash guide with a list of compatible drives: https://forum.makemkv.com/forum/viewtopic.php?f=16&t=19634. I use the LG WH14NS40. You just pop in your disc, tell MakeMKV to take a gander at it, select what you want it to rip. Usually the main file is listed twice, and then there's some smaller files if there's additional features on the drive. Though on stuff like my Kingdom of Heaven Blu-Ray there was the theatrical release and the directors cut, I selected both based off the size (director's cut being longer, it's slightly larger file size) and ripped those.

Ah nice, I didn't know they gave out a code. I didn't mind the $30 to support 'em anyways though.

If your case has a 5.25" slot then it's a $90 initial outlay ($60 for a blu-ray drive and $30 for a copy of makeMKV) + ~15 minutes to flash the firmware, then you're good to rip any blurays or DVDs you acquire. Not too terrible.

Looks like there's neat pliers made specifically for grabbing various networking connectors from packed switches: https://www.fs.com/products/105341.html. Looks like it should work for grabbing SFPs as well?

No. The CPU cores are fast, the GPU cores are fast, the built-in accelerators for encode and decode are fast, the memory is fast, all those parts are packaged together so inter-communication is fast, and Apple pours a massive amount of work into optimizing their OS and software for these chips, as well as building a fast translation layer (Rosetta 2) for apps that are not built for the ARM architecture. Probably some other bits I missed, the point being that there is a lot that goes into how fast the M chips are. ^^^ My 7980XE runs quad-channel DDR4, giving me much better memory bandwidth than a regular dual-channel DDR4 PC. That doesn't magically outweigh Skylake-X cores being slow, nor does it magically fix the mesh architecture being high latency (vs Intel's traditional ringbus arch). There are probably certain tasks (that I don't run, I just play games) where memory bandwidth is the main bottleneck, and in those it would do better than something with less. Has been. Properly stacked HBM2 moves 1TB/s (my RVII GPU had that). HBM2e and HBM3 are even faster. But it's only a benefit to a certain point (as YoungBlade was saying). Fast memory just means you can move things in and out of it quickly, you still have to have something on the receiving end capable of processing fast enough to take advantage of that increased bandwidth to see any large difference.

I would certainly hope that the EA Sports WRC game, "the all-new official videogame of the FIA World Rally Championship" would have rally in it. From the steam page, it seems to:

It does: https://help.ea.com/au/help/wrc/wrc-cross-play/. Uses Codemaster's physics (they make the DiRT series, very good rally games) too so it should be pretty good.

Games will load 1-8+ cores to 80-100%, most tend towards the lower end so total CPU usage will appear low, but the chip is boosting to full clocks and games will see gains from the cores running faster. That said, typically a 200Mhz+ boost is what's noticeable to the end user without a bunch of careful benchmarking and comparing single digit fps differences, so I'm going to echo @NikolakiH on that being a waste of power for an imperceptible boost. From what I have read you do not want to run 1.45v or higher as a constant voltage on these Ryzen chips. I believe it was due to current draw or something, when running stock they do pull that voltage or higher, but usually not under heavy load (voltages will drop back down when running auto). I don't have much experience overclocking anything past Zen+ (where the limit was 1.4v or so, if memory serves I ran 1.425v early on but that was later discovered to degrade the Zen/Zen+ chips slightly) so I can't personally confirm that. In general, overclocking most Ryzen chips seems useless unless you need the slight gains in allcore performance (as noted, your games are not loading all cores), as PBO will typically reach the same boost clocks on the cores games are actually using, and can often do so with less voltage than stock, keeping the chip cooler and safer.

M.2 slot is enabled in the BIOS? Not all boards enable them by default.

Read should be fine AFAIK, but yes writes will be hit when the drive is overly full. I don't recall if SSDs get grumpy about multiple partitions too, I've never split up drives that way so I can't speak from experience (I use separate physical drives as I've picked up a bunch over the years for cheap). From googling, looks like partitions should be a non-issue. It's likely just the low free space that's bogging the drive down. BIOS fast boot skips initializing some devices until the OS is up, so since you've confirmed the SSD is what's slowing it down then it likely wouldn't make much of a difference. It seems to be purely down to how full your drive is, the only real solution would be to keep more free space. The rule of thumb I've always seen is 10% free, so ~50GB on a 500GB drive. Windows itself will complain if it goes past that, I use 1TB drives so whenever they drop below 100GB free, Windows explorer marks them in red.

^^^ Also the United States exists, where many ISPs still have 1TB/mo data caps you must pay extra to remove. And where rural areas often have very slow internet (when I lived in Collinwood TN circa 2015 we had 1.5mbps down in the middle of the night when no one else in the county was on, and ~400-500kbps during the day). So downloading updates sucks for data cap reasons, and just logging in can take a bit - especially if the connection is meh quality and it has to retry - with slow internet speeds.

Startup programs are the bloatware that would effect boot times, thus why I asked about them. What optimizations? If you cut enough out of Windows you start to break it, not make it better. I don't. I leave the OS itself alone and just disable all the startup programs I don't want in Task Manager. Boot times are slow as I am on HEDT and have the fast boot stuff disabled, but they are consistent. Shouldn't effect things, I've had $25 SATA SSDs boot as fast or faster than my nice samsung NVMe drives. The OS loads a ton of tiny files, so a drive's sustained read/write for large files (the number they stick in the marketing as it looks the best) isn't really relevant. How full is the drive? How much slower have your boot times gotten, and are they consistently slower or is it random? And have you looked around your BIOS for fast boot/wake related settings and tried enabling/disabling them to see if it helps consistency?

Not at all! Temp is very dependent on the instruction set used, and what's actually run with those instructions (see Prime95 smallFFT vs larger FFT sizes resulting in different peak CPU temps). AVX instructions are very hot to run, enough so that some chips have an AVX offset to downclock the chip when running those instructions for thermal/current draw reasons. Basically the type of load matters, not just the % number a monitoring utility spits out.