According to amd only epic cpus benefit from the standalone microcode update. All others need an updated bios with the new microcode. Zen 1 and zen 2 don’t need microcode updates. Also having the new microcode doesn’t mitigate the vulernability on its own as far as I can tell, the kernel needs to be the one doing the mitigation.
Seems like a lot of these “performance enhancing features” simply ignored security principles or tried to sidestep them, only for the features to introduce glaring security hole in the overall ISA, forcing people to then sidestep the supposed performance features so that it never mattered to begin with.
Are Intel, AMD and others pulling a fast one on us for the sake of gaining positive benchmarks?
If they were held liable, CPU manufacturers wouldn’t use these crappy hacks to increase performance, which helps their bottom line. Now I’m a cynic, so I’ll say that they might’ve done this on purpose.
I went to DL it and try it out, and… dude, I hate Sourceforge. I know how privileged this sounds, but I’m not spending a half hour DLing something that I can DL in a minute and a half somewhere else.
It’s funny because in a lot of contexts I’m pretty anti-comma but here we’ve got a confusing headline that would’ve traditionally included the comma for clarity.
As far as I know it essentially boils down to systemd doing too many things.
Critics argue that systemd is too complex and monolithic, making it harder to troubleshoot. They worry about a single point of failure, as all services are managed by one daemon, and voice concerns about tight integration with the Linux kernel, which could limit portability to other systems.
Looks interesting but I think claims of VMetal in regards of gaming are kind of bullshit. I have been looking into running windows/Linux in VM and have it full access to GPU (GPU pass through) and there are 2 major problems.
Nvidia and amd being scummy and having firmware limiters in consumer GPUs to segment their products. This should be “fixable” by modifying firmware but … Yeah… Not very comfortable thing to do
Motherboards not having support or worse have broken implementation of iommu.
Maybe my “research” was bad. Let’s hope someone will correct me.
Not sure what firmware limiters you’re talking about? I’m using a cheap ASUS board (B450i-gaming), a Zen 2 CPU and a 6600 XT, and single GPU passthru works just fine for me on Arch using this guide. (I haven’t tried VMetal or this new release of Nitrux yet). Yes, some manufactures are iffy about IOMMU support, mostly Intel-CPU and Intel-based boards in my experience, but if you’re using AMD you should be fine.
There is something called an ACS override patch, but that’s a kernel patch not a GPU firmware patch, and from my understanding, that’s for dual-GPU users. Regardless, it doesn’t modify your firmware in any way.
Not sure what firmware limiters you’re talking about?
The same limitation mentioned it the guide you shared:
This solution basically hands over the GPU to guest OS upon booting the VM and hands it back to the host OS upon powering off the VM. The obvious downside of this is that you can’t use the host OS (at least graphically) while the guest is running. It is therefore highly recommended that you set up SSH access to your host OS just in case of issues.
Also thanks for sharing it. I will try it some other time. 🙂
Immutable in this context usually means the root filesystem is readonly at runtime, and all changes are performed by updating a set of declarative config files that describe the desired state of the system. Changes would be prepared and applied after a reboot. Something like that.
Has anyone done an unbiased comparison of systems/openrc/whatever that compares each doing what it does best? So far the only difference is just me not having a clue how to do what I want when distro hopping.
At the end of the day, they all achieve exactly the same experience. The only real difference is that systemd does a lot of the work for you. Personally I wasn’t a huge fan of micromanaging my services manually on openrc or s6
They both achieve the same thing in different ways, but it comes down to philosophy if that even matters to the user. You can build gentoo for example with openrc or systemd, and depending on what you do you’ll need to integrate things differently (using elogind with openrc since logind is systemd specific). In gentoo it affects how you’ll compile. Its really a non issue, but it all depends on what level folks wanna spend on something that could potentially be different. I use openrc daily on multiple setups and it took maybe a week of using it to get the hang of it. Its just operationally different.
Does it really matter for most folks? No. But if its something you find interesting, it can be a nice change based on personal views and principles. I run openrc because I personally as a programmer don’t believe in how “proprietary” systemd is designed, nor agree with the decisions the maintainers have made. That’s just my opinion. At the end of the day it doesn’t technically matter.
This is a more hardcore viewpoint but it covers a lot of the issues folks have with it. suckless.org/sucks/systemd/
Cool I’ll check that out. I don’t think I’m at a level where the differences significantly affect me but I don’t really know what they are to begin with.
I installed Ubuntu in VirtualBox the other day. Right after install, I tried to open the default terminal. It wouldn’t open. Tried many times, restarted, tried again. Would not open.
Searched the web, found like a million other people having the same issue.
Installed Pop OS. It’s beautiful. Everything works, beautiful design, good wallpapers, good default theming.
Cool stuff. Especially excited by the fact that the Linux Mint team will continue looking at Wayland support. With Fedora starting to fall out of favour because of the telemetry stuff, I could see Mint become THE distro to recommend to someone. It really is heading that way.
I started using it again with version 115 and it’s so fast and easy to search mails, I spend a couple of hours deleting, archiving and cleaning up my inbox. Who knew that could be kind of fun. :)
I’ve been using it forever and I actually don’t think that much has changed. It has finally gotten some necessary builtin features that previously needed plugins (carddav, caldav) and the UI has been cleaned up a bit, but generally there are no game changers here. So why didn’t you use it before 115?
I have been using web interfaces to email for over 15 years now in think. I’ve been on Fastmail for the last 5 or so.
Before the web interfaces were popular, I was using email clients on the desktop, and I think I was using Thunderbird back then, but don’t really remember exactly…
9to5linux.com
Newest