NVIDIA released the first preview build of its CUDA toolkit for the RTX Spark platform as a native Windows-on-Arm software toolkit, enabling developers to perform native platform optimization for NVID…
Triton is a compiler with a Python frontend that parses a function's AST, runs it through an MLIR pipeline, and emits a GPU binary, never executing the Python function as Python. The compiler handles …
A developer has built Zamin, a scripting language with a Rust-based bytecode VM and GPU-accelerated machine learning. The language compiles to bytecode, features a register-based VM, and includes a ma…
A developer known as PyGo has built a deep learning framework where Go is the top-level API, Python handles autograd and the model zoo, and C++/CUDA does the raw compute. The architecture embeds CPyth…
A developer built Aakaar, a deep learning framework from scratch using native C++ and CUDA, wrapped in Python. In a benchmark against PyTorch on EMNIST, Aakaar achieved slightly faster runtime while m…
Alibaba's chip design unit T-Head open-sourced SAIL, the full software stack for its Zhenwu AI chips, at the World AI Conference in Shanghai on Saturday, aiming to lower migration barriers for develop…
PixelUp, a 100% offline AI video upscaler for Windows, uses optimized FSRCNN and ESPCN deep learning models with universal hardware acceleration via CUDA, Vulkan, or CPU to upscale low-resolution vide…
Alibaba Group Holding's chip design unit T-Head announced at the World AI Conference in Shanghai on Saturday that it will open-source its SAIL software stack, the foundational architecture for its Zhe…
AMD reported Q1 2026 revenues of $10.25 billion, a 38% year-over-year increase, with data center revenue reaching approximately $5.775 billion, up 57%. The chipmaker is expanding AI infrastructure thr…
Chapel 2.9 shipped on June 18 with LLVM 22 and CUDA 13 support, a prototypical dynamic library loading feature for distributed parallel code, a browsable Mason package registry, and union types that h…
A developer shared a configuration example for running Qwen 3.6 with llama-turboquant, a Docker-based CUDA inference container. The setup includes GPU offloading, flash attention, speculative decoding…
A hardware-software co-design prototype called PIM-HBM explores zero-copy memory view fusion and fault telemetry for LLM training, using JAX/XLA integration and branchless CUDA kernels to mitigate com…
A FreeBSD Foundation intern, Sourojeet Adhikari, is porting AMD's ROCm compute stack to FreeBSD, aiming to get at least simple vector addition working on the BSD platform. The project was highlighted …
A developer demonstrated how to separate only CUDA kernels into .cu files while keeping the rest of the code in .cpp files, using a minimal noopKernel example. The approach confines CUDA-specific code…
Trellis2.c, a native executable for local 3D generation with Vulkan and CUDA backends, has been released on GitHub. The project aims to provide a lightweight alternative to Python/PyTorch runtimes, si…
Researchers present Fleet, a hierarchical task-based abstraction for multi-die GPUs that introduces Chiplet-tasks to exploit chiplet-level locality and synchronization. On AMD Instinct MI350 with Qwen…
Nvidia released an open-source toolkit of over 80 AI agent skills under its Metropolis framework on June 1, aimed at developers building vision AI for manufacturing, infrastructure, robotics, and indu…
More than 192,000 organizations downloaded hundreds of millions of CUDA packages from Anaconda over the past year, with annual CUDA download volume growing 75% year over year, according to Anaconda an…
A cubin (CUDA binary) is a standard ELF64 file with NVIDIA-specific sections that encode everything needed to load and launch a kernel, including machine code, parameter layout, and register allocatio…
PrismML released Bonsai 27B, claiming it is the first major AI model of its size that can run on an iPhone, iPad, or Mac, with a 1-bit version fitting in about 4 GB of memory. PrismML CEO Babak Hassib…