Nexxen wrote on 2026-03-28, 02:40:

darry wrote on 2026-03-28, 01:39:

Bought 16GB of DDR4 3200 for 169.99 CAN$ before taxes, for a friend's no longer delayable PC upgrade. Hint: Old CPU is Xeon e5450 on a P35 board.

Not so long ago it would have been 69,99.
Well, P35... it was really about time 😀

69.99$ ? Ipaid about that for 2x16GB DDR4 3200 a little while back.

The Xeon build is on its last legs. Only the PSU, GPU and SSDs/HDDs will get reused. It had a good run. I would have switched it to NVME for the SSDs, but with the prices these days, nope, unless it is really absolutely necessary.

I upscaled my iCarly DVDs to 8K resolution at 120 fps, then encoded the resulting video frames losslessly. Oh, and I remixed the audio to Dolby Atmos...after upsampling it to 32 bits and 384 KHz with a little help from X-Fi Cyrstalizer.

The file size per episode is now approximately 30 times greater than the original DVDs, the WiFi is barely fast enough to sustain the required data rate, and the 3D glasses need to be recharged every four episodes, but my word. This is the definitive way to watch iCarly. The show's sharp wit, masterful storytelling, and breathtaking cinematography have all--I feel--been amplified to a new level of glorificnessness that truly threatens to alter one's psyche and mystify the dog.

Finishing cleaning the water loop on my main rig. Noticed a month or so ago that the water was getting nasty and that temps seemed a bit high / the fans were ramping up way more than seemed normal.

Started last night.

Also replaced 2 of the 3 fans on my video card last night as they were going out. Ordered fans probably 2.5 months ago and just hadn't had the time to mess with it till now.

I tried to fix the WiFi on one of my desktops. Since it's a desktop, the best fix is to wire the house for Ethernet, but I don't want to live in this town for much longer. This is going to turn into a long Linux rant.

The problem is that EndeavourOS has random disconnects, and dmesg is flooded with "missed beacons" warnings. However, Ubuntu 24.04 and the latest Fedora running live don't have any problems. It has been that way for a few months, but I haven't tried to fix it until the weather started warming up and the new "upgrade" Ryzen is heating the room up too much (with integrated graphics). I looked around, and it seems like the only troubleshooting step for Intel WiFi is to set an option called 11n_disable, which is extremely poorly documented. The only documentation I can find is a single-line comment in the source code and a couple of forum posts from people who figured its function out on their own. Based on the name, the implication is that it will force my nice modern WiFi card to 802.11g, which is not a legitimate solution at all. If the driver is that broken, just say that. Even if you ignore the fact that a driver for modern hardware is potentially so broken that it can't handle a 20-year-old networking standard or any newer standards, why does this only happen in one distro and not others?

Anyway, I tried setting 11n_disable to 8, and it didn't help. I don't want to try 1 (the choices are 0, 1, 2, and 8 for some reason) since that's the choice that fully disables the ability to connect using any standard newer than 802.11g. I have no idea what's in EndeavourOS that's different and triggering this, and all troubleshooting threads I've found die after setting that option.

Also, I'm trying to get a Windows 11 VM to update to 24H2 (even though 25H2 is actively being rolled out). My VM ran out of disk space, and failing to install due to insufficient disk space causes Windows Update to redownload the entire update again, fail again, and download again in an infinite loop on my DSL Internet. Then when I freed up enough space to install, it gave me error 0x8024a22a and tried to download again. Why does everyone assume that you have infinitely fast Internet? Even 1Gbps isn't good enough for them. I've mentioned how bloated AMD/Xilinx dev tools are (such as the power estimator turning from an Excel spreadsheet into a full application that's a 9GB download), but that VM has MPLAB installed, and its folder in Program Files is 10GB just to program little microcontrollers. I'm going to uninstall that since I have no desire to do any kind of programming anymore (and if I did, there are Linux tools that don't need the VM).

These days I tried to fix my i3-12100 config with its Intel Arc 310 ECO using a different PWM fan controller instead of the video card one. Basically the known problems of this card is that the heatsink is seriously too thin and small to dissipate a 2Ghz GPU and the ridicolous fan used has to work so much to keep it cool enough making huge noise and so variable it'd easily drive any user crazy.
So I used a Corsair Commander XT to drive the GPU fan and deciding a better fan curve cause it's not possible to do that on Linux drivers with video card driver itself. Of course it was not enough for the noise but more stable speed for sure. Then I downclocked the video card @ 50% and now the card should ask for 20W instead of 31W before and now the fan seems at least acceptable.

Compiled Pale Moon for the second time from source on Windows...at least I hope it built, will see tomorrow. No idea what Frankenstein did actually come out, was fighting with the build system to use the old MS build tools that still produce Win7 compatible binaries and since Git submodules are alien to me along with the fact that I didn't change from the master branch...

I engaged in the time-old tradition of contacting a major equipment maker's support line, getting the runaround from lv1 support, then fruitlessly repeating the same troubleshooting I already did with their factory service manual, before finally being blessed with a factory service rep appointment to replace a toast component.

It sure would be nice to be able not waste a bunch of time when these things happen. I understand that the lengthy process exists to catch the most usual modes of failure, but really. Asking me to access the device's web interface when the nvram board is bad (which is what stores the configuration data of the device, including its ip address), seems a bit much.

How do I know its actually bad, actually bad?

The device does not respond to magic buttons to enter service/diagnostic mode. The device does not give a link light when connected to a switch. The USB port is physically damaged.

Need I go on?

Xerox support felt I did. 😁