Autotuning
Autotuning automatically evaluates and selects or adjusts algorithm, implementation, or configuration choices to improve a measured objective, such as performance.
Microsoft Azure in Action
Automatic tuning is a database optimization capability that analyzes workload behavior and recommends or applies configuration changes to improve performance.
- How it is used
- Uses workload analysis to guide Azure SQL Database performance decisions, particularly index creation and removal.
- How deep it goes
- Explains three tuning options and where to control them, without detailing the algorithms behind their decisions.
- Examples
- Uses an ordering database to suggest indexes on order-item identifiers and order dates.
- What you can do after
- Operate/configure it through the Azure SQL Database Automatic Tuning menu by choosing settings for Force Plan, Create Index, and Drop Index.
GPU Programming with Triton
Autotuning is the process of evaluating alternative implementation configurations and selecting one based on measured performance.
- How it is used
- Selects a measured launch configuration for GPU kernels when a single hand-chosen setup may not suit every matrix shape.
- How deep it goes
- Measures supplied tile, warp, and pipeline-stage configurations and reuses the winner by key, without detailing the search algorithm.
- Examples
- Compares matrix-product configurations across shapes, including tile choices for wide outputs and small matrices.
- What you can do after
- Apply Triton's autotune decorator to benchmark candidate configurations and reuse a selected setup for matching keys.
Also covered in
Related concepts
- CUTLASS profilerused by · The CUTLASS Profiler is a command-line executable for launching and benchmarking instantiated CUTLASS kernels across problem sizes and configurations.
- kernel launch configurationconfigures · A kernel launch configuration specifies the execution parameters, including grid and thread-block dimensions, for launching a GPU kernel.
- indexinguses · Indexing is the process of assigning or using positions, keys, or lookup structures to identify and access elements or records.
- recommendationproduces · A recommendation is a suggested item, choice, or course of action offered to guide a decision.
Questions about autotuning
What is autotuning?
Autotuning automatically evaluates and selects or adjusts algorithm, implementation, or configuration choices to improve a measured objective, such as performance.
Which Manning books explain autotuning?
Microsoft Azure in Action and GPU Programming with Triton.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.

