LLVM
PulseAugur coverage of LLVM — every cluster mentioning LLVM across labs, papers, and developer communities, ranked by signal.
8 day(s) with sentiment data
-
Alibaba's XuanTie C950 RISC-V processor wins major innovation award
Alibaba's DAMO Academy XuanTie C950 processor has received the "China Chip" Annual Major Innovation Breakthrough Product Award. This award recognizes the XuanTie C950's significant advancements in high-performance RISC-…
-
Game scripting language daslang achieves 5x faster speech-to-text than llama.cpp
A game scripting language called daslang, originally developed at Gaijin Entertainment, has been enhanced with an inference engine named dasllama. This new engine reportedly achieves speech-to-text transcription speeds …
-
Windows 11 readies Intel's APX for 2027 performance boost
Microsoft has introduced an experimental setting in Windows 11 for Intel's Advanced Performance Extensions (APX), a feature designed to boost application performance. While no current chips support APX, Intel announced …
-
Mojo 1.0 released, aiming to unify Python prototyping with high-performance deployment
Mojo, a programming language developed by Modular and co-founded by Chris Lattner, has reached version 1.0, offering Python-like syntax with systems-level performance and memory safety. This release aims to address the …
-
Ovis: English-based programming language compiles to native code
Ovis is a novel programming language that allows users to write code in plain English, which is then compiled into native executables using LLVM. The language aims to bridge the gap between natural language and high-per…
-
Software complexity concerns spark fears of a "Longhorn Reset"
A user on Mastodon expressed frustration with the perceived "slop" or unnecessary complexity in core software components like Linux, systemd, and compilers such as GCC and LLVM. They suggested that this issue is widespr…
-
AI code resurrects AVX-512 support for Mesa's LLVMpipe driver
AI-generated code is being used to reintroduce support for AVX-512 instructions in Mesa's LLVMpipe software driver. This update aims to improve the performance of the Gallium3D/OpenGL driver on modern systems that suppo…
-
llama.cpp releases GPU fix; new research on random projections published
The llama.cpp project has released version b11847, which includes fixes for bf16 WMMA operand vector types on RDNA 3 and RDNA 4 GPUs. This update addresses compilation errors related to short and __bf16 types, ensuring …
-
Amazon refreshes Kindle, new 'Carrie' adaptation, and Rust compiler speedups · 3 sources tracked
Amazon has updated its Kindle e-reader lineup, introducing flush screens on the standard Kindle, Kindle Paperwhite, and Kindle Colorsoft models. The refresh also includes new color options and two page-turning accessori…
-
New framework uses ML to speed up HPC performance autotuning
Researchers have developed a new framework to accelerate the process of performance autotuning in High Performance Computing (HPC) systems. This framework utilizes a machine learning-based ensemble LLVM Intermediate Rep…
-
LLVM debates enabling ClangIR by default, citing build time concerns
LLVM developers are discussing a proposal to enable ClangIR, a new intermediate representation built on MLIR, in the default build configuration. The primary concern raised is the potential increase in build times for d…
-
LLVM developers debate AGENTS.md for AI agent integration
Developers within the LLVM project are discussing the potential integration of an "AGENTS.md" document. This document aims to provide guidance and establish best practices for the development and deployment of AI agents…
-
EvoX Genesis enables autonomous software evolution with persistent projects · 4 sources tracked
A new research paper introduces EvoX Genesis, a system that organizes software development around a persistent project rather than persistent agents. This approach allows for long-horizon development, demonstrated by th…
-
AI's complexity risks further disconnecting us from technology
The author argues that the allure of AI is leading humanity to further complicate its technology stack rather than simplifying it by returning to fundamental principles. This increasing complexity, even within open-sour…
-
GCC adopts policy against LLM-generated code contributions
The GCC steering committee has adopted an AI contributions policy, prohibiting "legally significant" contributions that are generated by or derived from LLMs. This policy defines "legally significant" as approximately 1…
-
CRISP framework boosts LVLM efficiency by pruning visual tokens
Researchers have developed CRISP, a novel framework designed to improve the efficiency of Large Vision-Language Models (LVLMs) by pruning visual tokens before they are processed by the language model. This two-stage app…
-
AMD GFX1250 AI Accelerator Details Emerge from LLVM Submissions
Analysis of LLVM compiler submissions reveals details about AMD's upcoming GFX1250 architecture, codenamed MI455X, designed for AI acceleration. This architecture reportedly merges RDNA4's software model with CDNA's per…
-
Fortran code gains automatic differentiation via LFortran and Enzyme
Researchers have developed a method to enable automatic differentiation for legacy Fortran code, allowing it to be integrated into modern machine learning frameworks like JAX and PyTorch. This approach uses LFortran to …
-
New framework ANYPoC boosts LLM bug detection with executable proof-of-concept tests
Researchers have developed ANYPoC, a novel framework designed to enhance the reliability of Large Language Model (LLM)-based bug detection in software. This system generates executable proof-of-concept (PoC) tests to va…
-
Developer builds BASIC09 compiler using LLVM and AI assistance
A developer has successfully created a functional BASIC compiler utilizing the LLVM compiler infrastructure. This project specifically targets the BASIC09 dialect, which was originally developed for the OS-9 operating s…