简历内推可发送至: [email protected]
年假多 WLB Linux 性能优化方向，上海、北京、成都均有 office
Job Area:
Engineering Group, Engineering Group > Software Engineering
General Summary:
Responsible for Linux system performance analysis and optimization on upstream kernel based products on Qualcomm platforms. You will identify performance bottlenecks, drive root-cause analysis, develop fixes/improvements, and prevent regressions through benchmarking and automation. Focus areas include CPU scheduling, memory management, I/O subsystem, boot performance, and low-level HW performance characteristics (PMU, cache, DDR).
-
Analyze and optimize Linux upstream kernel performance across scheduler, memory, I/O, and boot-up sequence. -
Identify system-level performance bottlenecks using profiling/tracing tools/bootchart; propose and implement improvements in kernel and/or userspace. -
Conduct root-cause analysis for performance issues and regressions; provide clear analysis reports and corrective actions. -
Build and maintain performance benchmarking methodology: KPI definition, test plans, automation, baseline tracking, and regression gating. -
Optimize boot time: analyze boot critical path (kernel init, userspace init, systemd, services), reduce latency, and validate improvements. -
Drive performance-related patch development and upstream-quality engineering (review, test, documentation). -
Collaborate with cross-functional teams (platform, camera, graphics, multimedia, power, QA) to diagnose and resolve performance issues.
Skills /Experience Required:
-
Strong experience in Linux system performance analysis and optimization (kernel and/or system level). -
Solid understanding of Linux upstream kernel internals, especially: -
Scheduler (CFS/RT, task wakeup, CPU affinity, load balancing, latency) -
Memory management (page cache, reclaim, compaction, CMA/contiguous memory concepts, memory pressure) -
I/O subsystem (block layer, filesystems basics, I/O scheduling, latency/throughput tradeoffs) -
Boot performance fundamentals (init sequence, kernel/userspace interactions) -
Good understanding of low-level performance topics: PMU/perf events, CPU cache behavior, memory bandwidth/DDR. -
Proficient programming skills in C (kernel/userspace) and scripting in Python/Shell for automation. -
Hands-on experience with performance tools such as perf, ftrace, trace-cmd, eBPF/bpftrace, flame graphs, and standard benchmarking tools. -
Strong problem-solving skills, ability to communicate findings clearly, and ability to work effectively across teams.
Select Language