Hello folks，我是 Luga，今天我们来聊一下人工智能应用场景中大语言模型（LLM）底层算力资源支撑设施 - AMD ROCm。 在过去十多年里，GPU 的竞争往往被简化为制程、算力峰值和显存带宽的对比。但随着 AI、HPC ...

专注AIGC领域的专业社区，关注微软&OpenAI、百度文心一言、讯飞星火等大语言模型（LLM）的发展和应用落地，聚焦LLM的市场研究和AIGC开发者生态，欢迎关注！SGLang发布了Mini-SGLang。将30万行代码的庞然大物浓缩为5000行， ...

在人工智能逐步成为国家竞争核心变量的当下，算力正以前所未有的速度重塑技术路径与产业结构。13日举办的「AI ...

Abstract: Heterogeneous CPU-GPU systems are extensively utilized in high-performance computing. Compute Unified Device Architecture (CUDA) [1] is a model for programming the GPUs. A CUDA program ...

英伟达发布最新版CUDA 13.1，官方直接定性：这是自2006年诞生以来最大的进步。 核心变化是推出全新的CUDA Tile编程模型，让开发者可以用Python写GPU内核，15行代码就能达到200行CUDA C++代码的性能。 英伟达是不是亲手终结了CUDA的“护城河”？如果英伟达也转向Tile ...

Calling it the largest advancement since the NVIDIA CUDA platform was inroduced in 2006, NVIDIA has launched CUDA 13.1 with CUDA Tile, which the company said introduces a virtual instruction set for ...

Abstract: In a CPU-GPU based heterogeneous computing system, the input data to be processed by the kernel resides in the host memory. The host and the device memory address spaces are different.