double-precision floating-point format
PulseAugur coverage of double-precision floating-point format — every cluster mentioning double-precision floating-point format across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
Morloc compiler generates CLIs, MCPs, and APIs from single source
The Morloc compiler is a tool designed to generate multiple interfaces, such as command-line interfaces (CLIs), MCPs for AI agents, and application programming interfaces (APIs), from a single source of functions. It le…
-
Sparse MoE models show behavioral divergence due to numerical state tracing differences
A new arXiv paper details how different numerical state tracing methods can lead to behavioral divergence in sparse Mixture-of-Experts (MoE) models, even when mathematically equivalent. Researchers found that varying ag…
-
China launches first 100,000-card domestic AI supercomputing cluster
China has launched its first ten-thousand-card AI supercomputing cluster, the Sugon 8000 (Dengfeng), in Zhengzhou. This cluster is built entirely with domestic components and is designed for both large model training an…
-
New FFT method leverages FP8 tensor cores for high-precision GPU computation
A new research paper proposes an efficient method for calculating Fast Fourier Transforms (FFTs) using NVIDIA's Blackwell Ultra (B300) GPUs. The Ozaki-Bailey FFT technique leverages FP8 tensor cores for dense matrix mul…
-
Supermicro Unveils NVIDIA Vera Rubin NVL4 Blueprint for Converged HPC and AI
Supermicro has introduced a blueprint for NVIDIA's Vera Rubin NVL4 data center building block systems. This new offering provides native double-precision floating-point performance, catering to converged high-performanc…
-
FP8 with reconstruction schemes matches FP64 accuracy in HPC
A new research paper challenges the long-held belief that double-precision (FP64) hardware is essential for high-performance computing (HPC). The authors propose that using FP8 tensor cores, combined with specific recon…