平头哥平头哥-CPU物理设计专家-北京
任职要求
电子工程/微电子/VLSI等相关专业学士或硕士,3年以上RTL2GDS项目经验; *熟悉先进工艺节点、半导体器件、超大规模集成电路、定制化库及优化工作; *具有高性能CPU物理设计经验,熟悉处理器微架构、逻辑设计和计算单元(加法器/乘法器等)等方面经验者优先; *熟悉先进的时钟树、电源规划及电源完整性方面的开发技能; *在先进工艺节点上具有相对大型设计(>10m Flops)的项目经验,具备高性能低功耗方面的优化技能经验; *熟悉DVFS、DFT、DFY、DFM者优先; *熟悉RTL级和门级网表功耗分析、EMIR分析及修复者优先; 有使用以下工具的量产投片经验者优先: *Cadence Innovus and/or Synopsys ICC2 -> Cadence Innovus or Sy…
工作职责
作为物理设计团队的一员,你将参与打造基于先进工艺的下一代服务器SoC芯片,一方面将驱动整个RTL2GDS的物理设计交付流程,包括 Floorplan, Synthesis, P&R, Timing, PI, Power 以及Sign-offs;另一方面也将专注于芯片PPA(性能、功耗、面积)的优化工作。 工作职责包括但不限于以下几个方面: * 在先进工艺节点上实现复杂的CPU模块; * 与架构师及设计人员密切合作进行PPA的优化; * 交付流程涵盖:Floor planning,physical-aware synthesis, equivalence checks, partitioning, IO assignment and IP integration, power grid, P&R, CTS, timing closure, power analysis等 * 设计和实现Timing ECO并完成tapeout 签核 (sign-offs) * 优化改进芯片物理设计流程和方法学;
1.与CPU设计者密切合作,进行关键时序路径的分析与逻辑电路优化; 2.前后端协同设计,持续提升CPU PPA性能指标; 3.CPU Core的物理设计与交付,达到产品指标要求; 4.CPU 集成实现工程师(中端)综合实现

1.参与自动驾驶CPU(高性能AP核)的需求和规格的定义与分析; 2.完成CPU子系统的交付,包括RTL集成、时钟/复位设计、电源域划分、低功耗流程、静态时序分析与物理协同,支持验证团队测试并协助后端完成物理实现; 3.为CPU计算子系统打造产品竞争力,包括SOC场景需求分析、微架构及方案制定、性能和成本分析、时序面积功耗优化等工作; 4.支持系统级验证与硅后调试,完成量产问题跟踪、良率提升等相关工作;
• Collaborate with developers, researchers, and framework maintainers across industries to identify and resolve performance challenges in diverse workloads such as AI, data analytics, simulation, and numerical computing. • Profile, analyze, and optimize CPU performance from application-level algorithms down to low-level microarchitecture. • Contribute to open-source frameworks, key software stacks, reference implementations, and performance libraries to unlock full CPU potential. • Work closely with NVIDIA’s architecture, research, libraries, tools, and system software teams to improve our overall platform performance. • Provide insights that shape next-generation CPU designs, compiler toolchains, and development workflows for better developer productivity and throughput.
• Collaborate with developers, researchers, and framework maintainers across industries to identify and resolve performance challenges in diverse workloads such as AI, data analytics, simulation, and numerical computing. • Profile, analyze, and optimize CPU performance from application-level algorithms down to low-level microarchitecture. • Contribute to open-source frameworks, key software stacks, reference implementations, and performance libraries to unlock full CPU potential. • Work closely with NVIDIA’s architecture, research, libraries, tools, and system software teams to improve our overall platform performance. • Provide insights that shape next-generation CPU designs, compiler toolchains, and development workflows for better developer productivity and throughput.