Pinned Tweet
This is collection of out-of-tree #LLVM examples for teaching and learning. I've been working on them over the last few months and they are ready for sharing 🤓🤗
github.com/banach-space/llvm…
Andrzej Warzynski retweeted
If you want an LLM solutions that is deployable across a range of mobile devices - then target the Arm CPU, the most ubiquitous CPU on the planet! Great performance & programmability ensuring quick time to market.
We are only getting started.... stay tuned for more!
#llm #onarm
📢 Generative AI has arrived at the edge on mobile & it runs #onArm CPUs.
But this is just the start.
With smaller, more efficient LLMs, improved performance & software optimization, get ready for the proliferation of gen AI on mobile that runs #onArm: bit.ly/3S2Jmjt
Andrzej Warzynski retweeted
I've published my slides from my talk at #LLVM Developers Meeting 2023 about getting started with LLVM development under Visual Studio on Windows. jvste.ch/vs-windows-llvm-202…
Excited that this is finally available 🤗 Sadly I wasn’t able to watch it live 😫
Recording of my EuroLLVM talk is up 🎉
Compiling Ruby (with MLIR)
youtube.com/watch?v=NfMX-dFM…
Andrzej Warzynski retweeted
We've made the online version of the Low Level Software Security book a lot nicer-looking recently, see llsoftsec.github.io/llsoftse…. Please let us know what you think can be improved further in the book. There's a "report" 💡 button at the top now to make that easy. #LLSoftSecBook
Don’t miss your chance to chat with an #LLVM guru - I’ve learnt a great deal from Kristof!
I'm hosting an LLVM office hours every 2 weeks. Next one is next Wednesday the 23rd, at 9:30am CET, at meet.jit.si/KristofBeylsLLVM…. Come join me to talk about anything #llvm. More details, see llvm.org/docs/GettingInvolve….
Thanks @AartBik ! This is super fun compiler stuff and SVE is particularly close to my heart 🤗 🤩🎉
Replying to @AartBik
One i has been dotted: ArmSVE support is back as well!
discourse.llvm.org/t/mlir-sp…
Thank you for working on this! Lots of folks are eager to try this 👌🏻🤩
I've been working on CMake magic to fully automatically apply BOLT to Clang, similar to how PGO two-stage build is automated with CMake cache file. It appears that fully automatically BOLT-optimized Clang is ~30% faster than the Release (-O3) Clang: reviews.llvm.org/D132975
IMHO, CGDB (cgdb.github.io) is better - it has syntax highlighting and the interface is more intuitive than GDB+TUI.
Adding some ML heuristic in #LLVM has been discussed for ages. Here’s a working example 👇🏻. No spoilers 🤫ai.googleblog.com/2022/07/ml…
That was a very informative intro into concepts and other template magic in C++20, thank you! #must-watch-again
A ton of great ideas/thoughts on making the CS education more accessible, thank you @mohardek! Kudos to @CamBlackwood95 @colin_riddell for hosting 🙏!👇⚓️
anchor.fm/the-coder-career/e…
👇🏻🙌🏻 This reminds me, I’ve got some slides to prepare 🧐🤔🤫
'How compilers work: introduction to LLVM passes'
- Andrzej Warzynski (@_banach_space)
14:00-15:00, Tuesday, 5th July 2022
cpponsea.uk/2022/sessions/ho…
#CMake support for #LLVM #Flang has finally landed (avail in 3.24) 😎🔥👇
gitlab.kitware.com/cmake/cma…
This is quite bold, looking forward to the end result! 🙏🏻👇🏻
This Post is from an account that no longer exists. Learn more
Enjoyed reading this @sheredom, thanks! Legacy PM I still used for the back-end pipelines, so this is relevant 🤔🧐Dumb question - what’s HPC#?
This quoted post is unavailable.
That’s me yesterday! 🤗🙏🏻Please reach out if you’d like to hear more. DM’s always welcome 📝!
Replying to @_banach_space
@_banach_space presents work on the Flang driver at #EuroLLVM. Writing a compiler driver doesn't happen every day, so there's a lot of reverse engineering in the absence of comprehensive documentation.




















