dgilmore changed the topic of #fedora-riscv to: Fedora on RISC-V https://fedoraproject.org/wiki/Architectures/RISC-V || Logs: https://libera.irclog.whitequark.org/fedora-riscv || Alt Arch discussions are welcome in #fedora-alt-arches
Guoguo has joined #fedora-riscv
Guoguo has quit []
hezhengyu has joined #fedora-riscv
hezhengyu has quit [Client Quit]
hezhengyu has joined #fedora-riscv
davidlt has joined #fedora-riscv
davidlt has quit [Remote host closed the connection]
davidlt has joined #fedora-riscv
davidlt has quit [Remote host closed the connection]
davidlt has joined #fedora-riscv
davidlt has quit [Remote host closed the connection]
davidlt has joined #fedora-riscv
davidlt has quit [Remote host closed the connection]
davidlt has joined #fedora-riscv
fuwei has joined #fedora-riscv
tg has quit [Ping timeout: 255 seconds]
tg has joined #fedora-riscv
fuwei has quit [Quit: Konversation terminated!]
bjoto has quit [Ping timeout: 268 seconds]
bjoto has joined #fedora-riscv
bjoto has quit [Ping timeout: 264 seconds]
<davidlt> rwmjones_, I don't think Jakub is looking at GCC PRs
<rwmjones_> morning
<rwmjones_> how do you know?
<rwmjones_> & what do you suggest we do?
<davidlt> rwmjones_, ping him, there are a number of PRs opened and no one looked
<davidlt> rwmjones_, looking at PR history I don't think he merges PR (or did it in recent months)
<rwmjones_> ok I emailed
bjoto has joined #fedora-riscv
<davidlt> this should be already merged in webkit thus you could re-create the patch buy fishing out the right lines from upstream, I think.
<davidlt> ah, it slightly changed.
<davidlt> Again, you might want to double check relative bits with upstream.
<rwmjones_> davidlt: ok I'l ltake a look at these ones
<rwmjones_> what did we decide about luajit? I can't remember if we thought it could be enabled
<davidlt> there is no port
<davidlt> There is Lua (works), but LuaJIt is not merged.
<davidlt> PLTC Labs is working on it, and they have it out-of-tree, but it's not under review for upstream
<rwmjones_> ok
<rwmjones_> let me subscribe to that ...
<davidlt> so work seems to be ongoing, but it's now LuaJIT v3.0 thing
<rwmjones_> urrr pagure is down
<davidlt> well that will be bad
<rwmjones_> can you reach https://src.fedoraproject.org/ ?
<rwmjones_> https://www.fedorastatus.org/ says everything is fine
<davidlt> yes its fine
<rwmjones_> well it's defnitely unroutable from here
<rwmjones_> and we're back
<rwmjones_> qt5-qttools was already merged
<rwmjones_> davidlt: re qt5-qtwebkit
<rwmjones_> there's an upstream, merged PR containing our fix:
<rwmjones_> but weirdly I don't see the change upstream
<rwmjones_> oh wait, I'm on the wrong branch, what's going on ...
<rwmjones_> oh no, actually qtwebkit is obsolete, this is very strange
<davidlt> it was merged
<rwmjones_> right, but which branch!
<rwmjones_> $ git branch --contains d9824ec
<rwmjones_> no output
<rwmjones_> ok apparently it's a remote branch origin/qtwebkit-5.212
<davidlt> build_dependencies_docker (#982) docker_build (#982) qtwebkit-5.212 (#982)
<davidlt> the notes say to look at a new place for qtwebkit: https://github.com/movableink/webkit
<rwmjones_> yeah but that's not the one they're using it seems
<rwmjones_> anyway I think I got this now, thanks
<rwmjones_> interesting thread here about optimizing for vector extension (read the replies too): https://news.ycombinator.com/item?id=39545622
<davidlt> rwmjones_, We're having some trouble serving your request. Sorry!
<davidlt> HN seems to be off
<rwmjones_> oh yeah seems down now :-(
<rwmjones_> try later :-/
<rwmjones_> ok gotta go to the post office, biab
<rwmjones_> i've done all those packages, so if there are any more drop them in here & i'll take a look later
djdelorie has quit [Ping timeout: 260 seconds]
djdelorie has joined #fedora-riscv
* rwmjones_ is liking the AMD 7800 XT so far
<davidlt> rwmjones_, why 7800 XT?
<davidlt> IIRC this's a good value option, but it's not part of ROCm support (yet)
<rwmjones_> ROCm support is pretty solid actually
<rwmjones_> if you have the 6.0.2 version, the 6.0.0 version in Fedora is buggy as hell
<rwmjones_> I bought it because it's a bit more open source than nvidia stuff
<rwmjones_> just for doing stable diffusion, I've not actually plugged a monitor into it so far
<rwmjones_> weird that it's not listed, but it definitely works for me
<davidlt> They only added support for the bigger die
<davidlt> Aren't there "two" stacks, one open source and another one from AMD?
<rwmjones_> don't know, I'm using all free software here though, compiled it all myself
<davidlt> The open source one runs on a lot other even if it's officially not supported.
<davidlt> Yeah, ROCm SMI stuff was merged as it even works with RX 500
<davidlt> But AMD just few months ago officially started supporting consumer GPUs
<davidlt> I guess this mostly related to optimization and testing.
<davidlt> Anyways, AMD has a lot to catch up to Nvidia. Even crappy cards supported CUDA.
<rwmjones_> we got a comment on https://src.fedoraproject.org/rpms/gcc/pull-request/51 .. do you think I should push it?
<davidlt> rwmjones_, should be fine, but would be cool someone to cross-check gcc14-libtool-no-rpath.patch
<rwmjones_> yeah so it needed the bleeding edge rawhide kernel and updated rocmclr to work for me
<rwmjones_> I was getting a lot of random segfaults before that
<davidlt> It's not officially supported, but I so far AMD is adding a new card every few months top - down. So far all 3 7000-series consumer cards are the same die IIRC.
<rwmjones_> the no rpath patch looks fine to me .. let's give it a bit longer to see if we get any more comments
<davidlt> ok
<rwmjones_> I'll push it tomorrow morning if no further feedback
<davidlt> I daily drive now AMD, and SW quality is definitely not that great compared to Intel devices I have.
<rwmjones_> for sure it took me a while to get it working
<davidlt> I understand that AMD is constantly expanding, and pushing things, but thing like VP9 not working properly is annoying.
<davidlt> The white death screen is annoying too from deep sleep state.
<davidlt> There were discussions about power consumption SW vs decoder.
<davidlt> "GPU hangs"
<davidlt> All of my problems are already documented, and on the issue tracker, but these are the same issues within last 3 generations.
<davidlt> I really like the hardware, but SW experience is not the best.
fuwei has joined #fedora-riscv
masami has joined #fedora-riscv
masami has quit [Quit: Leaving]
masami has joined #fedora-riscv
masami has quit [Client Quit]
zsun has joined #fedora-riscv
zsun has quit [Client Quit]
fuwei has quit [Remote host closed the connection]
zsun has joined #fedora-riscv
zsun has quit [Quit: Leaving.]
<davidlt> ocaml land is cookig now
<rwmjones_> this should work fine but let me know if not
amusil has joined #fedora-riscv
amusil has quit [Quit: Client closed]
<davidlt> Krste is doing a webinar, Topic: Creating Agile AI Systems with Open RISC-V Solutions
<davidlt> that incl. "Overview of SiFive X390 and P870 solutions"
nirik has quit [Ping timeout: 264 seconds]
iooi has quit [Read error: Connection reset by peer]
iooi has joined #fedora-riscv
nirik has joined #fedora-riscv
davidlt has quit [Ping timeout: 246 seconds]