Agentic AI Reference Architecture

Observability & Evaluation · Software component

Kernel Profiler

Software componentObservability & EvaluationObservability & Evaluationarc:KernelProfiler

An exhaustive profiler that instruments individual accelerator kernels at instruction granularity to expose intra-kernel inefficiencies.

Responsibility. Analyzes individual kernel execution at instruction granularity.

Also known as: Kernel-level profiler, Exhaustive kernel instrumentation

Variant of Performance Profiler abstract

When to choose. Choose only for controlled deep-dive optimization of individual kernels (warp divergence, memory access, register use); its 10-100x slowdown precludes continuous production use.

is target of alternativeToalternative tomonitorsis target of alternativeTospecializesis constrained byExecution Profiler: is target of alternativeToExecution ProfilerGPU System Profiler: alternative toGPU System ProfilerAttention Kernel: monitorsAttention KernelInference Engine Profiler: is target of alternativeToInference Engine ProfilerPerformance Profiler: specializesPerformance ProfilerProfiling Tier Policy: is constrained byProfiling Tier Policy
Direct neighbourhood (hover for relationship types)

Relationships

is constrained by control

monitors assurance

alternative to variability

Design guidance

Quantitative guidance

As stated by the sources; verify before use.

Classification

Technologies
NVIDIA Nsight Compute
Quality attributes
Maintainability (ISO/IEC 25010)

Sources

  1. Ch4.4: T. Nguyen, "Performance Profiling and Optimization," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 4.4. ISBN: 9798244538229.