Zando Bob

Even an RTX 3080 is only held back 6% or so on PCIe 3.0 x8 (same bandwidth as PCIe 2.0 x16), so just the best GPU you can get your hands on tbh. sauce: https://www.legitreviews.com/nvidia-geforce-rtx-3080-fe-video-card-review-ampere_221996/3

You may have glasses to see really well but you still missed the point. Real estate. Aka scaling. For example, my 13" MacBook Pro with a 2560x1080 screen gives me 4 scaling options (1024x640, 1280x800, 1440x900, and 1680x1050): Important note: these do all look crystal clear on my end. But I very much notice the amout of information I can see on screen, and I can still read text fine at the highest setting, so I keep it on that as to have more usable screen real estate.

Probably There's still a lot of 17" laptops floating around, plus even on 13" if you have good enough eyesight to handle smaller scaling, higher resolutions mean more usable real estate (sure something as high as 4K isn't practical, but it is not a statement of fact that it is indiscernible) There are many mobile GPUs that perform far above the level of an APU, even a 2060 max-Q (common GPU in many gaming laptops I've seen) is far past what even Ryzen's good APUs can do Uh yeah... prebuilt desktops? With desktop GPUs in them? As you yourself said, all opinion. Graphi

Thermal paste isn't conductive, so I highly doubt you cooked anything. But because it's non-conductive, if there is any left on the CPU pins or mobo contacts, it could be preventing those from making contact, which would be why the mobo powers on but the PC doesn't do anything. Did you double and triple check that none got into the tiny PGA socket holes?

It's an ASRock B450M Pro4-F vs an MSI B450M PRO-M2 MAX.

^^^ Windows 10 can do it out of the box: https://account.microsoft.com/family/about. Specifically the app/game stuff as well: https://support.microsoft.com/en-us/account-billing/set-app-and-game-limits-across-your-kids-devices-a45e2d2d-4b55-a320-c8e5-daf610447a05.

Yep, but you'll need a different motherboard. I assume you double checked Supermicro's documentation, but those aren't funky chips so they should work fine. The chips are mislabeled as Socket 2011, but according to Intel ARK they are indeed 2011-3 Xeons. Not very high clockspeeds but if you just need to throw cores at something, should be solid. IDK what DDR4 ECC prices are, but usually once datacenters start updating at all, a bunch of used stuff gets dumped on the market and prices drop (tis why DDR3 ECC is so cheap).

Yeah that's the problem, you can never tell if a company decide to skate by on the good reputation of the last gen boards and skimp this time, pretty annoying. I think it depends on stock, retailers typically leave prices the same or hike em massively on new old stock, I don't know exactly why.

I don't know about the OC performance, but any decent OC board should handle a hexacore fine, considering they build them to (usually) be able to handle the higher SKU 8/10 core chips. You'd need to check the compatability list for CPU support, for PCIe support all generations are backwards/forwards compatible (the only difference is bandwidth, you're going from higher>lower instead of vice versa so you'll have no issues). They don't actually have the CPU support list they refer you to up lmao, but on the specification tab it does say the board supports 10th/11th gen chips:

It "should" be able to, depends on whether they routed a proper DisplayPort out through the TB3 hub. I'd assume they did at least one so it should be able to run at the very least one high refresh external display.

Star Citizen is more dependent on servers than your rig, I have an i7 6950X at 4.1Ghz with 32GB quad channel RAM, a 2060 Super, and am running the game off a 970 Evo SSD. I get 40 or so fps average, drops into the 30s or below in big city areas like Microtech and such, it'll occasionally sit in the 60s when flying around in mostly empty space. People with similar rigs often get worse or better fps than me, it's kind of the luck of the draw. They do keep up a dashboard of all the public telemetry data: https://robertsspaceindustries.com/telemetry. As you can see performance seems to vary wildly

Looks like an InWin Infinity mirror, this one is ROG certified so it's probably the one: https://www.in-win.com/en/gaming-chassis/805i/

i9s are all higher core count chips, so yeah that'd be why they get better fps with the same GPU. Your CPU is choking in titles like GTA V that are hungry for cores, and your GPU is now fast enough that you'll notice that.

That just reinstalls from the same stock image HP preloaded. He needs to make an install USB with a fresh image (Microsoft's own tool will let you download a new one, then use their tool or Rufus to build the install USB) and do a clean install from that.

Has he tried reinstalling windows? I usually find HP's hardware is fine (use a lot of their hardware at work), but whatever they do to the standard Windows 10 install when they tweak it utterly fucks reliability. It's gotten better in the past year or so but a clean reinstall from Microsoft themselves is usually the best for stability.

^^^ 100%. Sounds like the CPU is choking a bit, GTA V is a very CPU hungry title AFAIK. Is your friend running the same CPU, or does he have a better one? Game performance is mostly a combo of CPU/GPU, not just the GPU alone (both of them have to do certain calculations per frame).

I very much assume it'd be PCIe 3.0 considering that's what the CPU runs.

Filled out, pretty solid survey. My main beef with current cases is poor EATX support for really wide boards, or ones like EVGA's that put the 24 pin laid flat, forcing me to run the 24-pin power through HDD mounts, or else hop up to a full tower. Which tbh, there's also a poor selection of just decent full towers. I'd go for a Redux P12 (or up to the A12x25 if you don't mind the colors). iPPCs drone a bit in my experience, my Reduxes didn't (I have 6 NF-F12 iPPC 2000s and IIRC 5 or so NF-A14 iPPC 3000 PWMs in various towers, have used the Redux P14s and P12 as well).

Try HWiNFO64, much better tool than HWMonitor, it's usually more accurate. Shows a lot more useful numbers than both Ryzen Master and HWM as well.

For comparison here's my (bought December 2019) i5 based MacBook Pro 8GB RAM/256GB SSD. Mostly office work/dropbox/lil photoshop and some VMs (mostly off an external SSD now):

^^^ I'd assume this, once the experiment/series is over they usually just tear down the PCs and put them in storage, often reuse them for other builds.

It would go considerably worse than a Hackintosh. I believe with some work you can get macOS running at almost baremetal performance with some Linux VM tools, but I don't know if you can get iMessage, AirDrop, Handoff (the main features of the actual ecosystem, without all the interconnected bits iOS/iPadOS and macOS are just standalone operating systems) working near as easily as you can with a properly done hackintosh (if you want guaranteed compatibility you have to tailor your parts selection around known working combos, not what represents the best value). That and you're still working wi

Did you disable Origin In-Game from the launcher's settings? Also worthy trying DX11 and DX12 to see which works best, as @Sam Flynn suggested.

1.35v is the Intel recommended max vCore, community-based testing says 1.40-1.45v is fine as a daily for 32nm 1366 chips. 2011 chips will take similar voltages without issue, they're on the excellent (clocks/voltage wise) Sandy/Ivy arches. 2011v3 chips are much better off at 1.35v (especially for Broadwell-E as it's first gen 14nm and doesn't take voltage well), though I've seen people claim they had no issues running 1.4v through Haswell-E. 2066 I don't have much experience with, AFAIK it's better to stick around 1.35v for heat reasons but they should be fine up to 1.4v? But they're also mesh

Should try and get your hands on a 32nm Westmere-EP chip (X5675s are usually the best general bin/$), pretty rare but some can do 5.0 around 1.6v. Not healthy for the chip ofc but if you're shooting for clocks, 32nm has the best general bins. Though golden 45nm chips can do stupid numbers IIRC, one of the peeps here has ran through a bunch of 1366 chips and has good bin info, I forget their username though. 2.0v will probably still be death eventually even for a 45nm chip, but IIRC Intel's own spec said they could take 1.5v (vs the 1.35v for 32nm and everything newer) so they do h