Open Source LLM

Mold 1.1.1 Released With Optimized Memory Usage, New Options

A new version of the high performance, open-source Mold linker is now available with more feature additions and performance optimizations.

Last year marked the release of Mold 1.0 as a high performance linker to take on GNU Gold and LLVM LLD. Mold's development has been led by Rui Ueyama who originally spearheaded LLVM's linker work. Even going back to the v1.0 times, Mold's performance has been exceptional:

Continue reading...

Open Source LLM|Intel

(PR) Intel Releases OpenVINO 2022.1 to Advance AI Inferencing for Developers

Since OpenVINO launched in 2018, Intel has enabled hundreds of thousands of developers to dramatically accelerate AI inferencing performance, starting at the edge and extending to the enterprise and the client. Today, ahead of MWC Barcelona 2022, the company launched a new version of the Intel Distribution of OpenVINO Toolkit. New features are built upon three-and-a-half years of developer feedback and include a greater selection of deep learning models, more device portability choices and higher inferencing performance with fewer code changes


Open Source LLM

Mir 2.7 Released With MirOil For Helping To Migrate Lomiri / Unity 8

Canonical this morning released Mir 2.7 as the newest version of its display stack that is centered around easing Wayland usage for various use-cases.

Arguably most interesting with Mir 2.7 is the introduction of MirOil, which is a support API for helping to migrate Lomiri onto Mir. Lomiri is the UBports Ubuntu Touch project for what was Unity 8. With MirOil, this should help transitioning Lomiri (Unity 8) to work atop a modern Mir stack and in turn helping to get UBports' Ubuntu Touch working atop

Open Source LLM

LLVMpipe/Lavapipe Patches Land To Allow For Greater CPU Saturation

LLVMpipe as the software OpenGL implementation that runs atop the CPU and Lavapipe as the Vulkan equivalent have seen a recent uptick in activity.

The work talked about last week for faster vertex and fragment shader processing have been merged into Mesa 22.1. That is the work by David Airlie he originally started on some two years ago when tackling tessellation shaders but then recently revisited. More recently developers have noticed this work can help boost the ParaView data visualization software performance by 13~67% depending upon the test scene
Continue reading...

Open Source LLM

Mold 1.1 High Performance Linker Brings Native LTO, RISC-V Support

Debuting last December was Mold 1.0 as a high performance, modern linker to compete with GNU's Gold and LLVM's LLD . That project was started by Rui Ueyama who originally worked on LLVM's LLD and has been working aggressively on performance optimizations. Sunday night marked the release of Mold 1.1 as the latest major update to this linker.

Mold 1.0 continued with point releases through a few weeks ago while Mold 1.1 is now ready with larger feature additions. Mold 1.
Continue reading...

Open Source LLM

LLVMpipe Patches Pending For Faster Vertex/Fragment Processing

Red Hat's David Airlie continues carrying out almost magical work on the open-source Linux graphics stack. Reviving some work he originally started a while back, patches pending for Mesa allow speeding up the software-based LLVMpipe OpenGL driver for vertex and fragment shader processing.

Airlie this week wrote a blog post outlining this vertex/fragment processing improvement he started originally two years ago. While workong on OpenGL tessellation shaders for LLVMpipe and noticing the unbearably slow performance out of Uniginee Heaven, he was working on improvements to the scene
Continue reading...

Open Source LLM

Intel OSPray Studio 0.10 Open-Source, Interactive Visualization Software Updated

Intel has released a new version of their open-source, interactive visualization software OSPray Studio that is built atop their OSPray ray-tracing rendering engine.

OSPray Studio is one of countless great open-source projects out of Intel. This Apache 2.0 licensed interactive visualization software continues advancing in-step with the underlying OSPray engine. Intel counts Bentley Motors, VW, and others among their users and collaborators. OSPray allows importing not only glTF models but also UDIM textures, OBJ/MTL data, and plug-ins
Continue reading...

Open Source LLM

Radeon AOMP 14.0-2 Compiler Released For OpenMP GPU Offloading

Along with this week bringing ROCm 5.0 , on Friday AMD engineers released Radeon AOMP 14.0-2 as their downstream of LLVM/Clang focused on carrying the latest patches for OpenMP offloading to Radeon GPUs.

AOMP 14.0-2 is their newest release of this LLVM/Clang downstream that carries AMD's latest GPU offloading patches until they are ready for mainline/upstream integration. Back in November they introduced AOMP 14.0 as they began tracking the LLVM/Clang 14.

Open Source LLM

Europe's AMD-Powered LUMI Supercomputer Continues With Code Porting, Open-Source Tuning

The LUMI supercomputer in Finland is still being assembled with its 2,560 nodes consisting of a 64-core AMD Trento CPU and four AMD Instinct MI250X GPU accelerators per node. This 375+ PFLOPs was supposed to come online by the end of 2021 but was challenged by the supply chain crisis and is now aiming for general availability by the middle of the year. While the hardware is still coming together, their HPC engineers have been hard at work optimizing the
Continue reading...

Open Source LLM

Oracle Linux 8 Now Available From The Microsoft Store

For use with Windows Subsystem for Linux (WSL), Oracle this week released Oracle Linux 8.5 onto the Microsoft Store.

Oracle Linux 8 is joining the likes of Ubuntu, openSUSE, Alpine, Fedora WSL Remix, and other WSL-based offerings available through the Microsoft Store for easy deployment.

Continue reading...