package tezos-benchmark

You can search for identifiers within the package.

in-package search v0.2.0

tezos-benchmark
- Tezos_benchmark
  - Ast
    
    Ast
    
    At_least_10
    
    X
    
    Make
    
    Size
    
    Optimize
    
    X
    
    S
    
    To_ast
    
    Ast
    
    Transform
    
    F
    
    Ast
    
    X
  - Base_samplers
    
    Adversarial
  - Benchmark
    
    Benchmark_base
    
    S
    
    Simple
    
    Simple_with_num
  - Benchmark_helpers
  - Builtin_benchmarks
    
    Timer_latency_bench
  - Builtin_models
  - Codegen
    
    Parser
  - Config
  - Costlang
    
    Arg_names
    
    Beta_normalize
    
    X
    
    Eval
    
    Eval_linear_combination_impl
    
    Eval_to_vector
    
    Fold_constants
    
    X
    
    Free_variables
    
    Hash_cons
    
    X
    
    Hash_cons_vector
    
    Identity
    
    X
    
    Let_lift
    
    X
    
    Num
    
    Pp
    
    S
    
    Subst
    
    P
    
    X
    
    Transform
    
    X
    
    Ty
    
    Type
    
    Void
  - Crypto_samplers
    
    Finite_key_pool_S
    
    Make_finite_key_pool
    
    Arg
    
    P_Finite_key_pool_S
    
    Param_S
    
    V0
    
    Finite_key_pool_S
    
    Make_finite_key_pool
    
    Arg
    
    V1
    
    Finite_key_pool_S
    
    Make_finite_key_pool
    
    Arg
    
    V_latest
    
    Finite_key_pool_S
    
    Make_finite_key_pool
    
    Arg
  - Csv
  - Dep_graph
    
    Graph
    
    Graphviz
    
    Solver
    
    Solved
    
    Unsolved
  - Fixed_point_transform
    
    Apply
    
    P
    
    X
  - Free_variable
    
    Map
    
    Set
    
    Sparse_vec
    
    Op
    
    Table
  - Generator
    
    S
  - Inference
  - Maths
  - Measure
    
    Time
  - Model
    
    Applied
    
    X
    
    Binary_operation
    
    Def
    
    X
    
    Instantiate
    
    M
    
    Def
    
    X
    
    X
    
    Instantiated
    
    Model_impl
    
    Def
    
    X
  - Namespace
    
    Hashtbl
    
    Map
    
    Set
  - Ocamlformat
  - Override
  - Plot
    
    Bar
    
    Color
    
    Data
    
    Fill
    
    Histogram
    
    Line
    
    Pointtype
    
    Scatter
    
    Style
    
    Tics
  - Pyinference
    
    LinearModel
  - Pyinit
  - Pytools
  - Registration
  - Scikit_matrix
  - Sparse_vec
    
    Make
    
    M
    
    Op
    
    S
    
    Op
    
    String
    
    Op
  - Structure_samplers

Legend:
Library
Module
Module type
Parameter
Class
Class type

This module defines several signatures to create a benchmark. * The most generic is S; it provides a rather complete control on benchmark creation. * Simple and Simple_with_num can be used to ease benchmark creation, with simpler parameters. Registration functions in Registration will convert these simple interfaces back to S. * All the common parameters for these signatures are declared in Benchmark_base.

type group =

| Standalone
| Group of string
| Generic

Some benchmarks depend on others, and some are for generic parameters that most benchmarks depend on. We need this information in order to correctly infer the values of parameters after a benchmark run; the user provides it with a group.

* Standalone: benchmarks that don't depend on others (except generic ones). This is the value to use if you're not sure whether you should group your benchmark. * Group: benchmarks that belong to the given inference group. Note that setting a benchmark with a group that is referenced only in this benchmark will produce the same inference results as with Standalone. * Generic: for generic parameters only.

type purpose =

| Other_purpose of string
| Generate_code of string

Described the purpose of the benchmark. * Generate_code of destination: generates code at the given destination file, prefixed by "src/proto_alpha/lib_protocol/" and suffixed by "_costs_generated.ml". * Other_purpose of purpose: any other purpose. The goal is to explain why the function is benchmarked since it does not produce a cost function.

type 'config parameters = {

bench_number : int;
config : 'config;

}

Benchmark parameters.

module type Benchmark_base = sig ... end

Common declarations used by the different module types of benchmarks

module type Simple = sig ... end

A simplification of S below, when only one model is defined and no particular process is needed to generate benchmarks.

module type Simple_with_num = sig ... end

A simplification of S below, when only one model is defined.

module type S = sig ... end

The module type of benchmarks

type t = (module S)

type simple = (module Simple)

type simple_with_num = (module Simple_with_num)

val pp : Stdlib.Format.formatter -> t -> unit

type ('cfg, 'workload) poly =
  (module S
  with type config = 'cfg
   and type workload = 'workload)

type packed =

| Ex : ('cfg, 'workload) poly -> packed

val ex_unpack : t -> packed

val name : t -> Namespace.t

Get the name of a benchmark

val info : t -> string

Get the description of a benchmark

val tags : t -> string list

Get the tags of a benchmark

val get_free_variable_set : (module S) -> Free_variable.Set.t

Returns the free variables occur in the models of the benchmark.