GPU架构建模专家(北京)30K-50K

学历:本科 | 工作年限:五年以上 | 年龄:不限
最后刷新:2024/03/20 08:09:42

微信扫一扫:分享

微信里点“发现”,扫一下

立即沟通 申请职位
职位描述
职位介绍
Job Responsibilities:
1. Build the cutting-edge simulator for GPU development and verification. The simulator will be used throughout the entire product lifecycle of our product design;
2. Work with design engineers, micro-architects, and other team members to understand specification and implement system C/C++ models from RTL modules. Within the simulator, tackle interesting technical challenges to meet the aggressive quality, functional and performance targets set for this project;
3. Developing tools to automate the building and running of simulation models;
4. Work with the Performance Verification team to validate the Model, and to root cause disconnects between Model projections, design (RTL), and ultimately, Silicon. • Passionate about delivering bug-free first silicon;
5. Discover ways to make our simulation algorithms more robust and high performance with both arch and engineer.

Job Requirements:
1. Bachelor degree or above;
2. GPU/Graphics concepts: 3D, Media, Imaging & Display etc.;
3. Strong C/C++ knowledge with expertise in at least one modern scripting language (i.e. Python, Perl, Json);
4. Deep familiarity with GPU or CPU simulators, simulation strategies, and identification of architectural performance bottlenecks using visualization;
5. Ability to understand and implement GPU specs and features;
6. Familiarity with either Linux or Windows build/debug toolchains, and software performance optimization;
7. Knowledge of OS internals, compiler technology, and ARM/x86 architecture a plus;
8. Experience with functional and/or performance modeling is an advantage;
9. Knowledge of GPU shader compiler/optimization is an advantage.
公司介绍

沐曦致力于为异构计算提供全栈GPU芯片及解决方案,可广泛应用于人工智能、智慧城市、数据中心、云计算、自动驾驶、数字孪生、元宇宙等前沿领域,为数字经济发展提供强大的算力支撑。公司于2020年9月成立于上海,并在北京、南京、成都、杭州、深圳和武汉等地建立了全资子公司暨研发中心。沐曦拥有技术完备、设计和产业化经验丰富的团队,核心成员平均拥有近20年高性能GPU产品端到端研发经验,曾主导过十多款世界主流高性能GPU产品研发,包括GPU架构定义、GPU IP设计、GPU SoC设计及GPU系统解决方案的量产交付全流程。沐曦打造全栈GPU芯片产品,推出MXN系列GPU(曦思)用于AI推理,MXC系列GPU(曦云)用于AI训练及通用计算,以及MXG系列GPU(曦彩)用于图形渲染,满足数据中心对“高能效”和“高通用性”的算力需求。沐曦产品均采用完全自主研发的GPU IP,拥有完全自主知识产权的指令集和架构,配以兼容主流GPU生态的完整软件栈(MACAMACA),具备高能效和高通用性的天然优势,能够为客户构建软硬件一体的全面生态解决方案,是“双碳”背景下推动数据中心建设和产业数字化、智能化转型升级的算力基石。沐曦自成立之日起即保持着高速的成长性,展示出强大的人才号召力、产业资源聚集力、产品研发执行力和产业生态上下游布局能力。目前,沐曦已与服务器OEM、大数据中心、互联网运营商等行业客户建立了紧密的合作关系,并与众多知名高校和研究机构密切开展产学研合作,快速推进产业上下游生态的建设。

工作地点
北京-海淀区IC-PARK
职位申请 快速申请 登录申请
确定 取消
芯片半导体职位来 高芯圈
登录 / 注册