Congratulations :)

Thank you! That makes much more sense.

ELI5…?

I’ve got 3D pipes running on my spare Win10 machine :) fills me with nostalgia every time I see it, even still

I started with C++ too, and then ended up finding a job writing firmware pretty much all in C. There really hasn’t been anything we’ve run into that’s made us consider switching to C++; being able to (and needing to) have complete control over your memory means you can do some pretty fancy stuff with the tiny amounts of memory on our ASICs.

We’ve been eyeballing switching to rust a little bit, but really only for other applications; the root of our main code base is over 25 years old at this point and a rewrite would take a Herculean effort.

Meh? I write pretty much exclusively in C and honestly I still like C++ better, and wouldn’t mind switching to Rust either

the fitness gram pacer test is a multistage…

I’ll let you guess if that was my most or my least favorite

Huh, that’s certainly interesting! The hacky solution ended up having to do with power states which is kinda annoying - I have to set the GPU to use max power state because if it goes into the min state and then I walk away for 5-10 mins, it drops out of the PCIe slot and I need to reboot. SSH still works but you can’t reattach it w/o a reboot. I’m running a PCIe gen 5 mobo though and I heard about some potential problems with that, so maybe that was related. Could also be the fact that I ran a Quadro RTX 4000 on the same system/OS for a year or so and didn’t want to do a full reinstall, so it probably had somewhat to do with leftover drivers and crap

I set up my 4070 TS (the brand new one) on Ubuntu 22.04 about two months ago and my god was it a pain in the ass. Took like two days to do and even after that it would still hit a screen freeze issue every thirty minutes that took another week to find a half-assed solution for…

Shitty k8s cluster/space heater?

…absolutely, positively, super false. I work in a sector where we’re constantly dealing with huge capacity enterprise SSDs - 15 and 30 terabytes at times. Always using RAID. It’s not even a question. Not only can you have controller malfunctions, but even though you’ve got what’s known as “over provisioning” on the SSDs, you still need to watch out for total disk failures!

Appreciate the response! After many, many hours of research, I came to the same conclusion. I tried a whole multitude of solutions that worked for others and none of them seemed to work - except for a weird hacky “solution” to just permanently set the power state of the GPU to max. Unfortunately, that means it consumes ~50 watts idle instead of the 5-10 it managed beforehand… but the fact that it fixed the system lockups made it worth it. I think the issue was something having to do with the GPU not properly waking up from lower power modes - so I super appreciate the advice :)

Off to a fairly rough start, unfortunately :/

Spent seven hours today trying and failing to get docker to work with our Jenkins deployment at work, and on top of that, my brand new GPU keeps “falling off the bus” (Ubuntu, 4070 Ti Super, randomly screen freezes and need a reboot to fix - but PC still runs so I can SSH in & check dmesg and whatnot). Sometimes it’s every 12 hours or so, or even more, but sometimes (today, for instance), it feels like it’s every ten minutes. Which … sucks.

Side note… if anybody knows how the heck to fix a GPU falling off the bus… please let me know, lol. It only happens when I’m using the PC (as in, if it’s on but the mouse ain’t moving, it doesn’t seem to happen), and I’m running the latest & greatest NVIDIA 550 drivers. Ubuntu 22.04. Reseated GPU, running a 1000W EVGA PSU and the Kill-a-watt attached to it never goes above 450 or so. And the crash never seems to happen when it’s under a huge amount of load, like doing AI stuff… only ever seems to happen when I’m browsing files and such. Anyone ever run into this before?? All of the google answers seem to say it’s a bad PSU or similar, but the PSU has been working just fine & dandy in other PCs, and this system wasn’t doing this at all with my old NVIDIA GPU (swapped last week)…

You hit the nail on the head with that last sentence - too big to give a f**k. Gotta love the big corpos.

lol, I wish. Luckily I have plenty of computing power at home already! But unfortunately we’re definitely not allowed to take anything - have asked a number of times :/ has nothing to do with storage and everything to do with capital BS

At least in the enterprise sector, you’re absolutely right. My company’s already got a massive list of all of the PCs that need to be discarded due to Win10 EOL. It freakin sucks because they’re very powerful PCs, but the damn lack of a TPM2.0 chip means they are basically garbage for our uses. And they don’t let employees take anything home :/ what a waste of

On the Pentalobe screw front, albeit somewhat random, I do know that all Samsung SATA & SAS 2.5” SSDs use Pentalobe screws to hold them together. Unsure if there are other Samsung products that use them as well but I deal with their drives on a weekly basis.

Even if you were able to make your own PCB and somehow solder everything onto it, one of the things that makes complex boards like motherboards so tough to make is signal path lengths. Ever see how some of the traces on motherboards are squiggly and take up more space than the straight ones? That isn’t just for fun - all of the traces have to be incredibly specific lengths for a whole number of reasons, including signal timing and interference with other traces.

Yeah… I realized that after I typed out my comment. Especially w/ F1 and all

Red Bull should stick to energy drinks…