1
Operating Systems and Porting / Re: [BUG] Linux 7.0-rc1
« Last post by tle on Today at 07:44:59 pm »That's concerning taking into account that Fedora 44 will likely ship this version.
Recent Posts
2
Operating Systems and Porting / [BUG] Linux 7.0-rc1
« Last post by MPC7500 on March 05, 2026, 12:08:22 pm »3
General OpenPOWER Discussion / Queensland Health, IBM's 12 year ban and the future of POWER
« Last post by pocock on March 02, 2026, 05:45:40 pm »I put up a fresh blog about Queensland Health's IBM fiasco. The whole company, including POWER components, was subject to a ban for 12 years in the state of Queensland public sector.
It is worth reading the official report and thinking about how it translates to behaviour in the Linux ecosystem.
Has anybody got any feeling about the future of the POWER platform?
Are any other vendors likely to come forward with a 100% open CPU like the POWER9?
It is worth reading the official report and thinking about how it translates to behaviour in the Linux ecosystem.
Has anybody got any feeling about the future of the POWER platform?
Are any other vendors likely to come forward with a 100% open CPU like the POWER9?
4
Open Source and Libre Systems / Forum should be called Threads Daniel Pocock starts and then locks
« Last post by TimKelly on February 28, 2026, 04:36:13 am »Daniel, if you are going to keep starting threads and then locking them when people engage, we might get the idea you don't really want to hear other sides or viewpoints.
5
Open Source and Libre Systems / Re: Open source aspirations of credibility exploited to legitimize Bitcoin
« Last post by pocock on February 28, 2026, 02:47:29 am »
I started putting some of my observations in blog posts, this one begins to cover the relationship with open source phenomena
I don't have a comment overall, but I find it ironic that very early on in his post, Daniel posts an 2011 email from the debian-private mailing list in which Philip Hands says "Likewise IANAL, but as I understand it, if one does not defend a trademark, one risks losing it."
If you accept Bitcoin from anybody who wants to use your trademark and then the Bitcoins lose all their value, what was the point?
A reminder to the readers that after Daniel had been expelled from Debian and stripped of his developer status,
I resigned from some voluntary activities at a time when I lost two family members. People spreading rumours about relationships with interns are lying, please read the proof about the relationships
There is nothing in copyright law to say people can expel each other. If authors stop working together they still have to recognise the status of their peers. Not respecting the status of a co-author is not expulsion, it is plagiarism.
he registered multiple domains with "debian" in them, forcing Debian to file suit against Daniel to defend their trademark (Daniel was ordered to turn over the domains).
DNSLytics found over 2,850 domain names containing the Debian trademark. They did not start a trademark dispute about any of those domains. They only attacked approximately a dozen domains registered by me because they want to censor debian-private emails revealing the the real history of Debian
Daniel states Debian has spent $120,000 in legal fees as a result of his actions.
In every serious development team I've worked in, if you have a dispute between people, you have a meeting and you look each other in the eye and you sort it out.
Debian didn't spend $120,000 because of my actions. They spent $120,000 to avoid looking me in the eye. They spent $120,000 to avoid apoligising to my family. They spent $120,000 because they want other developers to be afraid of them in future. They hope they can give people orders and people will obey them out of fear. That is modern slavery.
FSFE received a bequest of EUR 150,000. That is money that probably was meant for the real FSF. The Fellowship had elected me as their representative and I didn't want to see the money wasted so I tried to provide some constructive suggestions about how to use that money. The FSFE lost over $500,000 because they didn't listen to when we tried to discuss it
6
GPU Compute / Accelerators / Re: Intel Arc A770 - failed experiment
« Last post by Hasturtium on February 26, 2026, 09:54:49 am »Booted into a fresh install of Fedora 43 after patching, and it is not currently supported there via xe or i915 as far as I can tell. lspci enumerated it correctly and ID'd it as Intel DG2 [Arc A380], but there was no acceleration, no identification of its audio, and several libraries I'd expect Fedora to host for Intel GPUs aren't for ppc64le. So I'd guess even the xe driver has not yet reached mature status on ppc64le. Looks like my Radeon W6600 will continue to perform... not that I've pushed it terribly hard in that machine outside of running 30 year old open source games at 4K with antialiasing, or the odd bout of Quake II RTX.
7
Open Source and Libre Systems / Re: Open source aspirations of credibility exploited to legitimize Bitcoin
« Last post by TimKelly on February 26, 2026, 06:35:13 am »
I started putting some of my observations in blog posts, this one begins to cover the relationship with open source phenomena
I don't have a comment overall, but I find it ironic that very early on in his post, Daniel posts an 2011 email from the debian-private mailing list in which Philip Hands says "Likewise IANAL, but as I understand it, if one does not defend a trademark, one risks losing it."
A reminder to the readers that after Daniel had been expelled from Debian and stripped of his developer status, he registered multiple domains with "debian" in them, forcing Debian to file suit against Daniel to defend their trademark (Daniel was ordered to turn over the domains). Daniel states Debian has spent $120,000 in legal fees as a result of his actions.
8
Applications and Porting / Re: Discussion about Blender Cycles on PPC64
« Last post by tle on February 26, 2026, 02:15:25 am »9
Applications and Porting / Re: Discussion about Blender Cycles on PPC64
« Last post by MPC7500 on February 25, 2026, 08:50:38 pm »10
Applications and Porting / Discussion about Blender Cycles on PPC64
« Last post by tle on February 25, 2026, 07:15:30 pm »I was sitting down with a friend of mine who was a professional 3D modeler. By observing him working with Blender on my Blackbird workstation for a simple modeling job of a snow man, I realize the biggest blocker that push Blender to the finish line is the lacking of GPU acceleration rendering, specifically the Blender Cycles engine.
As you might have known that AMD ROCm is huge and complex so I rule out any possibility of AMD ever support PPC64. Other engines like CUDA is out of question because NVIDIA does not even offer driver for PPC64. The same goes with Intel unfortunately. The only option is OpenCL which is vendor-neutral but sadly its non-CUDA closeness design and buggy implementation forced Blender team to drop OpenCL few years ago.
What are options out there? LuxCoreRender is one potential solution because it supports OpenCL. In theory we could achieve it like this: LuxCoreRender → OpenCL API → RustiCL (Mesa) → radeonsi (Gallium) → amdgpu (kernel) → AMD GPU. I am unsure how reliable RustiCL is so this approach might be not working at all. However let's assume that it may work, there is still a huge effort to get LuxCore running on PPC64.
I am wondering if any community member has attempted to tackle this blocker issue before.
As you might have known that AMD ROCm is huge and complex so I rule out any possibility of AMD ever support PPC64. Other engines like CUDA is out of question because NVIDIA does not even offer driver for PPC64. The same goes with Intel unfortunately. The only option is OpenCL which is vendor-neutral but sadly its non-CUDA closeness design and buggy implementation forced Blender team to drop OpenCL few years ago.
What are options out there? LuxCoreRender is one potential solution because it supports OpenCL. In theory we could achieve it like this: LuxCoreRender → OpenCL API → RustiCL (Mesa) → radeonsi (Gallium) → amdgpu (kernel) → AMD GPU. I am unsure how reliable RustiCL is so this approach might be not working at all. However let's assume that it may work, there is still a huge effort to get LuxCore running on PPC64.
I am wondering if any community member has attempted to tackle this blocker issue before.
