日前，曾在ATI、AMD、苹果和英特尔任职的的传奇GPU架构师Raja Koduri对外公布了其 成立的GPU公司Oxmiq Labs。 据Raja Koduri在其社交媒体介绍，Oxmiq Labs Inc. 是一家全新的 GPU 软件和 IP 初创公司，从 隐形中脱颖而出。"我们组建了一支由 GPU 和 AI 架构师组成的世界级团队，他们拥有 500 多年 的综合经验。OXMIQ 的可许可 IP 从头开始重新构建 GPU。"Raja Koduri说。 "我们希望成为下一代的 Arm"，Raja Koduri在接受路透社采访时表示。 Oxmiq Labs指出，公司正在采用软件优先的方法来构建其芯片设计，并开发了一种工具，允许为 Nvidia 的 CUDA 编写的软件程序在非 Nvidia 硬件上运行，"而无需修改代码或重新编译"。 "我们可能是硅谷 25+ 年来第一家新的 GPU 初创公司，GPU 并不容易"，Raja Koduri说。 一款基于RISC-V的GPU 据Oxmiq Labs所说，现代计算已从根本上转向多模态体验，文本、音频、视频、图像和 3D 环境 无缝交互，GPU 架构成为这一转变的基石。 ...

华为版本CUDA，全面开源开放！ 最新消息，华为宣布为其昇腾AI GPU开源其CANN软件工具包。 红交 发自 凹非寺 量子位 | 公众号 QbitAI 华为轮值董事长徐直军在主题演讲中强调：华为AI战略的核心是算力，并坚持昇腾硬件变现。 基于这一背景下，徐直军在会上宣布华为昇腾硬件使能CANN全面开源开放，Mind系列应用使能套件及工具链全面开源，支持用户自主的深 度挖潜和自定义开发，加速广大开发者的创新步伐， 让昇腾更好用、更易用 。 CANN，神经网络计算架构，提供多层编程接口，帮助用户构建针对华为昇腾的AI应用程序。 它是一个由多种技能栈和算子加速库等组成的软件生态系统，换句话说，它就像是华为版本的CUDA，为GPU提供相同的接口。 巧合的是，同一天，一传奇GPU大佬创业公司终于浮出水面，他们不搞消费级GPU，而是做CUDA类似软件生态系统。 因为一旦开发者想要迁移到其他产品。那么就需要重写代码，使用生态相对不成熟的替代库，同时也会失去英伟达围绕CUDA建立起来的庞大 技术社区的支持。 之前有项目尝试将CUDA功能（通过转换层）引入其他 GPU供应商，但由于英伟达阻拦，这些项目大多未能成功。2024年 ...

Core Insights - Huawei has announced the open-sourcing of its CANN software toolkit for Ascend AI GPUs, emphasizing its AI strategy centered around computing power and monetization of Ascend hardware [3][6][11] - The CANN architecture, akin to NVIDIA's CUDA, provides a multi-layer programming interface to help users build AI applications specifically for Huawei's Ascend GPUs [4][6] Group 1: Huawei's CANN Open-Sourcing - The CANN toolkit has been upgraded to version 8.0, offering both a community version for early feature access and a commercial version tailored for enterprise users [6] - CANN now supports various deep learning frameworks, including PyTorch, MindSpore, TensorFlow, and others, enhancing its compatibility and usability for developers [9] - Huawei has initiated the "CANN Open Source Ecosystem Co-Building Initiative," indicating a strong commitment to developing an open ecosystem around Ascend technology [11] Group 2: Competitive Landscape - A new startup, Oxmiq Labs, founded by a legendary GPU architect, aims to create a software ecosystem similar to CUDA, focusing on GPU hardware and software IP [12][14] - Oxmiq's software stack includes OXCapsule for workload management and OXPython, which allows CUDA-based applications to run on non-NVIDIA hardware without modification [21][23] - The competitive environment is intensifying, with multiple players challenging NVIDIA's dominance in the GPU market, ultimately benefiting developers through increased options and innovation [7][23]

Core Viewpoint - Huawei has announced the open-source release of its CANN software toolkit for Ascend AI GPUs, aiming to enhance its competitiveness against NVIDIA's CUDA ecosystem [1][3][7]. Group 1: Huawei's AI Strategy - Huawei's rotating chairman Xu Zhijun emphasized that the core of Huawei's AI strategy is computing power, with a commitment to monetizing Ascend hardware [3]. - The open-source CANN architecture and Mind series application enablement toolkit will allow developers to explore the potential of Ascend GPUs independently [7][12]. Group 2: CANN Software Toolkit - CANN, a neural network computing architecture, provides multiple programming interfaces to help users build AI applications for Huawei's Ascend [4]. - The CANN toolkit is likened to a Huawei version of CUDA, offering similar interfaces for GPU support [5]. - CANN has been upgraded to version 8.0, with both community and commercial versions available, supporting 12 operating systems [7]. Group 3: Competitive Landscape - On the same day, a new startup led by legendary GPU architect Raja Koduri emerged, focusing on developing a software ecosystem similar to CUDA, but not targeting consumer-grade GPUs [6][13]. - Oxmiq Labs, founded by Koduri, aims to provide a vertically integrated platform combining GPU hardware IP with a comprehensive software stack for AI and graphics workloads [20][22]. Group 4: Oxmiq's Software Strategy - Oxmiq's software stack includes OXCapsule, a unified runtime for workload management, and OXPython, which allows CUDA-based applications to run on non-NVIDIA hardware without modification [29][31]. - The software is designed to be independent of Oxmiq's hardware, which is a core part of its strategy [31]. Group 5: Industry Implications - The competition against NVIDIA's CUDA ecosystem is intensifying, with multiple players entering the field, ultimately benefiting developers [32].

公众号记得加星标⭐️，第一时间看推送不会错过。 日前，曾在ATI、AMD、苹果和英特尔任职的的传奇GPU架构师Raja Koduri对外公布了其成立的GPU公司Oxmiq Labs。 据Raja Koduri在其社交媒体介绍，Oxmiq Labs Inc. 是一家全新的 GPU 软件和 IP 初创公司，从隐形中脱颖而出。"我们 组建了一支由 GPU 和 AI 架构师组成的世界级团队，他们拥有 500 多年的综合经验。OXMIQ 的可许可 IP 从头开始重新 构建 GPU。"Raja Koduri说。 "我们希望成为下一代的 Arm"，Raja Koduri在接受路透社采访时表示。 据该公司在官网介绍，OXMIQ 还提供完整的GPU 硬件 IP堆栈，为从边缘设备和自主机器人中的物理 AI 到企业边缘基础 设施和泽级数据中心的各种硅片解决方案提供支持。其可扩展的 GPU 核心OXCORE将标量、矢量和张量计算引擎集成在 一个模块化架构中，可根据特定工作负载进行定制，从而支持nano-agents、原生 Python 加速以及与 SIMT/CUDA 范式 的兼容性。 OXCORE可通过OXQUILT chiplet ...