planets on Random thoughts of Peter 'CzP' Czanik

Syslog-ng 4.12.0 available for Ubuntu 26.04 (Resolute)

Tue, 14 Jul 2026 14:41:41 +0200

Recently I was asked if syslog-ng supports Ubuntu 26.04 (Ubuntu Resolute). Yes, and with the arrival of the syslog-ng 4.12.0 release we also provide ready-to-use packages for it. The release notes mention it, and info is in the Readme on GitHub.

I tend to mention FreeBSD and openSUSE more often in my blogs (personal preference), so today I installed Ubuntu 26.04 and tested syslog-ng myself.

syslog-ng logo

Syslog-ng Java destination disabled

Thu, 09 Jul 2026 15:19:41 +0200

For many years, syslog-ng used Java, where C libraries were unavailable. However, over the years native C libraries became available for Elasticsearch and Kafka, and HDFS practically disappeared. As a “scream test”, I am going to disable Java support in all of my syslog-ng packages.

Once upon a time, Java support was added to syslog-ng to be able to load Elasticsearch Java drivers. Later, Kafka, HDFS, and a generic HTTP destination were also added. Unfortunately, using Java was a major pain. Loading libraries required some manual configuration. Packaging the Java destination in official Linux distribution packages was possible, however, packaging the actual drivers written in Java was impossible. For a while, I maintained unofficial packaging for these drivers, but as C alternatives appeared, I removed these components. Nobody complained. Recently all drivers, except for HDFS, have been removed from the source code as well. Again: nobody complained.

Right now, we still have HDFS support in the source code, but not for long. I have been posting about it for years now, and nobody asked us to keep it. This is the last driver making use of the Java destination of syslog-ng. We will delete it soon, too.

We do not delete code related to Java right now. However, as a “scream test” I am going to disable the Java destination in all my packages. I have done it already in the official openSUSE syslog-ng package. Fedora Rawhide is next. I will also remove it from my git snapshot packages.

If you use Java with your own driver code, let us know! Otherwise, the Java destination will be not just disabled in packages but removed from code as well.

syslog-ng logo

syslog-ng 4.12.0, syslog-ng PE 8.2.0 and SSB 7.8.0 are now available

Tue, 16 Jun 2026 16:50:23 +0200

Today, the syslog-ng team released three different syslog-ng versions, which provided a good opportunity to wear my “Release is coming” t-shirt. This coordinated release is due to an SQL injection security bug fix, even if most likely it affects less than a handful of people (I mean, is there anyone who still uses SQL to store logs in 2026…?)

Release is coming

What’s new?

syslog-ng OSE 4.12.0 arrives with performance optimizations, making syslog-ng more scalable in several situations. Many user-reported bugs were fixed as well. RPM-based containers are now available, utilizing Alma Linux under the hood.
syslog-ng PE 8.2.0 focused on bug fixes and now also allows users to use extremely complex configuration files.
SSB 7.8.0 adds support for Proxmox and Nutanix virtualization.

What’s next?

For more details, check the respective release notes. The syslog-ng OSE release notes are available at https://github.com/syslog-ng/syslog-ng/releases/tag/syslog-ng-4.12.0 Release notes for the commercial products should be available soon at https://docs.oneidentity.com

syslog-ng logo

The status of OpenSSL 4.0 support in syslog-ng

Thu, 04 Jun 2026 14:01:40 +0200

OpenSSL 4.0 was released just over a month ago. So, how is its support progressing in syslog-ng? Well, Git master already supports it, and the patch is easy to backport to earlier releases. At the same time, version 4.12 will support OpenSSL 4.0 out of the box.

A month ago, someone from Gentoo Linux reached out to the syslog-ng team about OpenSSL 4.0 support. I asked the community about their expectations, knowing that version 4.0 is not an LTS version. However, I quickly learned that all major distros were preparing to use it in their rolling development versions. A few days later, we also received hints on how to add support for it in syslog-ng. And thus, a PR was born, which is now merged into syslog-ng git master.

What does this mean for you?

If you need OpenSSL 4.0 support RIGHT NOW using a RELEASED syslog-ng version, then use syslog-ng 4.11 and the related pull request from https://github.com/syslog-ng/syslog-ng/pull/5688
Use the latest syslog-ng git snapshot, as the above PR is already merged.
Wait a few more weeks, as syslog-ng 4.12 will be released soon with OpenSSL 4.0 support.

syslog-ng logo

Originally published at https://www.syslog-ng.com/community/b/blog/posts/the-status-of-openssl-4-0-support-in-syslog-ng

Nightly syslog-ng containers based on Alma Linux

Tue, 26 May 2026 14:02:31 +0200

For many years, the syslog-ng project provided container images based on Debian. Most of our users run syslog-ng on RHEL & compatibles, and have asked for an RPM-based container. So, nightly containers based on Alma Linux are now also available.

A while ago, I prepared a small test project to run syslog-ng in an Alma Linux container: https://www.syslog-ng.com/community/b/blog/posts/experimental-syslog-ng-container-image-based-on-alma-linux However, that was only an experiment which I never updated. Fast forward to today: nightly syslog-ng containers based on the latest syslog-ng git snapshot package builds are now available on the Docker Hub!

syslog-ng logo

Learning syslog-ng

Tue, 19 May 2026 13:15:16 +0200

How can you learn syslog-ng? There are many possibilities, depending on your time and budget. Possibilities range from tutorial series through reading the documentation to instructor-led training. Find out which one is for you!

syslog-ng logo

The syslog-ng Insider 2026-05: OTEL; central log collection; old Mac

Thu, 14 May 2026 13:46:57 +0200

Dear syslog-ng users,

This is the 140th issue of syslog-ng Insider, a monthly newsletter that brings you syslog-ng-related news.

Streaming syslog-ng data to your lakehouse using OTEL

Version 4.11.0 of syslog-ng contains contributions from Databricks related to OAuth2 authentication. Recently, they published a blog about how this enables their customers to send logs to their data lake using syslog-ng and the OpenTelemetry protocol.

https://www.syslog-ng.com/community/b/blog/posts/streaming-syslog-ng-data-to-your-lakehouse-using-opentelemetry

Central log collection - more than just compliance

I often hear, even at security conferences that “no central log collection here” or “we have something due to compliance”. Central logging is more than just compliance. It makes logs easier to use, available and secure, thus making your life easier in operations, security, development, but also in marketing, sales, and so on.

https://www.syslog-ng.com/community/b/blog/posts/central-log-collection—more-than-just-compliance

Compiling syslog-ng on an old Mac

I have an aging, but fully functional MacBook. I bought it for syslog-ng testing, but I also use for watching movies. Homebrew no more fully supports old, Intel-based Macs. This blog helps to compile the latest syslog-ng release on these old, but otherwise functional machines.

https://www.syslog-ng.com/community/b/blog/posts/compiling-syslog-ng-on-an-old-mac

syslog-ng logo

Your feedback and news, or tips about the next issue are welcome. To read this newsletter online, visit: https://syslog-ng.com/blog/

Fedora 44, CentOS 7 and Amazon Linux syslog-ng questions

Wed, 06 May 2026 13:28:09 +0200

Fedora 44 was announced last week: syslog-ng 4.11 is part of it. While checking the Fedora Copr build service for Fedora 44, I realized that CentOS 7 and Amazon Linux 2023 packages are also there. I have a few questions about those for you!

syslog-ng logo

Fedora 44

The availability of the Fedora 44 release was announced last week. Vesion 4.11 of syslog-ng, the current latest release, is part of it. As usual, I did a quick test: everything works as expected.

RHEL 6

The removal of RHEL 6 packages from Copr was announced many years ago. Then, the countdown was silently canceled. I have just checked: RHEL 6 packages are no longer available, so I deleted all my related repositories. Also, I deleted a couple of temporary test repos along the way.

CentOS 7

When talking to product manager friends around the world, I realized that syslog-ng is not an “enterprise” application. “Enterprise” developers are still actively maintaining packages for RHEL 6, when even RHEL 7 has reached end of life a long time ago. Of course, this is just a satirical definition of “enterprise”, at least in my view…

Support for RHEL 7 / CentOS 7 was dropped in syslog-ng a month before the distro became end-of-life. Copr announced the deletion of CentOS 7 packages, but after a while, the countdown suddenly disappeared. Packages built 10+ years ago are still here, and CentOS 7 is still a valid build target.

Question: is there anyone still using my syslog-ng packages on RHEL 7 / CentOS 7? Otherwise, I would be happy to delete anything related from my Copr repositories. I could delete many repositories, save storage, and I would not have to deselect them as a build target during package builds.

Amazon Linux 2023

Another question mark in my mind is Amazon Linux 2023 support. If we can believe the download statistics provided by Copr, then this is one of the most popular syslog-ng repos on Copr. However, over the years, I only received a single feedback about it, which was on Twitter years ago: “Thanks, I use it.” That is all. While there were regular requests to create these packages, nobody asked for features, updates, whatever. The repo is still stuck at syslog-ng version 4.8.

Question: should I update syslog-ng to a more recent version, as time permits?

What is next?

Share your thoughts with us in this syslog-ng GitHub discussion: https://github.com/syslog-ng/syslog-ng/discussions/5691 or reach out to me on Twitter / Mastodon / LinkedIn.

Originally published at https://www.syslog-ng.com/community/b/blog/posts/fedora-44-centos-7-and-amazon-linux-syslog-ng-questions

Support for OpenSSL 4.0?

Tue, 28 Apr 2026 15:20:56 +0200

Although OpenSSL 4.0 released just two weeks ago, the syslog-ng project has already received a GitHub issue complaining that we do not support it. So, before we would allocate too much effort on it: what should we expect?

OpenSSL 4.0 was announced on April 14: https://openssl-library.org/post/2026-04-14-openssl-40-final-release/ However, this announcement mentions that it is NOT a long-term support (LTS) release.

This raises the question that if it is not an LTS release, then can we stay on version 3.x and skip 4.x altogether? When will Linux distributions start using it? Looking at Repology, there are already a few places where OpenSSL 4.0 is available. This includes Gentoo, the community where the GitHub issue originated from, and also various FreeBSD ports. The current list is available at: https://repology.org/project/openssl/versions

Fedora is planning to use OpenSSL 4.0 as default starting from the next release: https://fedoraproject.org/wiki/Changes/OpenSSL40 However, OpenSSL 3.x will most likely stay supported for backwards compatibility.

I am also curious if there are any other projects which have added support for OpenSSL 4.0. If so, then what are your experiences? Was porting your code to use OpenSSL 4.0 difficult?

I am all in for supporting the latest technologies, but currently, even if we have an open request for OpenSSL 4.0 support, I do not feel that I have enough information to prioritize its development.

Share your thoughts with us in this syslog-ng GitHub discussion: https://github.com/syslog-ng/syslog-ng/discussions/5685 or reach out to me on Twitter / Mastodon / LinkedIn.

syslog-ng logo

Originally published at https://www.syslog-ng.com/community/b/blog/posts/support-for-openssl-4-0

New performance tuning possibilities in syslog-ng

Wed, 22 Apr 2026 12:57:45 +0200

On April’s fool’s day, I shared that syslog-ng can reach 7 million EPS. This test lab result was in part possible thanks to a few performance enhancements coming to syslog-ng version 4.12.

How 7 million EPS is possible? Before diving deeper, let me repeat it: 7 million EPS is just a lab testing result, not (yet) possible in the real world. However, the technologies enabling this are already available on the development branch of syslog-ng, or have been available for ages, just not tested or promoted enough.

syslog-ng logo

Discussing RTO in my Genesi t-shirt...

Thu, 16 Apr 2026 09:58:48 +0200

This Monday I talked to a couple of friends about work while wearing my Genesi t-shirt. A teacher going back to school after Spring break and an IT guy explaining the nightmare of RTO threat. I love coincidences :-) Why do I say that?

Genesi t-shirt

As I wrote a few years ago about working from home: “After graduating from university, I worked from home for a small US-based company. I never met my boss while working there and met only one of my colleagues at a conference in Brussels. I eventually met my boss some seven years later, when I gave a talk at a conference in Washington, D.C.” The company was Genesi, and that was the work culture which defines me. I received the t-shirt on the photo during my visit to Washington, D.C.. Luckily, I’m still living mostly this way, visiting the office 1-2 times a week: working hybrid.

Imagine the contrast I felt, when I realized that I’m talking to someone who works on a very strict fixed schedule. For a teacher vacation is only possible when there is no school, like Spring break in Hungary last week. There is a fixed schedule all year around. Compare that to my Genesi years: no regular meetings, communicating by e-mail & chat, and working when it was the right time for me: sometimes in the morning, other days during the night. It was fantastic, especially with small kids. I have been working on flexible hours ever since, limited only by meetings.

COVID made remote work less of a niche. Sometimes even mandatory. Many people in IT started to work remotely. Most of our work does not require a fixed place or time. On-line meetings became the norm, teams are often not location based anymore but scattered around the globe. As long as you have an Internet connection and a noise canceling microphone you can join a meeting from anywhere, even from the top of a mountain. It is easy to get used to this flexibility and very difficult to give it up.

RTO became a periodical threat. It’s a lot cheaper to announce RTO and let people leave voluntarily than sending them away. Quite a few friends write me every once in a while that they have to return to the office starting in a few weeks time. Then, a few weeks later they happily share: they gave me an exemption, so they do not want me to leave…

Wearing my Genesi t-shirt all these problems feel so distant. I hope that it stays this way!

Streaming syslog-ng data to your lakehouse using OpenTelemetry

Wed, 15 Apr 2026 14:12:22 +0200

The syslog-ng project received two contributions from Databricks in the last weeks of 2025. The first one turned the already existing OAuth2 support generic and extensible, so it can be used anywhere, not just with Microsoft Azure (but of course, Azure compatibility was preserved). The next pull request was built on the first one and enabled OAuth2 support for gRPC-based destinations, like OpenTelemetry, Loki, BigQuery, PubSub, ClickHouse, etc. These changes were released as part of the syslog-ng 4.11.0 release. You can read more about these in the release notes at https://github.com/syslog-ng/syslog-ng/releases/tag/syslog-ng-4.11.0

Besides an excellent overview about syslog-ng, the related Databricks blog also provides step-by-step instructions on how to use syslog-ng with their product. You can read it at: https://community.databricks.com/t5/technical-blog/streaming-syslog-ng-data-to-your-lakehouse-powered-by-zerobus/ba-p/153979

syslog-ng logo

Originally published at https://www.syslog-ng.com/community/b/blog/posts/streaming-syslog-ng-data-to-your-lakehouse-using-opentelemetry

My new toy: April 1 syslog-ng performance tests

Wed, 01 Apr 2026 12:35:28 +0200

Almost 15 years ago, Balabit had a campaign, stating that syslog-ng could process 650k messages a second. Now I am happy to present 7 million EPS (events per second). Timing the announcement to April 1 is not a coincidence :-)

While the 650k EPS measurement was true, it was misleading. This value was measured right after syslog-ng 3.2 introduced multi-threading, in lab environment, under optimal circumstances, using synthetic log messages. However, there was no fine print explaining this, just the statement that syslog-ng could process 650k EPS. It was fixed after a while, but it took years to recover from the effects of this marketing campaign, and engineers ten years later still had a nervous breakdown when someone mentioned “650k”. Why? Because from that moment, everyone expected syslog-ng to collect logs at that message rate in a production environment with complex configurations. Which was of course not the case.

Fast-forward to today, I’m happy to share that:

syslog-ng can collect logs at 7 million EPS

Is this measurement value valid? Yes.
Does it apply to real world? No.
Does it sound good? Definitely :-)

My latest syslog-ng benchmark results

The tool: sngbench

I love playing with various non-x86 systems. I have various ARM, POWER, MIPS systems at home, and sometimes I access other architectures, like RISC-V remotely. And, of course, not just different architectures, but different operating systems: various Linux distributions, MacOS, FreeBSD, sometimes also other BSD variants. I’m a server guy, and for the past 15+ years: a syslog-ng guy. Sometimes I had access to an exotic system on the other side of the world only for less than an hour, but I almost always tested syslog-ng.

For many years I had a bunch of shell scripts and configs to benchmark syslog-ng performance. Not for real world production loads, but rather for comparing architectures and operating systems. I needed a script which could do measurements with minimal dependencies and do it quickly, in one go. This is how sngbench was born, based on my previous ugly scripts. It has quite a few advantages and shortcomings:

Minimal dependencies: bash and syslog-ng
No complex setup: everything runs on the same host
network bandwidth is not a limiting factor
loggen and syslog-ng processes are competing for resources
Two bundled configurations: a performance tuned and the default syslog-ng.conf from openSUSE with minimal modifications to add a TCP source
By default, very short (20 seconds) measurements, so disk I/O is not a limiting factor
Many different test scenarios: from a single TCP connection to 4 * 128

Of course this describes just the “factory defaults”. You can easily change the test scenarios and configurations too.

How I reached 7 million EPS, and why it is not relevant

I was testing syslog-ng code which was not yet even merged to the development branch. First, I tested these patches with various settings. Along the way I remembered that Splunk guidelines mention so-rcvbuf tuning also for TCP connections. Previously I only used that for optimizing UDP performance. Now I have done it for TCP. Wonders happened :-)

But, of course, the main question is: can you achieve this performance in production? TL;DR: No.

My tests are run from localhost. Network bandwidth is not an issue. Tests are run in short bursts. This is peak performance; when it comes to writing logs to files or forwarding to a cluster of Splunk or Elasticsearch endpoints around the clock, that would be slower. Also, in my fastest test case, logs came from four different loggen instances, over 32 TCP connections each, at a constant rate. In the real world, logs come in bursts and connections are opened and closed regularly.

Test environment and tests

I used my AI mini workstation with Fedora Linux 44 Beta. First, I took a base line with stock syslog-ng 4.11.0 included in the distribution. Then I used my syslog-ng git snapshot packages for Fedora from https://copr.fedorainfracloud.org/coprs/czanik/syslog-ng-githead/. Initially it also had jemalloc support compiled in. Later I disabled it and purely focused on the yet to be merged parallelize() optimizations from GitHub. I experimented with enabling and disabling parallelize(), adding various batch_size() values, and finally also so-recbuf().

AI in a miniature box :-)

This blog is part of a longer series about my adventures with my new machine and AI. You can reach me to discuss this blog on one of the contacts listed in the upper right corner. You can read the rest of the blogs under the toy tag.

My new toy: Back to high-end audio

Tue, 31 Mar 2026 10:12:38 +0200

My AI mini workstation from HP has seen some non-AI workloads this weekend. I installed Capture One for photo editing and a couple of software synthesizers. And realized along the way that while built-in speakers are nice, high-end audio is a lot better! :-)

For months, I have been listening to music on devices that are designed for speech: a pair of Jabra headphones and the speakers of my various laptops. There were many reasons for this, including peer pressure, and some hearing loss at a way too loud concert. I was also lazy to use my high-end devices and tried to persuade myself that audio equipment designed for meetings is good enough for music too. Well…

This weekend, I installed various software synthesizers on my new computer. Not that I learned music or could play any instruments, but I still enjoy experimenting with music (well, with noise, actually :-) ). As I connected the machine to the big screen in the living room, I also connected it to my HiFi system. Suddenly, I realized how much better it sounds than my laptop or anything I’ve listened to in the past few months.

While making noise with a couple of software synths and listening to music from my TIDAL subscription, I also recharged my Focal headphones. My Focal Bathys is not as good as my HiFi, but also has a wonderful sound regardless.

So I guess that after a few months long detour, I am back to using high-end audio gear whenever it is technically possible. I love the extra detail I can hear on my Heed Enigma speakers or on my Focal headphones. Of course, nothing can replace listening to live music at concerts, but high-end gear is much better at approximating the vibe of various live events than anything below it.

AI in a miniature box :-)

My new toy: Open WebUI first steps

Thu, 26 Mar 2026 13:42:45 +0100

Once I got hardware-accelerated AI working under Linux on my AI mini workstation from HP, my next goal was to make it easier to use. From this blog, you can read about my initial experiments with Open WebUI on Fedora Linux.

Open WebUI talking about central log collection :-)

Everything in containers

As Open WebUI is not yet available as a package in Fedora, my initial approach was to use containers. I found a Docker compose setup which was tested on Fedora Linux 43 according to its documentation: https://github.com/jesuswasrasta/ollama-rocm-webui-docker. As I (also) use Fedora 43, it sounded like a good choice.

It worked; however, I quickly realized that hardware acceleration for AI was not working. Instead of that, most CPUs were running close to 100%. It was a good test for cooling: I could hear the miniature box from the next room through closed doors :-)

ollama eating CPU :-)

As it turned out, the content of the HSA_OVERRIDE_GFX_VERSION environment variable was incorrect. When I set it according to the docs, hardware acceleration still did not work. Removing the environment variable ollama found the hardware, but never answered a prompt anymore.

Ollama from the system

My next experiment was that I kept using Open WebUI from the container, but I installed ollama from the Fedora package repository directly on the system. The good news? Some smaller models ran really fast, using hardware acceleration. The bad news: most models failed to load with an error message that the given model format is unknown.

Update to Fedora 44 beta

I guessed that ollama was too old in Fedora 43. Solution? Update the whole system to Fedora 44 beta. It seems to have helped. A lot more models work now, including the largest freely available Granite models from IBM.

Why Granite?

First of all: I’m an IBM Champion, and thus using IBM technologies is for granted. But also because I learned some background stories from a friend working at IBM on LSF, which makes it also a personal choice.

What I’ve been showing here is AI inferencing on my HP AI system. But before the model can be used (for inferencing), it needs to be trained. These models are trained on large, GPU rich conpute clusters. To get an idea of the scale of such clusters, you can learn more in this research paper (https://arxiv.org/abs/2407.05467). It duscusses the IBM Blue Vela system which supports IBMs’ GenAI mission. What’s interesting is the Blue Vela uses a more traditional HPC software stack including IBM LSF for workload management and Storage Scale (GPFS) for rapid access to large data sets.

AI in a miniature box :-)

Compiling syslog-ng on an old Mac

Wed, 25 Mar 2026 15:43:49 +0100

syslog-ng logo

My new toy: first steps with AI on Linux

Wed, 25 Mar 2026 12:48:34 +0100

Ever since I bought my AI mini workstation from HP, my goal was to run hardware accelerated artificial intelligence workloads in a Linux environment. Read more to learn how things turned out on Ubuntu and Fedora!

I have been using various AI tools for a while now. Generating pictures about some impossible situations, like a dinosaur climbing the Hungarian parliament building, finding information where a simple web search is useless, or explaining syslog-ng code to me. All these are nice, sometimes even useful, however I prefer to know what is behind the magic. Well, at least part of it :-) I want to get a bottom up view of various components and processes, and getting my hands dirty. Hopefully this miniature but powerful box will help me in getting known with AI better.

AI in a miniature box :-)

Testing AI on Ubuntu

As mentioned in my installing Ubuntu blog, the 24.04 LTS installer did not work on this machine. I found a nice tutorial about AI on the Ryzen AI Max+ 395 which mentioned using 25.10, so I installed that version instead of the LTS. It installed without any troubles, 3D graphics worked out of the box.

However, AI is a different story. ROCm, hardware acceleration for AI workloads on AMD chips, is only packaged for Ubuntu LTS releases. The workaround described in the tutorial was to use distrobox. Unfortunately, the steps described in the tutorial did not work. Containerization brought in various problems with permissions, software availability, and so on. Most likely an experienced distrobox user could resolve these. In my case, after reading the distrobox documentation for hours, I just gave up.

Getting started with hardware accelerated AI on Fedora

Next, I turned to Fedora Linux 43. The wiki page of the Fedora Heterogeneous Computing Special Interest Group proved to be a good starting point. Fedora has ROCm packaged as part of the distro, and the wiki page gives clear instructions how to get started.

Once I set up user rights and installed the necessary packages, I was able to get some info about my hardware. You can see the output of rocminfo and rocm-clinfo at the bottom of this blog. I did not want to shorten those, but given the many lines of output, I was not sure if anyone would read the rest of my blog :-)

Testing with llama

Of course, seeing info about the hardware is nice, but it’s even better to see it in action. The Ubuntu ROCm tutorial mentioned llama, so I started with that one. Luckily Fedora includes it as a ready to install package, so I did not have to compile it from source. I also installed huggingface-hub, also from a package:

dnf install python3-huggingface-hub llama-cpp

This allowed me to download the model mentioned in the tutorial, and ask a few questions from the downloaded LLM. For now I just used the sample command line, but based on the output llama found the hardware and used it. Next up: learn more about the available models.

You can find the output of the following command at the end of this blog:

llama-cli   -m ~/models/llama-2-7b.Q4_K_M.gguf   --no-mmap   -ngl 99   -p "Explain quantum computing in simple terms:"   -n 256

Testing with pytorch

When I mentioned a friend that hardware accelerated AI seems to work on my Linux box, he suggested to me to try it with PyTorch. Luckily this was available as a ready to install package for Fedora as well:

dnf install python3-torch

I was quite a bit surprised, as the above command installed 8 GB worth of RPM packages (texlive accounting for a good part of it). I do not know much about PyTorch, but did a quick test anyway. Here is the really complex Pyhon code I built based on the documentation:

import torch
x = torch.rand(5, 3)
print(x)
print('Is hw AI accel available')
print(torch.cuda.is_available())

And here is the output from the above code:

tensor([[0.1034, 0.0183, 0.1233],
        [0.1787, 0.0097, 0.8426],
        [0.2872, 0.6351, 0.8468],
        [0.8226, 0.2991, 0.8539],
        [0.2061, 0.6422, 0.8146]])
Is hw AI accel available
True

It’s simple, but looks promising :-)

Outputs

Ooutput of rocminfo and rocm-clinfo

czanik@fedora:~$ rocminfo 
ROCk module is loaded
=====================    
HSA System Attributes    
=====================    
Runtime Version:         1.1
Runtime Ext Version:     1.7
System Timestamp Freq.:  1000.000000MHz
Sig. Max Wait Duration:  18446744073709551615 (0xFFFFFFFFFFFFFFFF) (timestamp count)
Machine Model:           LARGE                              
System Endianness:       LITTLE                             
Mwaitx:                  DISABLED
XNACK enabled:           NO
DMAbuf Support:          YES
VMM Support:             YES

==========               
HSA Agents               
==========               
*******                  
Agent 1                  
*******                  
  Name:                    AMD RYZEN AI MAX+ PRO 395 w/ Radeon 8060S
  Uuid:                    CPU-XX                             
  Marketing Name:          AMD RYZEN AI MAX+ PRO 395 w/ Radeon 8060S
  Vendor Name:             CPU                                
  Feature:                 None specified                     
  Profile:                 FULL_PROFILE                       
  Float Round Mode:        NEAR                               
  Max Queue Number:        0(0x0)                             
  Queue Min Size:          0(0x0)                             
  Queue Max Size:          0(0x0)                             
  Queue Type:              MULTI                              
  Node:                    0                                  
  Device Type:             CPU                                
  Cache Info:              
    L1:                      49152(0xc000) KB                   
  Chip ID:                 0(0x0)                             
  ASIC Revision:           0(0x0)                             
  Cacheline Size:          64(0x40)                           
  Max Clock Freq. (MHz):   5187                               
  BDFID:                   0                                  
  Internal Node ID:        0                                  
  Compute Unit:            32                                 
  SIMDs per CU:            0                                  
  Shader Engines:          0                                  
  Shader Arrs. per Eng.:   0                                  
  WatchPts on Addr. Ranges:1                                  
  Memory Properties:       
  Features:                None
  Pool Info:               
    Pool 1                   
      Segment:                 GLOBAL; FLAGS: FINE GRAINED        
      Size:                    131136832(0x7d0fd40) KB            
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:4KB                                
      Alloc Alignment:         4KB                                
      Accessible by all:       TRUE                               
    Pool 2                   
      Segment:                 GLOBAL; FLAGS: EXTENDED FINE GRAINED
      Size:                    131136832(0x7d0fd40) KB            
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:4KB                                
      Alloc Alignment:         4KB                                
      Accessible by all:       TRUE                               
    Pool 3                   
      Segment:                 GLOBAL; FLAGS: KERNARG, FINE GRAINED
      Size:                    131136832(0x7d0fd40) KB            
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:4KB                                
      Alloc Alignment:         4KB                                
      Accessible by all:       TRUE                               
    Pool 4                   
      Segment:                 GLOBAL; FLAGS: COARSE GRAINED      
      Size:                    131136832(0x7d0fd40) KB            
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:4KB                                
      Alloc Alignment:         4KB                                
      Accessible by all:       TRUE                               
  ISA Info:                
*******                  
Agent 2                  
*******                  
  Name:                    gfx1151                            
  Uuid:                    GPU-XX                             
  Marketing Name:          Radeon 8060S Graphics              
  Vendor Name:             AMD                                
  Feature:                 KERNEL_DISPATCH                    
  Profile:                 BASE_PROFILE                       
  Float Round Mode:        NEAR                               
  Max Queue Number:        128(0x80)                          
  Queue Min Size:          64(0x40)                           
  Queue Max Size:          131072(0x20000)                    
  Queue Type:              MULTI                              
  Node:                    1                                  
  Device Type:             GPU                                
  Cache Info:              
    L1:                      32(0x20) KB                        
    L2:                      2048(0x800) KB                     
    L3:                      32768(0x8000) KB                   
  Chip ID:                 5510(0x1586)                       
  ASIC Revision:           0(0x0)                             
  Cacheline Size:          128(0x80)                          
  Max Clock Freq. (MHz):   2900                               
  BDFID:                   50432                              
  Internal Node ID:        1                                  
  Compute Unit:            40                                 
  SIMDs per CU:            2                                  
  Shader Engines:          2                                  
  Shader Arrs. per Eng.:   2                                  
  WatchPts on Addr. Ranges:4                                  
  Coherent Host Access:    FALSE                              
  Memory Properties:       APU
  Features:                KERNEL_DISPATCH 
  Fast F16 Operation:      TRUE                               
  Wavefront Size:          32(0x20)                           
  Workgroup Max Size:      1024(0x400)                        
  Workgroup Max Size per Dimension:
    x                        1024(0x400)                        
    y                        1024(0x400)                        
    z                        1024(0x400)                        
  Max Waves Per CU:        32(0x20)                           
  Max Work-item Per CU:    1024(0x400)                        
  Grid Max Size:           4294967295(0xffffffff)             
  Grid Max Size per Dimension:
    x                        4294967295(0xffffffff)             
    y                        4294967295(0xffffffff)             
    z                        4294967295(0xffffffff)             
  Max fbarriers/Workgrp:   32                                 
  Packet Processor uCode:: 34                                 
  SDMA engine uCode::      18                                 
  IOMMU Support::          None                               
  Pool Info:               
    Pool 1                   
      Segment:                 GLOBAL; FLAGS: COARSE GRAINED      
      Size:                    65568416(0x3e87ea0) KB             
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:2048KB                             
      Alloc Alignment:         4KB                                
      Accessible by all:       FALSE                              
    Pool 2                   
      Segment:                 GLOBAL; FLAGS: EXTENDED FINE GRAINED
      Size:                    65568416(0x3e87ea0) KB             
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:2048KB                             
      Alloc Alignment:         4KB                                
      Accessible by all:       FALSE                              
    Pool 3                   
      Segment:                 GROUP                              
      Size:                    64(0x40) KB                        
      Allocatable:             FALSE                              
      Alloc Granule:           0KB                                
      Alloc Recommended Granule:0KB                                
      Alloc Alignment:         0KB                                
      Accessible by all:       FALSE                              
  ISA Info:                
    ISA 1                    
      Name:                    amdgcn-amd-amdhsa--gfx1151         
      Machine Models:          HSA_MACHINE_MODEL_LARGE            
      Profiles:                HSA_PROFILE_BASE                   
      Default Rounding Mode:   NEAR                               
      Default Rounding Mode:   NEAR                               
      Fast f16:                TRUE                               
      Workgroup Max Size:      1024(0x400)                        
      Workgroup Max Size per Dimension:
        x                        1024(0x400)                        
        y                        1024(0x400)                        
        z                        1024(0x400)                        
      Grid Max Size:           4294967295(0xffffffff)             
      Grid Max Size per Dimension:
        x                        4294967295(0xffffffff)             
        y                        4294967295(0xffffffff)             
        z                        4294967295(0xffffffff)             
      FBarrier Max Size:       32                                 
    ISA 2                    
      Name:                    amdgcn-amd-amdhsa--gfx11-generic   
      Machine Models:          HSA_MACHINE_MODEL_LARGE            
      Profiles:                HSA_PROFILE_BASE                   
      Default Rounding Mode:   NEAR                               
      Default Rounding Mode:   NEAR                               
      Fast f16:                TRUE                               
      Workgroup Max Size:      1024(0x400)                        
      Workgroup Max Size per Dimension:
        x                        1024(0x400)                        
        y                        1024(0x400)                        
        z                        1024(0x400)                        
      Grid Max Size:           4294967295(0xffffffff)             
      Grid Max Size per Dimension:
        x                        4294967295(0xffffffff)             
        y                        4294967295(0xffffffff)             
        z                        4294967295(0xffffffff)             
      FBarrier Max Size:       32                                 
*******                  
Agent 3                  
*******                  
  Name:                    aie2                               
  Uuid:                    AIE-XX                             
  Marketing Name:          AIE-ML                             
  Vendor Name:             AMD                                
  Feature:                 AGENT_DISPATCH                     
  Profile:                 BASE_PROFILE                       
  Float Round Mode:        NEAR                               
  Max Queue Number:        1(0x1)                             
  Queue Min Size:          64(0x40)                           
  Queue Max Size:          64(0x40)                           
  Queue Type:              SINGLE                             
  Node:                    0                                  
  Device Type:             DSP                                
  Cache Info:              
    L2:                      2048(0x800) KB                     
    L3:                      32768(0x8000) KB                   
  Chip ID:                 0(0x0)                             
  ASIC Revision:           0(0x0)                             
  Cacheline Size:          0(0x0)                             
  Max Clock Freq. (MHz):   0                                  
  BDFID:                   0                                  
  Internal Node ID:        0                                  
  Compute Unit:            0                                  
  SIMDs per CU:            0                                  
  Shader Engines:          0                                  
  Shader Arrs. per Eng.:   0                                  
  WatchPts on Addr. Ranges:0                                  
  Memory Properties:       
  Features:                AGENT_DISPATCH
  Pool Info:               
    Pool 1                   
      Segment:                 GLOBAL; FLAGS: KERNARG, COARSE GRAINED
      Size:                    131136832(0x7d0fd40) KB            
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:4KB                                
      Alloc Alignment:         4KB                                
      Accessible by all:       TRUE                               
    Pool 2                   
      Segment:                 GLOBAL; FLAGS: COARSE GRAINED      
      Size:                    65536(0x10000) KB                  
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:0KB                                
      Alloc Alignment:         4KB                                
      Accessible by all:       TRUE                               
    Pool 3                   
      Segment:                 GLOBAL; FLAGS: COARSE GRAINED      
      Size:                    131136832(0x7d0fd40) KB            
      Allocatable:             TRUE                               
      Alloc Granule:           4KB                                
      Alloc Recommended Granule:4KB                                
      Alloc Alignment:         4KB                                
      Accessible by all:       TRUE                               
  ISA Info:                
*** Done ***

and

czanik@fedora:~$ rocm-clinfo 
Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3649.0)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_amd_event_callback 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon 8060S Graphics
  Device Topology:				 PCI[ B#197, D#0, F#0 ]
  Max compute units:				 20
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 2900Mhz
  Address bits:					 64
  Max memory allocation:			 57070749280
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 8
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 16384
  Max image 3D height:				 16384
  Max image 3D depth:				 8192
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 Yes
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 128
  Cache size:					 32768
  Global memory size:				 67142057984
  Constant buffer size:				 57070749280
  Max number of constant args:			 8
  Local memory type:				 Local
  Local memory size:				 65536
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1236174432
  Max global variable size:			 57070749280
  Max global variable preferred total size:	 67142057984
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 32
  Error correction support:			 0
  Unified memory for Host and Device:		 1
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 0x7ffb97d11d80
  Name:						 gfx1151
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3649.0 (HSA1.1,LC)
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 
  Extensions:					 cl_khr_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_media_ops cl_amd_media_ops2 cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_depth_images cl_amd_copy_buffer_p2p cl_amd_assembly_program

Output from llama

root@fedora:~# llama-cli   -m ~/models/llama-2-7b.Q4_K_M.gguf   --no-mmap   -ngl 99   -p "Explain quantum computing in simple terms:"   -n 256
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
  Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32
build: 0 (unknown) with HIP version: 6.4.43484-9999 for x86_64-redhat-linux-gnu
main: llama backend init
main: load the model and apply lora adapter, if any
llama_model_load_from_file_impl: using device ROCm0 (Radeon 8060S Graphics) - 64031 MiB free
llama_model_loader: loaded meta data with 19 key-value pairs and 291 tensors from /root/models/llama-2-7b.Q4_K_M.gguf (version GGUF V2)
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv   0:                       general.architecture str              = llama
llama_model_loader: - kv   1:                               general.name str              = LLaMA v2
llama_model_loader: - kv   2:                       llama.context_length u32              = 4096
llama_model_loader: - kv   3:                     llama.embedding_length u32              = 4096
llama_model_loader: - kv   4:                          llama.block_count u32              = 32
llama_model_loader: - kv   5:                  llama.feed_forward_length u32              = 11008
llama_model_loader: - kv   6:                 llama.rope.dimension_count u32              = 128
llama_model_loader: - kv   7:                 llama.attention.head_count u32              = 32
llama_model_loader: - kv   8:              llama.attention.head_count_kv u32              = 32
llama_model_loader: - kv   9:     llama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  10:                          general.file_type u32              = 15
llama_model_loader: - kv  11:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  12:                      tokenizer.ggml.tokens arr[str,32000]   = ["<unk>", "<s>", "</s>", "<0x00>", "<...
llama_model_loader: - kv  13:                      tokenizer.ggml.scores arr[f32,32000]   = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv  14:                  tokenizer.ggml.token_type arr[i32,32000]   = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
llama_model_loader: - kv  15:                tokenizer.ggml.bos_token_id u32              = 1
llama_model_loader: - kv  16:                tokenizer.ggml.eos_token_id u32              = 2
llama_model_loader: - kv  17:            tokenizer.ggml.unknown_token_id u32              = 0
llama_model_loader: - kv  18:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:   65 tensors
llama_model_loader: - type q4_K:  193 tensors
llama_model_loader: - type q6_K:   33 tensors
print_info: file format = GGUF V2
print_info: file type   = Q4_K - Medium
print_info: file size   = 3.80 GiB (4.84 BPW) 
load: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect
load: special tokens cache size = 3
load: token to piece cache size = 0.1684 MB
print_info: arch             = llama
print_info: vocab_only       = 0
print_info: n_ctx_train      = 4096
print_info: n_embd           = 4096
print_info: n_layer          = 32
print_info: n_head           = 32
print_info: n_head_kv        = 32
print_info: n_rot            = 128
print_info: n_swa            = 0
print_info: is_swa_any       = 0
print_info: n_embd_head_k    = 128
print_info: n_embd_head_v    = 128
print_info: n_gqa            = 1
print_info: n_embd_k_gqa     = 4096
print_info: n_embd_v_gqa     = 4096
print_info: f_norm_eps       = 0.0e+00
print_info: f_norm_rms_eps   = 1.0e-05
print_info: f_clamp_kqv      = 0.0e+00
print_info: f_max_alibi_bias = 0.0e+00
print_info: f_logit_scale    = 0.0e+00
print_info: f_attn_scale     = 0.0e+00
print_info: n_ff             = 11008
print_info: n_expert         = 0
print_info: n_expert_used    = 0
print_info: causal attn      = 1
print_info: pooling type     = 0
print_info: rope type        = 0
print_info: rope scaling     = linear
print_info: freq_base_train  = 10000.0
print_info: freq_scale_train = 1
print_info: n_ctx_orig_yarn  = 4096
print_info: rope_finetuned   = unknown
print_info: model type       = 7B
print_info: model params     = 6.74 B
print_info: general.name     = LLaMA v2
print_info: vocab type       = SPM
print_info: n_vocab          = 32000
print_info: n_merges         = 0
print_info: BOS token        = 1 '<s>'
print_info: EOS token        = 2 '</s>'
print_info: UNK token        = 0 '<unk>'
print_info: LF token         = 13 '<0x0A>'
print_info: EOG token        = 2 '</s>'
print_info: max token length = 48
load_tensors: loading model tensors, this can take a while... (mmap = false)
load_tensors: offloading 32 repeating layers to GPU
load_tensors: offloading output layer to GPU
load_tensors: offloaded 33/33 layers to GPU
load_tensors:        ROCm0 model buffer size =  3820.94 MiB
load_tensors:          CPU model buffer size =    70.31 MiB
..................................................................................................
llama_context: constructing llama_context
llama_context: n_seq_max     = 1
llama_context: n_ctx         = 4096
llama_context: n_ctx_per_seq = 4096
llama_context: n_batch       = 2048
llama_context: n_ubatch      = 512
llama_context: causal_attn   = 1
llama_context: flash_attn    = 0
llama_context: freq_base     = 10000.0
llama_context: freq_scale    = 1
llama_context:  ROCm_Host  output buffer size =     0.12 MiB
llama_kv_cache_unified:      ROCm0 KV buffer size =  2048.00 MiB
llama_kv_cache_unified: size = 2048.00 MiB (  4096 cells,  32 layers,  1 seqs), K (f16): 1024.00 MiB, V (f16): 1024.00 MiB
llama_kv_cache_unified: LLAMA_SET_ROWS=0, using old ggml_cpy() method for backwards compatibility
llama_context:      ROCm0 compute buffer size =   288.00 MiB
llama_context:  ROCm_Host compute buffer size =    16.01 MiB
llama_context: graph nodes  = 1158
llama_context: graph splits = 2
common_init_from_params: setting dry_penalty_last_n to ctx_size = 4096
common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
main: llama threadpool init, n_threads = 16

system_info: n_threads = 16 (n_threads_batch = 16) / 32 | ROCm : NO_VMM = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : LLAMAFILE = 1 | REPACK = 1 | 

sampler seed: 2232334333
sampler params: 
	repeat_last_n = 64, repeat_penalty = 1.000, frequency_penalty = 0.000, presence_penalty = 0.000
	dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 4096
	top_k = 40, top_p = 0.950, min_p = 0.050, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.800
	mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000
sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist 
generate: n_ctx = 4096, n_batch = 2048, n_predict = 256, n_keep = 1

 Explain quantum computing in simple terms: what is it, how does it work, and what are its potential benefits?
This is a difficult question to answer because quantum computing is not yet a well-defined field of study, and many of the potential applications are still being researched. However, we can say that quantum computing is a type of computation that relies on the principles of quantum mechanics (the branch of physics that describes the behaviour of particles such as electrons and photons).
These particles obey a set of rules that are different from those obeyed by classical computers, which rely on the principles of classical mechanics. Quantum computing uses a particle’s quantum state (such as its spin) to store information. This means that quantum computers can perform computations that are not possible on classical computers.
In the simplest terms, quantum computing is a type of computation that takes advantage of the unique properties of quantum mechanics. These properties include superposition, entanglement, and non-locality. Superposition is the ability of a quantum system to exist in multiple states simultaneously.
This means that a quantum system can be in two different places at the same time, or have two different properties at the same time. Entanglement is the ability of two quantum systems to be inter

llama_perf_sampler_print:    sampling time =       4.27 ms /   265 runs   (    0.02 ms per token, 62075.43 tokens per second)
llama_perf_context_print:        load time =     631.46 ms
llama_perf_context_print: prompt eval time =      63.57 ms /     9 tokens (    7.06 ms per token,   141.57 tokens per second)
llama_perf_context_print:        eval time =    7110.09 ms /   255 runs   (   27.88 ms per token,    35.86 tokens per second)
llama_perf_context_print:       total time =    7184.25 ms /   264 tokens

Closing words

These are just my first steps. Most of the time I was not even fully aware what I was doing, just reused some sample command lines and code. These experiments were good enough to see that AI works on Linux as well, not just on Windows.

My new toy: FreeBSD on the HP Z2 mini revisited

Thu, 19 Mar 2026 10:43:33 +0100

Last week, I wrote about my initial FreeBSD experiences on my new toy, an AI workstation from HP. FreeBSD runs lightning fast on it, but the desktop was somewhat problematic. Well, I made lots of improvements this week!

A bit of debugging

While there are still some rough edges, there have been tons of improvements since last week. I do not have plans to use FreeBSD on the desktop in the long term, but still, I just could not believe that the FreeBSD GUI is this problematic on this device. I did some experimentation though and it helped a lot… :-)

The initial problem I realized while browsing the output of dmesg was that desktop-installer enabled the wrong kernel modules repository for me. The line leading there was this:

KLD amdgpu.ko: depends on kernel - not available or version mismatch

The next problem occurred when I fixed this problem: there was a kernel panic on boot, when amdgpu.ko was loaded.

I did a fresh FreeBSD install and instead of using the latest packages, I decided to go with the quarterly packages. This way, the desktop installer configured the right kmod repo – however, loading amdgpu.ko still caused a kernel panic. Another experiment I made was using the ATI driver instead of AMD. The installer says that AMD is for modern cards, and ATI is for older ones. Well, as it turned out, even if the chip is barely half a year old, it counts as “old”… :-)

I am still not convinced that proper hardware-based acceleration works: both X.org logs and the GNOME “About” page showed software rendering. However, I had no problem with graphics performance: TuxRacer worked perfectly well… :-) And the GNOME desktop also worked nicely and as stable, including video playback. The only pain point when using GNOME was that screen locking still did not work.

KDE to the rescue

Even if it’s just software rendering, the graphics problem seems to be resolved. However, the screen locking problem still bothered me, as I’m an IT security guy with a healthy dose of paranoia (which means that I lock my screen even when I’m home alone… :-)).

So even if I haven’t tried KDE for the past 5+ years, I gave it a try now. After so many years on XFCE and GNOME, the interface looks a bit weird. However, everything I tried on it seems to work just fine, including screen locking.

KDE on FreeBSD

Central log collection - more than just compliance

Wed, 18 Mar 2026 16:10:37 +0100

What are logs and what is central log collection?

Most operating systems and applications keep track of what they are doing. They write log messages. A syslog message might look similar:

Mar 16 13:13:49 cent sshd[543817]: Accepted publickey for toor from 192.168.97.14 port 58246 ssh2: RSA SHA256:GeGHdsl1IZrnTniKUxxxX4NpP8Q

Applications might store their logs separately and have their own log format, like this Apache access log:

192.168.0.164 - - [16/Mar/2026:13:17:01 +0100] "HEAD /other/syslog-ng-insider-2026-03-4110-release-opensearch-elasticsearch/ HTTP/1.1" 200 3764 "-" "SkytabBot/1.0 (URL Resolution)"

Central log collection simply means that log messages are collected to a central location instead or in addition to saving them locally.

In this blog we take a look at what ease of use, availability, and security of central log collection mean for you.

Ease of use

If you have a single computer in your organization, finding a log message about an event on that computer takes some time. Once you have 2 computers, you have to check 2 computers to find that event. It might take twice as much time, but it is still easier than implementing central log collection. Not to mention, which one is the central computer. :-)

Once you have a network of 10 computers, logging in to each of them to find a log message about an event becomes a huge overhead. It is still doable, but implementing central log collection is a lot easier already in the short term, than looking at the logs on the machines where they were created.

On a network of 100 computers, it is practically impossible to find relevant logs by security or operations, unless logs are collected centrally.

Availability

Collecting logs centrally means that log messages are available even when the sending machine is down. If you want to know what happened, you do not have to get the machine up and running again, but you can check the logs at the central location. If you see signs of a hardware failure, you can go with a spare part immediately, reducing the time and effort needed to repair the machines.

Security

When a computer is compromised, log messages are often altered or deleted completely. However, this tactic only works with logs stored locally. Collecting logs at a central location allows you to use the unmodified logs and to figure out how the compromise happened.

What is next?

It is time to introduce central logging to your organization if you have not yet done it yet. Of course I am a bit biased, but syslog-ng is the perfect tool to do so. You can get started by reading / watching the syslog-ng tutorial on https://peter.czanik.hu/posts/syslog-ng-tutorial-toc/.

syslog-ng logo

Originally published at https://www.syslog-ng.com/community/b/blog/posts/central-log-collection—more-than-just-compliance

My new toy: AI first steps with the HP Z2 Mini

Tue, 17 Mar 2026 14:40:33 +0100

In the past few weeks, I installed five different operating systems on my latest toy: an AI workstation from HP. I love playing with OSes, but my main goal with the new machine is to learn various aspects of AI. I took my first steps in this adventure on Windows.

Of course, you might ask: why on Windows? Well, it’s easy: because it’s easy… :-) There is nothing to install or configure there, as Windows has multiple built-in apps that support AI and can utilize the NPU (hardware-accelerated AI) support of the AMD Ryzen 395 chip. The Windows Task Manager application already has a graph for NPU utilization and I found good use of it already: the Task Manager was open on the NPU tab while I conducted my experiments, allowing me to verify that it was indeed in use, as I ran a couple of generative AI tasks.

I find it somewhat surprising, but Paint now also comes with integrated AI support. Of course, you must pay for image generation but generating simple graphics (or “stickers”, as they are called in Paint) is free. If the prompt field is empty, its prompt text suggests “a cat wearing sunglasses”, so that’s what I generated. It has no practical use for me, but looks fun regardless… :-) You can see an example in the first figure.

Windows Paint with AI

It’s only slightly crazier that the simplest Windows app, Notepad, now also has AI features. I asked Notepad to write me a three-paragraph-long blog about writing a blog with AI in a boring, formal style. It completed the task successfully: the result is pretty boring and full of bullshit. Here it is:

The integration of artificial intelligence into the realm of content creation, particularly in the sphere of blog writing, presents a multitude of advantages that merit serious consideration. Firstly, AI-driven tools offer an unprecedented level of efficiency, allowing content creators to produce substantial volumes of text without the need for exhaustive manual input. This technological boon is particularly beneficial for those with time constraints or heavy workloads, as it liberates them from the mundane task of typing and editing.

Moreover, the precision and consistency of AI in language generation cannot be understated. By leveraging advanced algorithms, AI systems can maintain a uniform tone and style throughout a piece, which is a common challenge for human writers. This consistency ensures that the content is cohesive and that the voice of the piece remains unaltered, thereby enhancing the overall reading experience.

Lastly, the ability of AI to learn from vast datasets and adapt to various writing styles is a significant asset. It empowers writers to explore new topics and styles with confidence, knowing that the AI can provide a solid foundation upon which to build. This adaptability not only streamlines the writing process but also encourages creativity and innovation in content creation.

Windows Notepad with AI

I also wanted to try a controversial Windows feature: Recall. Well, it does not work. When I started it, I got a nice error message stating that it needs Secure Boot. Linux requires it to be turned off, so I cannot test it now. But I must admit that I do not mind that… :-)

Windows Recall needs Secure Boot

If everything goes well, I’ll make my first steps next week to enable hardware-accelerated AI under Linux.