Taiwan's Zhen Ding Technology has joined Nvidia's MGX ecosystem to supply cable-free PCB-based interconnect technology that replaces hundreds of individual copper cables in AI server racks. The approa…
Researchers from UC Berkeley's UCCL project released mKernel, a library of persistent CUDA kernels that fuse intra-node NVLink communication, inter-node RDMA, and compute into a single kernel to addre…
NVIDIA released Nemotron-Labs Diffusion on May 23, 2026, a family of diffusion language models (DLMs) that generate entire blocks of tokens in parallel and iteratively refine them, rather than produci…
NVIDIA and SchedMD have released a new topology/block plugin for Slurm 23.11 that enables topology-aware job scheduling on NVIDIA GB200 NVL72 systems, allowing workloads to be aligned with NVLink doma…
Intel's data center CPU sales surged as AI inference demand shifted system designs from 8-to-1 GPU-to-CPU ratios to 4-to-1 or even 2-to-1 configurations, increasing CPU content per AI system. Intel CF…
Nvidia's Blackwell B200 GPU is the company's first chiplet design, using two reticle-sized dies that appear as a single GPU to software, moving away from the monolithic approach used in prior generati…