Java VM Software Performance Analysis Engineer

Job Description:

The Role:

We are seeking a JVM Performance Specialist to serve as the technical authority for managed runtimes on our RISC-V IP architecture. You will deep dive into JVM internals (JIT compilation, runtime execution, and garbage collection) to identify and resolve critical performance bottlenecks across pre-silicon, FPGA, and production environments.

By collaborating with our Compiler Team to optimize JIT code generation and fine-tuning GC algorithms, you will directly impact system scalability, throughput, and latency. This role goes beyond high-level tuning to bridge the gap between Java software execution and cutting-edge RISC-V hardware, ensuring our IP delivers peak performance for enterprise-critical workloads.

Responsibilities:

• Profile and analyze core datacenter workloads to pinpoint inefficiencies in CPU utilization, memory hierarchy, I/O throughput, and network interactions within distributed and containerized environments.
• Establish benchmarking methodologies and performance regression testing frameworks to ensure consistent execution quality of Java-based services on RISC-V.
• Act as the primary expert in diagnosing and resolving performance friction within the Just-In-Time (JIT) compiler and the HotSpot runtime environment.
• Collaborate closely with the Compiler Team to drive improvements in JIT code generation, ensuring the JVM effectively leverages the unique features of the RISC-V ISA.
• Optimize JVM parameters to achieve industry-leading latency and throughput.
• Translate performance data into actionable insights for hardware architects, helping to influence future RISC-V IP design to better support managed runtime environments.

Requirements:

• Multiple years of experience with JIT compiler, runtime execution engine, and memory management subsystems.
• Experienced with profiling tools such as perf, ftrace, eBPF, and async-profiler.
• Proficiency in C/C++ and a strong understanding of modern CPU microarchitecture (pipeline execution, branch prediction, and memory consistency models).
• Mastery of assembly (RISC-V, x86, or ARM) and the ability to analyze JIT-generated machine code to identify sub-optimal execution patterns.
• (Nice-to-Have) Experience optimizing large-scale Java frameworks (e.g., Cassandra, Hadoop) within containerized or cloud-native environments.
• (Nice-to-Have) Proven track record of tuning and optimizing GC algorithms (e.g., G1GC, ZGC, Shenandoah) to solve complex latency and throughput issues.
• (Nice-to-Have) Familiarity with industry-standard benchmarks such as SPECjbb, or DaCapo, and the ability to derive meaningful insights from them.

If you want to do incredible work and be challenged by exciting, truly innovative projects that will test the boundaries of your skills and creativity, then SiFive is the place for you!  Be proud of your work.  Do things better.  Join SiFive.