sirius

Run TPC-H benchmarks on Super Sirius or DuckDB CPU baseline — generate data, execute queries, validate results, and compare timings. Trigger when the user mentions benchmarking, TPC-H, performance testing, query runtimes, or wants to compare Sirius vs DuckDB speed.

Use this skill to analyze Sirius query execution logs — to find why a query is slow, diagnose validation errors (wrong results), spot excessive downgrading, identify pipeline gaps, or compare two runs of the same query. Trigger whenever the user mentions a Sirius log file, says "look at this log", asks about per-operator time, pipeline gaps, memory pressure, downgrade behavior, validation/correctness mismatches, or wants to compare two query runs. Also use as a hand-off target from race-check (for two-run comparisons), runtime-errors (for "where did the query stall before hang"), and validate (for diffing the bad run against a good run). This is the log-only counterpart to profile-analyzer (nsys) and optimization-advisor (source-level guidance).

Use this skill to generate benchmark datasets (TPC-H, TPC-DS, etc.). Trigger when the user needs test data at a specific scale factor for benchmarking or testing. Supports parquet and duckdb output formats.

Use this skill to find exactly which source code to optimize for better GPU performance. Maps nsys profile hotspots to specific Sirius source files and functions, classifies bottlenecks as GPU-bound, CPU-bound, or sync-bound, and recommends actionable code changes. Trigger when the user wants to know what to optimize, where to focus coding effort, or wants source-level optimization guidance. This skill focuses on actionable source code targets — for generating performance reports and measurements, use profile-analyzer instead.

Use this skill to understand why a Sirius query is slow, identify GPU bottlenecks, or detect performance regressions. Generates reports with kernel occupancy, memory bandwidth, operator attribution, and cross-run comparisons. Trigger when the user mentions profiling, nsys, GPU utilization, kernel analysis, performance reports, or wants to compare query timings across runs. This skill focuses on measurement and reporting — for mapping hotspots to source code fixes, use optimization-advisor instead.

Use this skill when the build fails, compilation errors occur, or you see undefined references, linker errors, CUDA compilation issues, missing headers, or template instantiation failures. Analyzes errors, suggests fixes, and iteratively rebuilds until success.

Discover and document a dependency library or submodule — analyzes all uses within the codebase, divides the library into logical modules, identifies which modules are used, and generates LLM-consumable API documentation for each module. Use when the user wants to understand a library dependency, map its modules, or generate API reference docs for a submodule.

Use this skill to update Super Sirius documentation after code changes. Trigger when the user says "update docs", "refresh documentation", "sync docs with code changes", or after merging PRs that changed the Super Sirius codebase. Inspects merged PRs since the last update and patches affected doc files.

Use this skill when a Sirius query crashes, segfaults, hangs, throws an exception, or unexpectedly falls back to CPU. Also use when you see CUDA errors, std::bad_alloc, or the process gets killed. Diagnoses issues using log analysis, cuda-gdb, AddressSanitizer, and NVIDIA Compute Sanitizer.

Use this skill when a Sirius query returns wrong results, missing rows, extra rows, or incorrect values compared to DuckDB CPU. Pinpoints the faulty operator using per-operator row counts and data checksums. Also detects CUDA stream synchronization issues that cause garbage data.

Use this skill to find which commit introduced a bug or regression. Uses git bisect with automated build and test. Trigger when a bug appeared recently, a query started failing, performance regressed, or the user wants to compare behavior between two commits.

Automatically identify which dependency library modules are relevant to a task and load their API documentation into context. Use PROACTIVELY before implementing features, fixing bugs, or writing new operators — analyzes the task description and loads cudf, rmm, duckdb, cucascade module docs to improve code quality. Trigger when the user asks to implement, add, fix, or modify GPU operators, pipeline components, memory management, joins, aggregations, sorting, expressions, or data I/O.

Use this skill to find the optimal Sirius configuration for TPC-H workloads at any scale factor. Trigger when the user wants to tune performance, optimize config parameters, find the best thread count, batch size, or cache mode, or benchmark different Sirius configurations against each other. Also use when the user mentions "config tuning", "parameter sweep", or "optimal settings".

Use this skill when query results are non-deterministic, differ between runs, or you suspect data races, deadlocks, or thread safety issues. Uses ThreadSanitizer (CPU) and NVIDIA Compute Sanitizer (GPU) to detect and diagnose race conditions.