NVIDIA 推出的 cuTile (CUDA Tile) 编程模型代表了一次重要的范式转变。 本文将以矩阵乘法（Matrix Multiplication, GEMM）为例，基于 samples/MatMul.py 源码，并参考 NVIDIA 技术博客《Simplify GPU Programming with NVIDIA CUDA Tile in Python》，深入剖析从 SIMT 到 ...

Qualcomm is finally getting serious about AI infrastructure, but its push into the datacenter hinges on the success of an ...

The latest ASIC that grabbed a lot of attention is OpenAI’s Jalapeño chip, its first in-house inference chip, built with ...

Founded by the mind behind the Swift programming language, Modular’s 'write once, run anywhere' stack looks to accelerate ...

Qualcomm is in advanced talks to acquire AI software startup Modular for about $4 billion, Bloomberg reported, in a bet to ...

Nvidia and Microsoft launched the RTX Spark, a new chip designed to bring the full Nvidia AI software stack to slim Windows ...

Qualcomm confirmed a $3.92 billion all-stock deal to buy AI software startup Modular, paired with a Meta Platforms CPU ...

Daisy-chaining two of Dell's Nvidia GB10 DGX Spark systems didn't just pump up my home AI lab—it fundamentally changed how I ...

This book teaches modern GPU kernel programming as a progression: understand the GPU hardware → learn to program it → write state-of-the-art kernels. It treats the Blackwell-class GPU — its memory ...

Abstract: This paper presents a performance modeling and optimization analysis tool to predict and optimize the performance of sparse matrix-vector multiplication (SpMV) on GPUs. We make the following ...