1
Recent Posts
2
Applications and Porting / Re: Firefox WebAssembly (WASM) on ppc64le? Jitsi Meet, Zoom, etc not working
« Last post by tle on April 01, 2026, 11:52:32 pm »I am going to try to make his JIT work run again on the newer version - still in progress. I will NOT touch the WASM in this work
ref: https://github.com/runlevel5/firefox/pull/1
ref: https://github.com/runlevel5/firefox/pull/1
3
Applications and Porting / Re: RustiCL on Fedora 39
« Last post by Hasturtium on March 31, 2026, 02:32:41 pm »Yeah, it's worked for a while, just took some twiddling and manual intervention to prevent it from trying to load Clover.
4
Applications and Porting / Re: RustiCL on Fedora 39
« Last post by tle on March 30, 2026, 05:47:25 pm »It's been awhile, perhaps you should give it another a go
https://www.phoronix.com/news/RadeonSI-Modern-CL-Conformance
https://www.phoronix.com/news/RadeonSI-Modern-CL-Conformance
5
Applications and Porting / Re: [GAMES] Unvanquished is now fully open-sourced
« Last post by tle on March 30, 2026, 05:29:53 pm »Another progress with supporting NaCl sandboxing feature
https://github.com/DaemonEngine/Daemon/pull/1946
That is to use box64 emulation
https://github.com/DaemonEngine/Daemon/pull/1946
That is to use box64 emulation
6
Applications and Porting / Re: [GAMES] UT99, UTGOLD, UT2004 on native PPC64LE!
« Last post by MPC7500 on March 30, 2026, 09:21:52 am »There is AntiDrv.
ReleaseNotes
But there is a Vulkan renderer for Unreal Tournament (UT99) repo.
Quote
* Our patch includes a brand new OpenGL 3.3 / ES 3.0 renderer, AntiDrv. This renderer outperforms the original OpenGLDrv on most platforms, adds new features (e.g., shader-based gamma/color correction), implements/fixes several features that were missing/broken in the original OpenGLDrv (e.g., gradient projectors and weightmap terrain rendering), makes features that were previously nvidia/amd-exclusive available on all platforms (e.g., CO_AlphaBlend_With_Mask combiners), and it can serve as a reference implementation for other new renderers
ReleaseNotes
But there is a Vulkan renderer for Unreal Tournament (UT99) repo.
7
Applications and Porting / Re: [GAMES] UT99, UTGOLD, UT2004 on native PPC64LE!
« Last post by tle on March 29, 2026, 05:15:35 pm »You are right, the game is indeed not CPU bound. It's not clear to me if OldUnreal team has any plan to add Vulkan renderer for UT2004.
8
Operating Systems and Porting / Re: [BUG] Linux 7.0-rc1
« Last post by MauryG5 on March 29, 2026, 08:45:24 am »Great, I'd like to try compiling and installing it on Debian as soon as the stable one is released, which hasn't worked since version 6.2 or later... I'd also like to see how RDNA 4 support is going because I'd like to upgrade to a Radeon Pro, but unfortunately, it seems like no one knows anything about it yet...
9
GPU Compute / Accelerators / Re: The new 9070 XT seems like a really interesting card for Blackbird owners.
« Last post by MauryG5 on March 28, 2026, 04:50:29 pm »Any news on where we're at to use RDNA4 on Power? I think it'll take the usual trifecta: appropriate firmware, updated drivers, and, of course, a recent kernel...
10
Applications and Porting / Re: [GAMES] UT99, UTGOLD, UT2004 on native PPC64LE!
« Last post by SiteAdmin on March 28, 2026, 11:26:07 am »We're glad you are enjoying the preview release! UT2004 has been a favorite on the Raptor side for some time and we are honored to be able to officially enable it on the OpenPOWER platform.
Regarding performance, if you look at the CPU usage while the game is running you will likely see we are not CPU bound in the game itself. This is likely one of the most graphics intensive applications currently available for the OpenPOWER platform, and we suspect we're inadvertently highlighting performance issues in MESA and the Linux kernel on the POWER9 platform. If there is interest in tracking those down in the respective open source components the help would be appreciated -- in fact, anyone that wants to assist in that area can Email Tim directly for better collaboration on the open source side of the stack. It's unclear if we're hitting a GPU bandwidth issue, copying buffers somewhere that we don't need to (MESA/kernel), or if something else entirely is just not optimized for the instruction set.
That all said, the FPS observed is generally within a small percentage of that seen on at least older amd64 systems at similar clocks, so the problem may be more endemic to Linux than the OpenPOWER architecture itself.
Regarding performance, if you look at the CPU usage while the game is running you will likely see we are not CPU bound in the game itself. This is likely one of the most graphics intensive applications currently available for the OpenPOWER platform, and we suspect we're inadvertently highlighting performance issues in MESA and the Linux kernel on the POWER9 platform. If there is interest in tracking those down in the respective open source components the help would be appreciated -- in fact, anyone that wants to assist in that area can Email Tim directly for better collaboration on the open source side of the stack. It's unclear if we're hitting a GPU bandwidth issue, copying buffers somewhere that we don't need to (MESA/kernel), or if something else entirely is just not optimized for the instruction set.
That all said, the FPS observed is generally within a small percentage of that seen on at least older amd64 systems at similar clocks, so the problem may be more endemic to Linux than the OpenPOWER architecture itself.
