Cool article on our work on LLM inference with TornadoVM! Now, GPULlama3.java supports also Mistral models and soon Gwen3 TornadoVM

The University of Manchester team extended Alfonso² Peterssen☕'s Llama3.java (github.com/mukel/llama3.j…) with GPU support via the GraalVM JIT and TornadoVM: infoq.com/news/2025/06/g… Nicely done!👏

Thanks to Alfonso² Peterssen☕ for the llama2.java port and Roger Allen (github.com/rogerallen/lla…) for the CUDA port, we were able to understand and shape TornadoVM for running LLMs and transformers architectures.

New blogpost: How to enable NVIDIA Nsight Compute CLI in Fedora/Linux. 🔗jjfumero.github.io/posts/2025/07/… #profiling #cuda #configuration

I had great time discussing with Adam Bien our latest work on #GPULlama3.java and how we used TornadoVM to have GPU inference enabled in Java for multiple models

New TornadoVM release! 🙌 The new optimizations for efficient data movements are key for long-running applications with multiple kernels, such as LLM inference.

A must-listen for Java developers! My colleague Michalis Papadimitriou joined Adam Bien to discuss how #TornadoVM brings GPU acceleration to Java-based LLM inference. If you're into high-performance Java, this one's for you! #Java #GPUs #LLM #AI #Performance #TornadoVM

SiPearl’s Rhea1 has now reached tape-out and will equip the CPU cluster of JUPITER, Europe’s first #exascale supercomputer. 🚀 We are proud to be part of this journey. Press release: sipearl.com/wp-content/upl… #SiPearl #Rhea1 #EuroHPC #EU #cloud HaDEA EUCloudEdgeIoT

Docker images now available for TornadoVM v1.1.1! See here how to run: 👉github.com/beehive-lab/do… These images support #NVIDIA #GPUs or #Intel #GPUs, making it easier than ever to get started with heterogeneous computing using #TornadoVM. #containers #Java #acceleration

Huge congratulations to the Sednai team for their groundbreaking work on GPU-accelerated vector search with pgvector, showcased at Swiss PGDay! 👏 Slides: 👉pgday.ch/common/slides/… #PostgreSQL #GPU #OpenSource Horizon Europe 🇪🇺 HaDEA SBFI SEFRI SERI EUCloudEdgeIoT

Amazing work by Stratos Psomadakis Stratos Psomadakis and the others@CSLab CSLab@NTUA! Proud and privileged to work with them!

The VMIL paper deadline is next week. Please submit abstracts already. We are looking forward to your work on language implementation. Don't have results yet, but would like feedback on an experimental design? We're looking forward to those, too! 🧑‍🔬🧪💻 conf.researchr.org/home/icfp-spla…

Congratulations to our colleagues from CSLab@NTUA ICCS - NTUA for winning the Best Paper Award at #HotStorage25! 🥇🎉 SnapBPF is an eBPF-based snapshot prefetcher for VM-based serverless functions - developed within our project. 🇪🇺 Well deserved! 👏 HaDEA EUCloudEdgeIoT

"WebAssembly / Wasm and Java"--an airhacks.fm podcast conversation with Fabio Niephaus is ready to listen: 👉 adambien.blog/roller/webasse… #java #airhacks #podcast #wasm #graalvm

🔧 Today, TANGO joins MobiSpaces and NeMo.Bil for a Standardisation Co-Creation Workshop – exchanging insights and aligning efforts to advance interoperability across mobility data spaces. 🎥 Stay tuned! #TANGOProject #HorizonEurope #Standardisation #DataSpaces

Devoxx Belgium 2025 : Java + AI + Agents Certainly the biggest European AI conference using Java 🔥 #Devoxx #AI #Java #Agents

In October we will give a deep dive talk about #TornadoVM, an #opensource #Java technology for hardware acceleration at Devoxx! Happy to be back at #DevoxxBE! Join us to learn more! Registration will open soon! AERO P2CODE

📢 Super exciting news! Two talks on #TornadoVM one of the technologies used and extended in our project will be part of the #DevoxxBE program in October! Congratulations to the #TornadoVM team from the Department of Computer Science, The University of Manchester! 👏 HaDEA

If you give a try to the LangChain4j agentic AI module please feel free to share any kind of feedback. It is totally new, very likely there is a lot of room for improvements, and we are eager to make it as good as possible.

Really enjoyed the conversation with Adam We dove into the latest TornadoVM features that make GPU-powered LLM inference in Java not just possible, but fast! With the new APIs, TornadoVM users can now fine-tune GPU data management just like CUDA developers but entirely in Java.