Mavryk_benchmarks_proto_alpha.Michelson_generation
type generator_config = {
target_size : Mavryk_benchmark.Base_samplers.range;
The target size of the terms, in number of nodes, is sampled uniformly in target_size
.
burn_in_multiplier : int;
The generators are based on a Markov chain, which must be "heated-up" until it reaches its stationary state. A prefix of samples are therefore thrown away: this is called the burn-in phase. The number of thrown away terms is proportional to burn_in_multiplier
and target_size
.
}
generator_config
specifies some parameters to the Mavryk_benchmark_alpha.Michelson_mcmc_samplers
Michelson code and data generators.
val default_generator_config : generator_config
Default configuration for the generators.
val generator_config_encoding :
generator_config Mavryk_base.TzPervasives.Data_encoding.t
Samplers
val make_data_sampler :
Stdlib.Random.State.t ->
generator_config ->
Mavryk_benchmark_alpha.Michelson_mcmc_samplers.michelson_data
make_data_sampler
constructs a Michelson data sampler based on the infrastructure available in Mavryk_benchmark_alpha.Michelson_mcmc_samplers
.
val make_code_sampler :
Stdlib.Random.State.t ->
generator_config ->
Mavryk_benchmark_alpha.Michelson_mcmc_samplers.michelson_code
make_code_sampler
constructs a Michelson code sampler based on the infrastructure available in Mavryk_benchmark_alpha.Michelson_mcmc_samplers
.
Samplers
is an instance of the direct-style (non-MCMC based) samplers implemented in Mavryk_benchmark_alpha.Michelson_samplers
.