Aaron44126

Figured out the whole priority/pin system in the apt package manager, so I can add the Ubuntu 23.04 repos but only have it pull down select package updates from those repos (Linux kernel and MESA packages, because I need newer versions of those than are currently offered on 22.04). ...This opens up some interesting mix & match possibilities for pulling different package sets from different ditros/distro version and not having to worry about keeping track of updates. I don't have any other specific use cases right now but I think maybe if I wanted a certain application to stay more up-to-date than it does in the regular Ubuntu repos, I could set it up to be updated through the Debian testing repo or something. I know they have Snap and Flatpak to help with keeping applications up-to-date, but my trials of those have ended in frustration with how sandboxed everything is. I do understand the purpose of the sandboxing, and that's great, but here's an example of the frustration. Yesterday, I tried to install Pinta through Flatpak and ended up with an "ow, my eyes" experience when it opened up a big bright screen and everything else on my system is in a "dark mode" theme. Apparently, at least when running KDE, Flatpak apps can't "see" the system theme and you have to jump through hoops to make that sort-of work. I tried downloading it from Github and compiled it myself and when I ran that one, Pinta just used the system-wide dark GTK theme that I have set and looks great. The whole Snap/Flatpak thing is all new to me because it wasn't a thing yet last time I tried Linux as a desktop OS and isn't something I had any need for during my normal use of Linux in a VM, mostly from the terminal. And, I am looking at Pinta because it is obviously inspired by paint.net, which is my preferred image editor on Windows. It uses a lot of the same GUI ideas and keyboard shortcuts, so less re-learning may be needed on my part. For system backup, I ran across Timeshift which seems to do what I want and was very easy to set up. (I did have to use the terminal to get it to set the snapshot location to my RAID/LUKS drive, which didn't pop up as an option in the GUI.) It backs up all of the system files to another drive, and you can just browse the snapshots in any file manager. Multiple "snapshots" will use hard links for unchanged files to save space. It excludes user files (everything in the home directory) so that you can roll back to a snapshot without having to worry about those being modified. (I already have a separate backup solution in place for that.) It can do fancier stuff if you use btrfs, but my volumes are all ext4 right now. Today, I finally set up Thunderbird for email. Working fine for mail, and I've used it before so it's familiar. Still haven't tried to figure out calendar & tasks but it looks like there are options to make that work. Sabrent says they have inspected my broken drive and are sending me a new one. No tracking details yet, though...

OK, I didn't realize that the default is 1440p instead of 1080p, but I believe what @win32asmguy did was a default run (so, 1440p). When you ran this test, were you running power through the dock? WD19TB3 does have 240W brick, but it can't deliver a full 240W to the system. You might have slightly better results if you have a 240W brick connected to the system directly. If you did have a 240W brick attached to the system directly, then it is interesting that your 1440p result is slightly lower than @win32asmguy's. You might also get slightly faster result if you switch the system to "Ultra performance" mode (thermal mode set in the Dell Power Manager or Dell Optimizer app).

Twitter policy states that logging in once every 30 days should be good enough to keep your account (presumably, regardless of whether or not you post). ...I've had my account for 12+ years and it has 0 tweets. So. "Reverse special treatment"?

Today, there are some new details on the AMD version of the Framework Laptop 13, featuring RDNA 3 integrated graphics. https://frame.work/blog/announcing-the-framework-laptop-13-powered-by-amd-ryzen

Can you do a default settings/1080p run just so we can have a sort of apples-to-apples comparison against @win32asmguy's RTX 4000 Ada GPU result? [Edit] You mentioned an external monitor run above. Were you using a Dell "dual USB-C" dock for power when you ran this? I know there can be a performance difference because of the difference in power (210W vs 240W, docked vs undocked).

Figured out "manually" turning the NVIDIA GPU on and off. Sort of. Took some finessing and I have a bit of work to do yet to make some quick scripts for this. (Maybe other distributions have a better default configuration and don't make this so awkward?) I discovered bbswitch which is a kernel module that can check the status of the NVIDIA GPU and also turn it on and off using ACPI commands. It agreed with nvidia-smi and the runtime_status thing above, stating that the NVIDIA GPU is on even though there are no apps using it. I could not use bbswitch to turn off the NVIDIA GPU because the driver was still loaded, and I could not unload the driver because it was "in use". I used Ubuntu's prime-select script to switch the system to "intel" mode rather than "on-demand" and rebooted. The system booted up on the Intel GPU, with the NVIDIA driver not loaded according to lsmod. bbswitch was still showing the NVIDIA GPU on, but now I was able to turn it off without issue. Then I went through these steps: Use bbswitch to turn the NVIDIA GPU on. (No problem.) Load the NVIDIA driver with modprobe. (Oops, prime-select actually blacklisted it, I had to go and clean that up.) ...That's really it, the NVIDIA driver is now usable (I can run a game on it). Also, nvidia-smi now does not show an Xorg process trying to use it. Now, I still need to figure out how to turn it off after I have loaded the NVIDIA driver once. If I try to unload it with modprobe, it shows an error that it is in use, even though no applications are using it according to nvidia-smi. Unloading the driver is a prerequisite for bbswitch being able to shut the card down. [Edit] Managed to do a power on / power off cycle; if I take care to not allow the "nvidia-drm" module to load then I have no problem unloading the rest of the NVIDIA kernel modules.

Some more messing around over the weekend. I installed "Crossover Linux" and under that I installed Office 365. I was mostly just sort of curious to se if it would work. I didn't want to try to figure out wintricks/etc. to get it working on my own. Crossover did make the install really easy. I'm most interested in being able to run OneNote, but that doesn't seem to be an option for now, it is pretty brittle and easy to crash (never mind the fact that you can't access cloud notebooks). Excel, on the other hand, seems surprisingly robust under Crossover. I used it for a while and had no issues at all. Quicken under Crossover didn't go well. I got stuck at the Intuit account login screen. It wouldn't take any keyboard input or copy/paste. I'd also like to be able to run Adobe applications but I didn't even try that; just from reading, it looks like its a no-go unless you use pretty old versions. It is working way better for me now than what I showed in the video, since I have worked out some performance issues with the VM, but that demonstration is definitely along the lines of what I would experience on a "native" Windows system that had been up for a long time with a whole lot of apps open. Thanks for the pointers here. I haven't messed with any of it yet, but I plan to in the near future. I will note that when I first started messing with this a couple of weeks ago, I observed that nvidia-smi was always reporting that the dGPU is powered on. I read comments online and saw mention that issuing a command to nvidia-smi will force the dGPU to temporarily power on, even if it is to just check the status? I also saw someone state that you can also use this to tell if the dGPU is powered on: cat /sys/bus/pci/devices/0000:01:00.0/power/runtime_status (Substituting appropriate PCI device ID) I checked that and it returns "active", so really agreeing with nvidia-smi. Closer inspection of nvidia-smi output reveals that the "Xorg" process seems to be using the dGPU. I will have to see if I can figure out what is going on there. Ha, you may have noticed that I am something of an "all or nothing" person with regards to this project. I'm not going to be switching back and forth between booting Windows and Linux. I don't think I've booted into my Windows install for over a week now (obviously not counting the VM I have going). Anyway, I already spent about half of my Windows gaming time in emulators so this is not something that I am averse to. Also, it looks like about one third of my Steam games have a native Linux port which is a higher ratio than I was expecting.

Having issues with KDE locking up. When it happens, KDE starts to act weird (mouse movement works, some apps are responsive but others are frozen, keyboard input generally doesn't work)... but the system is othersise still up and running – I can log in via SSH, I can still use my Windows VM (via RDP from another system), and I can switch to a console TTY once (if I switch back to the desktop TTY then the keyboard is fully unresponsive and I can't switch anymore). Symptoms seem similar to what these other people are reporting, there are many from March/April indicating that this problem just started "in the last few weeks" so I wonder if some sort-of recent package update has messed things up, or if it just doesn't like kernels newer than Ubuntu 22.04's base 5.15 (they would have dropped 5.19 out for the first time in late February). I haven't found a reported solution. I am really liking KDE in general, but I can't deal with unknown instability so I might be having to switch to something else or back to GNOME. 😕 [Edit] Maybe this is the solution? If so, it's not just Kubuntu, looks like a KDE/Mesa conflict that would be affecting multiple distros. I installed the Mesa update from the PPA referenced; we'll see if this continues to occur. I don't really like being on the bleeding edge like this so if it does fix the problem, maybe I'll try to instead use the Mesa packages from Ubuntu 23.04 (looks like they are new enough to bypass this bug). I already drug the kernel packages over from there.

We also hashed this over in Precision 7X70 pre-release thread last year. The 3600 MHz limitation with 128GB is a limitation of 12th gen Alder Lake, not the CAMM module. (There's an Intel PDF document linked in there confirming the limitation.) There is tons of misleading support material out there from multiple laptop manufactures (even Dell's own marketing material in some cases indicates that it should work at a higher speed than 3600 MHz).

Yes, it is true. Precision 7X70 and 7X80 systems use CAMM modules before the standardization is done, and as you say, there is a chance that the standard could evolve before it becomes more widely used so future CAMM modules would not work in these systems. Dell launched these systems with CAMM to demonstrate their commitment to the standard and that the modules work. You aren't forced to use CAMM in these systems; you can order these systems with a SODIMM interposer and use SODIMM modules instead. Several users here have done so in order to save some money or install their choice of modules. I understand the reservations but I can't really fault Dell's approach here. They're the only manufacturer to get ahead of the problem with SODIMM and put out something that works (other than soldered modules), and they got JEDEC to go along with it. (There's a downside to choosing SODIMM in these systems, though; with only two modules, 64GB is the memory cap... maybe 96GB soon, though, as it does look like 48GB SODIMMs aren't that far off.)

…From last year. I posted a link to this very article in the Precision 7770 prerelease thread here when it was newly posted.

Old news? There was a lot of chatter about this in the Precision threads when it first came up. (My own laptop has a CAMM module.) Since then, JEDEC has blessed CAMM as the next standard for laptop memory and other manufacturers are expected to start using it. (SODIMM is going to have signaling issues as RAM speeds continue to ramp up, so something has to be done, and this is better than soldered memory at least.) https://www.extremetech.com/computing/342342-jedec-says-dells-camm-laptop-memory-tech-will-replace-so-dimm

The key is "burned" into the BIOS these days, it should activate automatically or you can pull the key out with ProduKey; since version 1.70 it can read the key out of the BIOS and will display it in the list. (That tool might generate an antivirus false positive.)

You can try resetting the NVRAM. Hold the power button down for 30 seconds. If successful, the front LED will blink a few times, and then you can try powering on the system again; all BIOS settings will be reset. Other than that, the best thing is to call ProSupport. They are open 24/7. They might have some ideas to kick it into working. Otherwise, they will most likely dispatch a replacement motherboard if they have one available, or if not then you are looking at a system replacement, in which case getting a 7770 or 7780 is likely because I don't think that they have 7760 with Core i9 available to build anymore. (But that also means you're looking at a few weeks for them to build and ship the replacement system.) [Edit] They might dispatch a replacement motherboard and only after some time will you get a notice saying that the part is delayed. This would probably be Monday morning when they actually start looking for the part to ship out. You might not get an alert about this, but it will show on the dispatch status page. If this happens, call them back right away and ask for a system replacement. I have a feeling that if there is a motherboard parts delay, it might be longer than whatever their original projection is (if you even ever get one).

Google Authenticator uses the TOTP standard which is supported by any number of apps. If you go to a website that has Google Authenticator support for 2FA, you can scan the QR code or or type the secret key into (insert your favorite 2FA app here) and it will work just fine. Microsoft Authenticator is an option. It supports the same type of codes. I'm sure there are zillions more options for phone apps. Myself, I hate having this tied to my phone. I use KeePassXC as a password manager on my desktop. It supports also TOTP codes so I just file them in there. I can access them on my phone with the app "KeePassium" which supports the same database format and TOTP codes as well. This is for setting up a new 2FA entry from scratch. As for migrating existing codes off of Google Authenticator and onto a different app, I have no idea... I haven't used Google Authenticator in many years so I'm not sure what import/export functionality it has.

Another question for @Ionising_Radiation, since you've messed with this stuff before (more than I have anyway). I've been reading some stuff and have a few starting points to look at, but I am curious if you have any practical experience or tips to share. Looking at Optimus / dGPU stuff. Do you know anything about how to do things like ... Tell if the dGPU is powered on or not? Tell which app(s) are using the dGPU? Start an app on a particular GPU? Force the dGPU to power off (no apps allowed to use it), in a way such that it can be powered back on without a reboot? So far I have been basically let things operate "by default" and I didn't have to do anything in particular to get games working on the dGPU. But, I'd like to exercise more control over this, make sure that the dGPU is fully powered off when it should be, and let some simpler games like Baba Is You just run on the iGPU. ——— While I had tested some Linux Steam games right away, I did fire up some Windows Steam games (using Proton) for the first time yesterday to see what's what ... I know that there will be issues with some of them, but so far I am impressed with the "it just works" aspect of it all.

It's the battery status light so you will generally only see it light up if the battery is charging. You can tell if the system is powered on from the LED light around the power button, but like you say, if the lid is shut then there's no indication. At least that power button LED is an improvement. Precision 7560 (that I am typing on now) has no power LED indicator at all. Is there an issue with the RTX 4000 Ada getting too hot? I think even that chip is already large enough that there should be no issue keeping it cool at the power level that it would run at in a laptop. (I'll be pretty shocked if RTX 5000 Ada gets more than 5% better performance, if the workload is not vRAM-constrained. That's been about the performance difference expectation between the 4000-level and 5000-level GPUs for the past five or six Precision generations now.)

I was running a big data processing job on my 7560 and the fans were running full tilt for about 24 hours. Now that they have settled down, the right fan has some sort of periodic rocky grind sound going on when it is running at low speed, that wasn't there before. Great. I see a heatsink assembly replacement in my near future.

Windows LTSC is definitely closer to regular Windows 10 than it is to Windows Server. It does exclude a lot of "fluff" though, basically all of the bundled UWP apps and related background services, and it is my choice whenever I can swing it. (I have my own article about it, linked in sig.) I have tried Windows Server on a laptop before and I recall running into some things like it being difficult to get Bluetooth working because required components weren't included and it wasn't obvious how to install them. I've also run into situations where some software (i.e. backup tools) will want you to pay more for a license because you are running on a server OS. But, if you can get past that sort of thing, it should be fine to use. Dell will support Windows 10 LTSC but I doubt that they will support Windows Server, if you have some issue with it. (Sounds like you're comfortable supporting the OS on your own, though...?) I do not think that you would have any trouble setting up Hyper-V VM workloads on Windows 10 LTSC. Regarding 7680 vs 7780, it is basically as you say. The systems are very similar. The main difference is the physical size and 7780 getting an extra NVMe slot. There is a modest performance difference as well, with 7780 getting higher power limits and a slightly beefier cooling system, allowing for both the CPU and GPU to run at slightly higher speeds. One other thing to note is that 7680 has a 16:10 display while 7780 has 16:9, so you get a bit more "desktop working space" on the 7680 even though the screen is technically smaller. With Dell slimming up the bezels and everything, you will see Precision 7680 being a bit smaller than your M4800. Precision 7780 would be larger, but not as large as a Precision M6800.

Microsoft clarifies near-term Windows support roadmap: Windows 10, version 22H2 is the last version of Windows 10. It will be supported until October 14, 2025. A version of Windows 11 Enterprise LTSC will be released in the second half of 2024. (Windows 10 Enterprise LTSC will remain supported past the end of 2025; see details for each version at the top of the first post in this thread.) https://techcommunity.microsoft.com/t5/windows-it-pro-blog/windows-client-roadmap-update/ba-p/3805227

Finished up yet more scripting to make sure that certain apps/processes stay pinned only on the E cores and with low scheduling priority, and others like games stay pinned only on the P cores with high scheduling priority. I also added scripting to automatically move the Windows VM to the P cores when I am actively using it and to the E cores (with low scheduling priority) when I am not. I also decided to disable Linux's "autogroup" scheduling feature; I don't see any situation where fixes more problems than it creates, for me, since I guess I like to have a strong input in system-wide process scheduling. Playing around with learning how to insert special characters. On Windows, this is holding down the "Alt" key and then keying in a four-digit sequence on the numeric keypad. I have a lot of them memorized that I use regularly, like "—" and "×", and characters that pop up in Spanish like "é" and "ñ" and "¿". On Linux, you press the "compose key" and then key in a two (or more) character sequence to insert a special character. ...So, the Linux way will take some getting used to, but in the end it is definitely more intuitive and easy to "guess" what sequence you will need for more oddball characters that I can't remember the numeric sequence for, like "°" and "÷" and "→" and "©" and "…" and "€" and "≈" and "⅞". I like it so much that I went and installed WinCompose to be able to use the same sequences on Windows. I set the compose key to "right Windows" which is also Fn+Windows on Dell laptops. ————————————————————— I got in touch with a Sabrent person and with no fuss at all I was provided a shipping label to send the broken drive back to them. Only after receiving it will they ship out a replacement. I don't have to pay shipping; it is going to Florida via 2-day FedEx. I won't be able to ship it until tomorrow. I don't actually have anything that I feel comfortable shipping it in. I haven't kept any NVMe drive boxes. I ordered the cheapest NVMe 2280 drive that I could find (128 GB for $12 USD), which should arrive this evening, I will use its packaging to ship out this drive to Sabrent. And from now on, I will always keep at least one empty NVMe drive box around, I guess. Oh, and it turns out that I've been pronouncing "Sabrent" wrong since forever. On the phone recording, I heard that it has a long "a" sound as in "gave" — "Sabe-rent" — whereas I have always been saying it with a short "a" sound as in "rat". Not sure how long it will take me to stop saying it wrong, ha.

Not directly Twitter related, but Tesla/Musk related. Apparently, the new approach to dealing with past statements that come back to bite you is "I don't remember saying that" followed by: Tesla lawyers claim Elon Musk’s past statements about self-driving safety could just be deepfakes https://www.theverge.com/2023/4/27/23700339/tesla-autopilot-lawsuit-2018-elon-musk-claims-deepfakes

3×8TB in my Precision 7770 with no issues. I think it should work fine in Precision 7780. Well. I had trouble with them with RAID mode on (in BIOS setup). I recommend that you run the system in NVMe mode if your drives are "above spec" and use Storage Spaces or some other software solution if you want to combine drives into an array.

https://www.theverge.com/2023/4/26/23699102/microsoft-phone-link-imessage-ios-iphone-support-available-now It looks like Microsoft is rolling this out to the "general public" with the full rollout to be completed by mid-May. Meanwhile ... I'm still quite satisfied with Beeper, and I can report that it also works just as good on Linux, which I have recently switched over to on my "personal" laptop.

Well, it turns out that having fingerprint authentication enabled will block login to KDE (from SDDM, after a reboot), it just hangs after entering the password, like I experienced when I tried Kubuntu for the first time. Another case of KDE being "rough around the edges". I thought that maybe I wouldn't have to be as "careful" about process management on Linux, but... Last night I was just trying to watch TV in Firefox while also downloading some Steam games in the background. Turns out that Steam is happy to use 100% of 12 CPU cores while downloading and that was causing my video stream playback to stutter and hitch. Solved once I pushed the Steam process to E cores only. So, got to work out a solution to make things like that happen automatically. Process Lasso was really helpful for making this happen on Windows but I am not aware of a similar piece of software for Linux.