Denzil Ferreira :fedora:<p>Ok… so added a few env variables on docker and now it runs… for a few minutes and then it freezes the laptop. Unable to see dmesg or anything as the whole laptop goes dark and image is frozen… <a href="https://techhub.social/tags/amd" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>amd</span></a> <a href="https://techhub.social/tags/rocm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>rocm</span></a></p>
Recent searches
No recent searches
Search options
Only available when logged in.
veganism.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
Veganism Social is a welcoming space on the internet for vegans to connect and engage with the broader decentralized social media community.
Administered by:
Server stats:
293active users
veganism.social: About · Status · Privacy policy
Mastodon: About · Get the app · Keyboard shortcuts · View source code · v4.3.0-alpha.5+vegan
#ROCm
3 posts · 3 participants · 0 posts today
Denzil Ferreira :fedora:<p>Been fighting the whole day trying to get ROCm to play nice with 780M and PyTorch. Using latest <a href="https://techhub.social/tags/rocm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>rocm</span></a> and my laptop just freezes with gfx1103 and using HSA override to 11.0.0 and with 10.3.0 :blobcatknife: </p><p><a href="https://techhub.social/tags/amd" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>amd</span></a> really needs to fix this crap for their GPUs. Using Docker and their provided ROCm images. I know, 780M is not supported. But c’mon, ALL Nvidia cards can run <a href="https://techhub.social/tags/CUDA" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CUDA</span></a> just fine. <a href="https://techhub.social/tags/rant" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>rant</span></a></p>
RenézuCode<p>100 CPU threads & 240GB RAM to make @risc_v <a href="https://chaos.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> @amd <a href="https://chaos.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> and <a href="https://chaos.social/tags/t2linux" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>t2linux</span></a> <a href="https://www.twitch.tv/videos/2421181919" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="">twitch.tv/videos/2421181919</span><span class="invisible"></span></a></p>
pafurijaz<p>It seems that <a href="https://mastodon.social/tags/Vulkan" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Vulkan</span></a> could be the real alternative for using <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> on GPUs or CPUs of any brand, without necessarily having to rely on <a href="https://mastodon.social/tags/CUDA" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CUDA</span></a> or <a href="https://mastodon.social/tags/AMD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AMD</span></a>'s <a href="https://mastodon.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a>. I thought <a href="https://mastodon.social/tags/SYCL" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>SYCL</span></a> was the alternative. This might finally free us from of monopoly <a href="https://mastodon.social/tags/Nvidia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Nvidia</span></a>.<br><a href="https://mastodon.social/tags/Khronos" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Khronos</span></a></p>
Natasha Nox 🇺🇦🇵🇸<p>ffs, why does their docker only support Navi 31 and not Navi 32? 😩 <br><a href="https://hub.docker.com/r/rocm/pytorch" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="">hub.docker.com/r/rocm/pytorch</span><span class="invisible"></span></a></p><p>I just wish both <a href="https://chaos.social/tags/Nvidia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Nvidia</span></a> and <a href="https://chaos.social/tags/AMD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AMD</span></a> would stop with that whole licensing bullshit around <a href="https://chaos.social/tags/CUDA" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CUDA</span></a> and <a href="https://chaos.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> and just include that damn stuff in the default driver.<br>I just want to run <a href="https://chaos.social/tags/Codestral" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Codestral</span></a> on my local machine so I can use it with non-public code. Will be troublesome enough to cram it into 16gb VRAM. 😑 <br><a href="https://chaos.social/tags/computer" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>computer</span></a> <a href="https://chaos.social/tags/Linux" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Linux</span></a> <a href="https://chaos.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a></p>
Michael DiLeo<p>Last night I was up until 2AM trying to get <a href="https://gotosocial.michaeldileo.org/tags/trunas" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>trunas</span></a> <a href="https://gotosocial.michaeldileo.org/tags/amd" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>amd</span></a> drivers installed inside of a <a href="https://gotosocial.michaeldileo.org/tags/docker" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>docker</span></a> <a href="https://gotosocial.michaeldileo.org/tags/container" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>container</span></a> so that <a href="https://gotosocial.michaeldileo.org/tags/ollama" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ollama</span></a> would actually use the <a href="https://gotosocial.michaeldileo.org/tags/gpu" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>gpu</span></a>. I was so close. It sees the gpu, it sees it has 16GB of ram, then it uses the <a href="https://gotosocial.michaeldileo.org/tags/cpu" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>cpu</span></a>.</p><p>Trunas locks down the file system at the root level, so if you want to do much of anything, you have to do it inside of a container. So I made a container for the <a href="https://gotosocial.michaeldileo.org/tags/rocm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>rocm</span></a> drivers, which btw comes to like 40GB in size.</p><p>It's detecting, but I don't know if the ollama container has some missing commands, ie <code>rocm</code> or <code>rocm-info</code>, that it may need.</p><p>Another alternative is one I don't really want, and that's to install either <a href="https://gotosocial.michaeldileo.org/tags/debian" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>debian</span></a> or windows as a VM - windows because I did a test on the application that runs locally in windows on this machine before and it was super fast. It isn't ideal from RAM usage, but I may be able to run the models more easily with the <a href="https://gotosocial.michaeldileo.org/tags/windows" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>windows</span></a> drivers than the <a href="https://gotosocial.michaeldileo.org/tags/linux" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>linux</span></a> ones.</p><p>But anyway, last night was too much of <a href="https://gotosocial.michaeldileo.org/tags/onemoreturn" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>onemoreturn</span></a> for a weeknight.</p>
ℒӱḏɩę :blahaj:<p>The B-17 Bomber was amazing and helped win WWII. I flew on one in 2002 as a tourist - I have family members that were ball turret gunners - bad place to be.</p><p>This video was shot on Hi-8, and thankfully I digitized it (at 720x480) way back in that day. Now, I've up-scaled it with local AI (1408x954) and the improvement is astounding.</p><p>Sadly, this actual B17 crashed in 2019: <a href="https://en.wikipedia.org/wiki/2019_Boeing_B-17_Flying_Fortress_crash" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">en.wikipedia.org/wiki/2019_Boe</span><span class="invisible">ing_B-17_Flying_Fortress_crash</span></a></p><p><a href="https://tech.lgbt/tags/localai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>localai</span></a><br><a href="https://tech.lgbt/tags/stablediffusion" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>stablediffusion</span></a><br><a href="https://tech.lgbt/tags/rocm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>rocm</span></a><br><a href="https://tech.lgbt/tags/amd" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>amd</span></a><br><a href="https://tech.lgbt/tags/b17" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>b17</span></a><br><a href="https://tech.lgbt/tags/flyingfortress" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>flyingfortress</span></a></p>
N-gated Hacker News<p>🌟 Introducing Instella: the "state-of-the-art" open language model that's so open, it comes with more <a href="https://mastodon.social/tags/buzzwords" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>buzzwords</span></a> than actual functionality. ⚙️ Dive into the <a href="https://mastodon.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> rabbit hole, where even Ctrl+K can't save you from the avalanche of self-congratulatory <a href="https://mastodon.social/tags/jargon" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>jargon</span></a>. 🤖✨<br><a href="https://rocm.blogs.amd.com/artificial-intelligence/introducing-instella-3B/README.html" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">rocm.blogs.amd.com/artificial-</span><span class="invisible">intelligence/introducing-instella-3B/README.html</span></a> <a href="https://mastodon.social/tags/Instella" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Instella</span></a> <a href="https://mastodon.social/tags/OpenModel" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>OpenModel</span></a> <a href="https://mastodon.social/tags/Technology" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Technology</span></a> <a href="https://mastodon.social/tags/HackerNews" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>HackerNews</span></a> <a href="https://mastodon.social/tags/ngated" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ngated</span></a></p>
N-gated Hacker News<p>💥 Breaking: Another "groundbreaking" AI tool emerges! Hold onto your hats, folks, because <a href="https://mastodon.social/tags/AITER" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AITER</span></a> is here to revolutionize... wait for it... <a href="https://mastodon.social/tags/reading" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>reading</span></a> <a href="https://mastodon.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> blogs! 📚✨ With more <a href="https://mastodon.social/tags/buzzwords" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>buzzwords</span></a> than lines of code, AITER promises to do something with tensors, or maybe just tensor your patience. 🧠🔧<br><a href="https://rocm.blogs.amd.com/software-tools-optimization/aiter:-ai-tensor-engine-for-rocm™/README.html" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">rocm.blogs.amd.com/software-to</span><span class="invisible">ols-optimization/aiter:-ai-tensor-engine-for-rocm™/README.html</span></a> <a href="https://mastodon.social/tags/AItool" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AItool</span></a> <a href="https://mastodon.social/tags/revolution" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>revolution</span></a> <a href="https://mastodon.social/tags/HackerNews" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>HackerNews</span></a> <a href="https://mastodon.social/tags/ngated" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ngated</span></a></p>
Hacker News<p>Aiter: AI Tensor Engine for ROCm</p><p><a href="https://rocm.blogs.amd.com/software-tools-optimization/aiter:-ai-tensor-engine-for-rocm™/README.html" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">rocm.blogs.amd.com/software-to</span><span class="invisible">ols-optimization/aiter:-ai-tensor-engine-for-rocm™/README.html</span></a></p><p><a href="https://mastodon.social/tags/HackerNews" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>HackerNews</span></a> <a href="https://mastodon.social/tags/Aiter" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Aiter</span></a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/Tensor" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Tensor</span></a> <a href="https://mastodon.social/tags/Engine" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Engine</span></a> <a href="https://mastodon.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> <a href="https://mastodon.social/tags/AMD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AMD</span></a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/Tensor" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Tensor</span></a> <a href="https://mastodon.social/tags/Processing" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Processing</span></a></p>
GripNews<p>🌘 AITER：ROCm的人工智慧張量引擎<br>➤ 使用ROCm的AITER加速AI運算，極大提升效能<br>✤ <a href="https://rocm.blogs.amd.com/software-tools-optimization/aiter:-ai-tensor-engine-for-rocm™/README.html" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">rocm.blogs.amd.com/software-to</span><span class="invisible">ols-optimization/aiter:-ai-tensor-engine-for-rocm™/README.html</span></a><br>ROCm的人工智慧張量引擎（AITER）是AMD的高性能AI運算子庫，提供多樣化的功能和強大的性能優化，助力開發者最大化GPU效能。<br>+ 這篇文章清楚解釋了ROCm的AITER如何提升人工智慧應用的效能，對於想要優化GPU運算的開發者很有參考價值。<br>+ AI運算的性能優化對於未來各行各業的發展至關重要，ROCm的AITER帶來了新的解決方案，讓人工智慧應用更具競爭力。<br><a href="https://mastodon.social/tags/%E4%BA%BA%E5%B7%A5%E6%99%BA%E6%85%A7" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>人工智慧</span></a> <a href="https://mastodon.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> <a href="https://mastodon.social/tags/%E5%BC%B5%E9%87%8F%E5%BC%95%E6%93%8E" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>張量引擎</span></a></p>
Giuseppe Bilotta<p>Even now, Thrust as a dependency is one of the main reason why we have a <a href="https://fediscience.org/tags/CUDA" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CUDA</span></a> backend, a <a href="https://fediscience.org/tags/HIP" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>HIP</span></a> / <a href="https://fediscience.org/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> backend and a pure <a href="https://fediscience.org/tags/CPU" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CPU</span></a> backend in <a href="https://fediscience.org/tags/GPUSPH" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>GPUSPH</span></a>, but not a <a href="https://fediscience.org/tags/SYCL" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>SYCL</span></a> or <a href="https://fediscience.org/tags/OneAPI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>OneAPI</span></a> backend (which would allow us to extend hardware support to <a href="https://fediscience.org/tags/Intel" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Intel</span></a> GPUs). <<a href="https://doi.org/10.1002/cpe.8313" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="">doi.org/10.1002/cpe.8313</span><span class="invisible"></span></a>></p><p>This is also one of the reason why we implemented our own <a href="https://fediscience.org/tags/BLAS" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>BLAS</span></a> routines when we introduced the semi-implicit integrator. A side-effect of this choice is that it allowed us to develop the improved <a href="https://fediscience.org/tags/BiCGSTAB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>BiCGSTAB</span></a> that I've had the opportunity to mention before <<a href="https://doi.org/10.1016/j.jcp.2022.111413" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">doi.org/10.1016/j.jcp.2022.111</span><span class="invisible">413</span></a>>. Sometimes I do wonder if it would be appropriate to “excorporate” it into its own library for general use, since it's something that would benefit others. OTOH, this one was developed specifically for GPUSPH and it's tightly integrated with the rest of it (including its support for multi-GPU), and refactoring to turn it into a library like cuBLAS is</p><p>a. too much effort<br>b. probably not worth it.</p><p>Again, following <span class="h-card" translate="no"><a href="https://peoplemaking.games/@eniko" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@<span>eniko</span></a></span>'s original thread, it's really not that hard to roll your own, and probably less time consuming than trying to wrangle your way through an API that may or may not fit your needs.</p><p>6/</p>
Fishd<p>AI rabbit hole ... I've been playing with Ollama and some stability diffusion tools on my MacBook Pro M2 Max and my Linux desktop ... the desktop is way faster and only has an RX6800 in it, so of course I'm now thinking about an Rx7900XTX ... (I don't do Nvidia cards) ...</p><p>Anyone have experience with this upgrade? Is going from 16gb of VRAM to 24gb going to make a massive difference?</p><p>Using radeontop I can see it's using all 16gb at some points, but not consistently ... and I'm not sure if that's an issue or a feature. I believe <a href="https://infosec.exchange/tags/rocm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>rocm</span></a> still has some issues. </p><p><a href="https://infosec.exchange/tags/selfhosting" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>selfhosting</span></a> <a href="https://infosec.exchange/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ai</span></a> <a href="https://infosec.exchange/tags/sdxl" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>sdxl</span></a> <a href="https://infosec.exchange/tags/ollama" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ollama</span></a></p>
Benjamin Carr, Ph.D. 👨🏻💻🧬<p>Just how deep is <a href="https://hachyderm.io/tags/Nvidia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Nvidia</span></a>'s <a href="https://hachyderm.io/tags/CUDA" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CUDA</span></a> moat really?<br>Not as impenetrable as you might think, but still more than Intel or AMD would like<br>It's not enough just to build a competitive part: you also have to have <a href="https://hachyderm.io/tags/software" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>software</span></a> that can harness all those <a href="https://hachyderm.io/tags/FLOPS" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>FLOPS</span></a> — something Nvidia has spent the better part of two decades building with its CUDA runtime, while competing frameworks for low-level <a href="https://hachyderm.io/tags/GPU" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>GPU</span></a> <a href="https://hachyderm.io/tags/programming" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>programming</span></a> are far less mature like AMD's <a href="https://hachyderm.io/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> or Intel's <a href="https://hachyderm.io/tags/OneAPI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>OneAPI</span></a>.<br><a href="https://www.theregister.com/2024/12/17/nvidia_cuda_moat/" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">theregister.com/2024/12/17/nvi</span><span class="invisible">dia_cuda_moat/</span></a> <a href="https://hachyderm.io/tags/developers" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>developers</span></a></p>
ℒӱḏɩę :blahaj:<p>My little chonky squirrel, courtesy of my solar powered 7900XTX with Stable Diffusion. 2560x1536 resolution. <a href="https://tech.lgbt/tags/localai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>localai</span></a> <a href="https://tech.lgbt/tags/stablediffusion" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>stablediffusion</span></a> <a href="https://tech.lgbt/tags/radeon" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>radeon</span></a> <a href="https://tech.lgbt/tags/rocm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>rocm</span></a> <a href="https://tech.lgbt/tags/amd" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>amd</span></a></p>
butterflyofChick ⏚ꝃ⌁⁂<p>Hello <span class="h-card" translate="no"><a href="https://fosstodon.org/@argosopentech" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@<span>argosopentech</span></a></span> :)</p><p>When installing LibreTranslate or Locomotive from `pip`, it installs a lot of Nvidia stuff whereas I have an AMD Ryzen™ 5 5600G with Radeon™ Graphic.</p><p>Is there something equivalent I can modify to support <a href="https://mstdn.fr/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> :<br><a href="https://rocm.docs.amd.com/projects/install-on-linux/en/develop/install/3rd-party/pytorch-install.html" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">rocm.docs.amd.com/projects/ins</span><span class="invisible">tall-on-linux/en/develop/install/3rd-party/pytorch-install.html</span></a></p>
Oblomov<p>Even better, in the afternoon I managed to find a workaround for my <a href="https://sociale.network/tags/GPGPU" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>GPGPU</span></a> software building but hanging when trying to run it, which seems to be related to an issue with some versions of the <a href="https://sociale.network/tags/AMD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AMD</span></a> software stack and many integrated GPUs, not just the <a href="https://sociale.network/tags/SteamDeck" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>SteamDeck</span></a> specifically. So exporting the HSA_ENABLE_SDMA=0 environment vriable was sufficient to get my software running again. I'm dropping the information here in case others find it useful.</p><p><a href="https://sociale.network/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> <a href="https://sociale.network/tags/GPU" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>GPU</span></a> <a href="https://sociale.network/tags/APU" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>APU</span></a> <a href="https://sociale.network/tags/HIP" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>HIP</span></a></p><p>2/2</p>
Giuseppe Bilotta<p>One of the nice things of the refactoring that I had to do to introduce CPU support is that it also allowed me to trivially had support for <a href="https://fediscience.org/tags/AMD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AMD</span></a> <a href="https://fediscience.org/tags/HIP" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>HIP</span></a> / <a href="https://fediscience.org/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a>.<br>That, and the fact that AMD engineers have written a drop-in replacement for the Thrust library that we depend on in a couple of places. (This is also one of the things that is holding back a full <a href="https://fediscience.org/tags/SYCL" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>SYCL</span></a> port for <a href="https://fediscience.org/tags/GPUSPH" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>GPUSPH</span></a>, BTW.)</p>
Benjamin Carr, Ph.D. 👨🏻💻🧬<p><a href="https://hachyderm.io/tags/ZLUDA" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ZLUDA</span></a> Takes On Third Life: <a href="https://hachyderm.io/tags/OpenSource" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>OpenSource</span></a> Multi-GPU <a href="https://hachyderm.io/tags/CUDA" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CUDA</span></a> Implementation Focused On AI<br>ZLUDA is being rebuilt to focus on multi-GPU vendor support and will take a particular emphasis on <a href="https://hachyderm.io/tags/machinelearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>machinelearning</span></a> / <a href="https://hachyderm.io/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> type workloads. Previously ZLUDA was more focused on enabling professional creator workloads while now it will be more focused on CUDA-based AI/#ML software. The new ZLUDA code will be focused on <a href="https://hachyderm.io/tags/RDNA1" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>RDNA1</span></a> and newer support along with <a href="https://hachyderm.io/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> 6.1+ compute stack support. <br><a href="https://www.phoronix.com/news/ZLUDA-Third-Life" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">phoronix.com/news/ZLUDA-Third-</span><span class="invisible">Life</span></a></p>
Natasha Nox 🇺🇦🇵🇸<p>What would be the optimal (used) GPU for running LLMs and other Tensor / ROCm workload? I'm seeing Nvidia Tesla M10 cards with 32gb VRAM going for about 200€ in ebay. I don't know the differences between the M, P and K series though and if there are any caveats. Any recommendations?<br><a href="https://chaos.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> <a href="https://chaos.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>LLM</span></a> <a href="https://chaos.social/tags/DeepLearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>DeepLearning</span></a> <a href="https://chaos.social/tags/Tensorflow" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Tensorflow</span></a> <a href="https://chaos.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ROCm</span></a> <a href="https://chaos.social/tags/AMD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AMD</span></a> <a href="https://chaos.social/tags/NVidia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>NVidia</span></a></p>
VegansExploreLive feeds
Mastodon is the best way to keep up with what's happening.
Follow anyone across the fediverse and see it all in chronological order. No algorithms, ads, or clickbait in sight.
Create accountLoginDrag & drop to upload