hip

colibrì fork with AMD Radeon GPU (HIP) support for the DeepSeek V4 engine on Windows — prebuilt for RX 6000/7000/9000

Read-only mirror of the official repository: https://gitlab.com/trasgo-group-valladolid/controllers. Controllers is a library written in C11 that provides a simplified way to program applications that can exploit heterogeneous computational platforms including accelerators and/or multi-core CPUs.

Distributed GPU-accelerated tree search: Exploring parallel Branch-and-Bound based on the MPI+X standard (X=OpenMP, MPI, CUDA, HIP, etc), written in C language.

Matrix Exponential Approximation in HIP

Tiny autograd engine in pure C11.

神经音频编解码器 — Multi-backend neural audio codec. CPU (AVX/AVX2/AVX-512, NEON/SVE, RVV), GPU (CUDA, HIP/ROCm, Vulkan), LLVM JIT. Clean-room implementation.

AMD GPU Profiler for ROCm compatible GPU.

Gen 1 C host substrate for openOODA. 26 capability tokens, Landlock sandbox, PQC (ML-KEM-768 + ML-DSA-65), GPU HIP/ROCm.

A CUDA/HIP-accelerated video photo-mosaic generator achieving up to 184× speedup, enabling per-frame mosaic rendering for full-length videos.

libnomp is a loopy based runtime for C programming language to create domain specific compilers.

GPU-accelerated linear solvers based on the conjugate gradient (CG) method, supporting NVIDIA and AMD GPUs with GPU-aware MPI, NCCL, RCCL or NVSHMEM

A high-performance RCCL / NCCL (ROCm Communication Collectives Library) plugin for Thunderbolt 5 that enables GPU-to-GPU communication across Thunderbolt connections with RDMA support.