Merge pull request #8451 from imapp-pl/yul-phaser-program-cache

[yul-phaser] Program cache

test/CMakeLists.txt

@@ -154,6 +154,7 @@ set(yul_phaser_sources
     yulPhaser/Phaser.cpp
     yulPhaser/Population.cpp
     yulPhaser/Program.cpp
+    yulPhaser/ProgramCache.cpp
     yulPhaser/Selections.cpp
     yulPhaser/SimulationRNG.cpp
 
@@ -170,6 +171,7 @@ set(yul_phaser_sources
     ../tools/yulPhaser/Phaser.cpp
     ../tools/yulPhaser/Population.cpp
     ../tools/yulPhaser/Program.cpp
+    ../tools/yulPhaser/ProgramCache.cpp
     ../tools/yulPhaser/Selections.cpp
     ../tools/yulPhaser/SimulationRNG.cpp
 )

test/yulPhaser/AlgorithmRunner.cpp

@@ -20,6 +20,8 @@
 #include <tools/yulPhaser/AlgorithmRunner.h>
 #include <tools/yulPhaser/Common.h>
 
+#include <liblangutil/CharStream.h>
+
 #include <libsolutil/CommonIO.h>
 
 #include <boost/filesystem.hpp>
@@ -29,6 +31,7 @@
 using namespace std;
 using namespace boost::unit_test::framework;
 using namespace boost::test_tools;
+using namespace solidity::langutil;
 using namespace solidity::util;
 
 namespace fs = boost::filesystem;
@@ -92,7 +95,7 @@ BOOST_AUTO_TEST_SUITE(AlgorithmRunnerTest)
 BOOST_FIXTURE_TEST_CASE(run_should_call_runNextRound_once_per_round, AlgorithmRunnerFixture)
 {
 	m_options.maxRounds = 5;
-	AlgorithmRunner runner(Population(m_fitnessMetric), m_options, m_output);
+	AlgorithmRunner runner(Population(m_fitnessMetric), {}, m_options, m_output);
 
 	CountingAlgorithm algorithm;
 
@@ -112,6 +115,7 @@ BOOST_FIXTURE_TEST_CASE(run_should_print_the_top_chromosome, AlgorithmRunnerFixt
 		// NOTE: Chromosomes chosen so that they're not substrings of each other and are not
 		// words likely to appear in the output in normal circumstances.
 		Population(m_fitnessMetric, {Chromosome("fcCUnDve"), Chromosome("jsxIOo"), Chromosome("ighTLM")}),
+		{},
 		m_options,
 		m_output
 	);
@@ -131,7 +135,7 @@ BOOST_FIXTURE_TEST_CASE(run_should_save_initial_population_to_file_if_autosave_f
 {
 	m_options.maxRounds = 0;
 	m_options.populationAutosaveFile = m_autosavePath;
-	AlgorithmRunner runner(m_population, m_options, m_output);
+	AlgorithmRunner runner(m_population, {}, m_options, m_output);
 	assert(!fs::exists(m_autosavePath));
 
 	runner.run(m_algorithm);
@@ -145,7 +149,7 @@ BOOST_FIXTURE_TEST_CASE(run_should_save_population_to_file_if_autosave_file_spec
 {
 	m_options.maxRounds = 1;
 	m_options.populationAutosaveFile = m_autosavePath;
-	AlgorithmRunner runner(m_population, m_options, m_output);
+	AlgorithmRunner runner(m_population, {}, m_options, m_output);
 	assert(!fs::exists(m_autosavePath));
 
 	runner.run(m_algorithm);
@@ -159,7 +163,7 @@ BOOST_FIXTURE_TEST_CASE(run_should_overwrite_existing_file_if_autosave_file_spec
 {
 	m_options.maxRounds = 5;
 	m_options.populationAutosaveFile = m_autosavePath;
-	AlgorithmRunner runner(m_population, m_options, m_output);
+	AlgorithmRunner runner(m_population, {}, m_options, m_output);
 	assert(!fs::exists(m_autosavePath));
 
 	vector<string> originalContent = {"Original content"};
@@ -180,7 +184,7 @@ BOOST_FIXTURE_TEST_CASE(run_should_not_save_population_to_file_if_autosave_file_
 {
 	m_options.maxRounds = 5;
 	m_options.populationAutosaveFile = nullopt;
-	AlgorithmRunner runner(m_population, m_options, m_output);
+	AlgorithmRunner runner(m_population, {}, m_options, m_output);
 	assert(!fs::exists(m_autosavePath));
 
 	runner.run(m_algorithm);
@@ -198,7 +202,7 @@ BOOST_FIXTURE_TEST_CASE(run_should_randomise_duplicate_chromosomes_if_requested,
 	m_options.randomiseDuplicates = true;
 	m_options.minChromosomeLength = 50;
 	m_options.maxChromosomeLength = 50;
-	AlgorithmRunner runner(population, m_options, m_output);
+	AlgorithmRunner runner(population, {}, m_options, m_output);
 
 	runner.run(algorithm);
 
@@ -227,7 +231,7 @@ BOOST_FIXTURE_TEST_CASE(run_should_not_randomise_duplicate_chromosomes_if_not_re
 
 	m_options.maxRounds = 1;
 	m_options.randomiseDuplicates = false;
-	AlgorithmRunner runner(population, m_options, m_output);
+	AlgorithmRunner runner(population, {}, m_options, m_output);
 
 	runner.run(algorithm);
 
@@ -237,6 +241,33 @@ BOOST_FIXTURE_TEST_CASE(run_should_not_randomise_duplicate_chromosomes_if_not_re
 	BOOST_TEST(runner.population().individuals()[2].chromosome == duplicate);
 }
 
+BOOST_FIXTURE_TEST_CASE(run_should_clear_cache_at_the_beginning_and_update_it_before_each_round, AlgorithmRunnerFixture)
+{
+	CharStream sourceStream = CharStream("{}", current_test_case().p_name);
+	vector<shared_ptr<ProgramCache>> caches = {
+		make_shared<ProgramCache>(get<Program>(Program::load(sourceStream))),
+		make_shared<ProgramCache>(get<Program>(Program::load(sourceStream))),
+	};
+
+	m_options.maxRounds = 10;
+	AlgorithmRunner runner(Population(m_fitnessMetric), caches, m_options, m_output);
+	CountingAlgorithm algorithm;
+
+	BOOST_TEST(algorithm.m_currentRound == 0);
+	BOOST_TEST(caches[0]->currentRound() == 0);
+	BOOST_TEST(caches[1]->currentRound() == 0);
+
+	runner.run(algorithm);
+	BOOST_TEST(algorithm.m_currentRound == 10);
+	BOOST_TEST(caches[0]->currentRound() == 10);
+	BOOST_TEST(caches[1]->currentRound() == 10);
+
+	runner.run(algorithm);
+	BOOST_TEST(algorithm.m_currentRound == 20);
+	BOOST_TEST(caches[0]->currentRound() == 10);
+	BOOST_TEST(caches[1]->currentRound() == 10);
+}
+
 BOOST_AUTO_TEST_SUITE_END()
 BOOST_AUTO_TEST_SUITE_END()
 

test/yulPhaser/FitnessMetrics.cpp

@@ -76,15 +76,16 @@ protected:
 	Chromosome m_chromosome{vector<string>{UnusedPruner::name, EquivalentFunctionCombiner::name}};
 	Program m_program = get<Program>(Program::load(m_sourceStream));
 	Program m_optimisedProgram = optimisedProgram(m_program);
+	shared_ptr<ProgramCache> m_programCache = make_shared<ProgramCache>(m_program);
 };
 
 class FitnessMetricCombinationFixture: public ProgramBasedMetricFixture
 {
 protected:
 	vector<shared_ptr<FitnessMetric>> m_simpleMetrics = {
-		make_shared<ProgramSize>(m_program, 1),
+		make_shared<ProgramSize>(m_program, nullptr, 1),
-		make_shared<ProgramSize>(m_program, 2),
+		make_shared<ProgramSize>(m_program, nullptr, 2),
-		make_shared<ProgramSize>(m_program, 3),
+		make_shared<ProgramSize>(m_program, nullptr, 3),
 	};
 	vector<size_t> m_fitness = {
 		m_simpleMetrics[0]->evaluate(m_chromosome),
@@ -97,31 +98,66 @@ BOOST_AUTO_TEST_SUITE(Phaser)
 BOOST_AUTO_TEST_SUITE(FitnessMetricsTest)
 BOOST_AUTO_TEST_SUITE(ProgramBasedMetricTest)
 
-BOOST_FIXTURE_TEST_CASE(optimisedProgram_should_return_optimised_program, ProgramBasedMetricFixture)
+BOOST_FIXTURE_TEST_CASE(optimisedProgram_should_return_optimised_program_even_if_cache_not_available, ProgramBasedMetricFixture)
 {
-	string code = toString(DummyProgramBasedMetric(m_program).optimisedProgram(m_chromosome));
+	string code = toString(DummyProgramBasedMetric(m_program, nullptr).optimisedProgram(m_chromosome));
 
 	BOOST_TEST(code != toString(m_program));
 	BOOST_TEST(code == toString(m_optimisedProgram));
 }
 
+BOOST_FIXTURE_TEST_CASE(optimisedProgram_should_use_cache_if_available, ProgramBasedMetricFixture)
+{
+	string code = toString(DummyProgramBasedMetric(nullopt, m_programCache).optimisedProgram(m_chromosome));
+
+	BOOST_TEST(code != toString(m_program));
+	BOOST_TEST(code == toString(m_optimisedProgram));
+	BOOST_TEST(m_programCache->size() == m_chromosome.length());
+}
+
+BOOST_FIXTURE_TEST_CASE(optimisedProgramNoCache_should_return_optimised_program_even_if_cache_not_available, ProgramBasedMetricFixture)
+{
+	string code = toString(DummyProgramBasedMetric(m_program, nullptr).optimisedProgramNoCache(m_chromosome));
+
+	BOOST_TEST(code != toString(m_program));
+	BOOST_TEST(code == toString(m_optimisedProgram));
+}
+
+BOOST_FIXTURE_TEST_CASE(optimisedProgramNoCache_should_not_use_cache_even_if_available, ProgramBasedMetricFixture)
+{
+	string code = toString(DummyProgramBasedMetric(nullopt, m_programCache).optimisedProgramNoCache(m_chromosome));
+
+	BOOST_TEST(code != toString(m_program));
+	BOOST_TEST(code == toString(m_optimisedProgram));
+	BOOST_TEST(m_programCache->size() == 0);
+}
+
 BOOST_AUTO_TEST_SUITE_END()
 BOOST_AUTO_TEST_SUITE(ProgramSizeTest)
 
 BOOST_FIXTURE_TEST_CASE(evaluate_should_compute_size_of_the_optimised_program, ProgramBasedMetricFixture)
 {
-	size_t fitness = ProgramSize(m_program).evaluate(m_chromosome);
+	size_t fitness = ProgramSize(m_program, nullptr).evaluate(m_chromosome);
 
 	BOOST_TEST(fitness != m_program.codeSize());
 	BOOST_TEST(fitness == m_optimisedProgram.codeSize());
 }
 
+BOOST_FIXTURE_TEST_CASE(evaluate_should_be_able_to_use_program_cache_if_available, ProgramBasedMetricFixture)
+{
+	size_t fitness = ProgramSize(nullopt, m_programCache).evaluate(m_chromosome);
+
+	BOOST_TEST(fitness != m_program.codeSize());
+	BOOST_TEST(fitness == m_optimisedProgram.codeSize());
+	BOOST_TEST(m_programCache->size() == m_chromosome.length());
+}
+
 BOOST_FIXTURE_TEST_CASE(evaluate_should_repeat_the_optimisation_specified_number_of_times, ProgramBasedMetricFixture)
 {
 	Program const& programOptimisedOnce = m_optimisedProgram;
 	Program programOptimisedTwice = optimisedProgram(programOptimisedOnce);
 
-	ProgramSize metric(m_program, 2);
+	ProgramSize metric(m_program, nullptr, 2);
 	size_t fitness = metric.evaluate(m_chromosome);
 
 	BOOST_TEST(fitness != m_program.codeSize());
@@ -131,7 +167,7 @@ BOOST_FIXTURE_TEST_CASE(evaluate_should_repeat_the_optimisation_specified_number
 
 BOOST_FIXTURE_TEST_CASE(evaluate_should_not_optimise_if_number_of_repetitions_is_zero, ProgramBasedMetricFixture)
 {
-	ProgramSize metric(m_program, 0);
+	ProgramSize metric(m_program, nullptr, 0);
 	size_t fitness = metric.evaluate(m_chromosome);
 
 	BOOST_TEST(fitness == m_program.codeSize());
@@ -143,7 +179,13 @@ BOOST_AUTO_TEST_SUITE(RelativeProgramSizeTest)
 
 BOOST_FIXTURE_TEST_CASE(evaluate_should_compute_the_size_ratio_between_optimised_program_and_original_program, ProgramBasedMetricFixture)
 {
-	BOOST_TEST(RelativeProgramSize(m_program, 3).evaluate(m_chromosome) == round(1000.0 * m_optimisedProgram.codeSize() / m_program.codeSize()));
+	BOOST_TEST(RelativeProgramSize(m_program, nullptr, 3).evaluate(m_chromosome) == round(1000.0 * m_optimisedProgram.codeSize() / m_program.codeSize()));
+}
+
+BOOST_FIXTURE_TEST_CASE(evaluate_should_be_able_to_use_program_cache_if_available, ProgramBasedMetricFixture)
+{
+	BOOST_TEST(RelativeProgramSize(nullopt, m_programCache, 3).evaluate(m_chromosome) == round(1000.0 * m_optimisedProgram.codeSize() / m_program.codeSize()));
+	BOOST_TEST(m_programCache->size() == m_chromosome.length());
 }
 
 BOOST_FIXTURE_TEST_CASE(evaluate_should_repeat_the_optimisation_specified_number_of_times, ProgramBasedMetricFixture)
@@ -151,17 +193,17 @@ BOOST_FIXTURE_TEST_CASE(evaluate_should_repeat_the_optimisation_specified_number
 	Program const& programOptimisedOnce = m_optimisedProgram;
 	Program programOptimisedTwice = optimisedProgram(programOptimisedOnce);
 
-	RelativeProgramSize metric(m_program, 3, 2);
+	RelativeProgramSize metric(m_program, nullptr, 3, 2);
 	size_t fitness = metric.evaluate(m_chromosome);
 
 	BOOST_TEST(fitness != 1000);
-	BOOST_TEST(fitness != RelativeProgramSize(programOptimisedTwice, 3, 1).evaluate(m_chromosome));
+	BOOST_TEST(fitness != RelativeProgramSize(programOptimisedTwice, nullptr, 3, 1).evaluate(m_chromosome));
 	BOOST_TEST(fitness == round(1000.0 * programOptimisedTwice.codeSize() / m_program.codeSize()));
 }
 
 BOOST_FIXTURE_TEST_CASE(evaluate_should_return_one_if_number_of_repetitions_is_zero, ProgramBasedMetricFixture)
 {
-	RelativeProgramSize metric(m_program, 3, 0);
+	RelativeProgramSize metric(m_program, nullptr, 3, 0);
 
 	BOOST_TEST(metric.evaluate(m_chromosome) == 1000);
 }
@@ -171,7 +213,7 @@ BOOST_FIXTURE_TEST_CASE(evaluate_should_return_one_if_the_original_program_size_
 	CharStream sourceStream = CharStream("{}", "");
 	Program program = get<Program>(Program::load(sourceStream));
 
-	RelativeProgramSize metric(program, 3);
+	RelativeProgramSize metric(program, nullptr, 3);
 
 	BOOST_TEST(metric.evaluate(m_chromosome) == 1000);
 	BOOST_TEST(metric.evaluate(Chromosome("")) == 1000);
@@ -181,11 +223,11 @@ BOOST_FIXTURE_TEST_CASE(evaluate_should_return_one_if_the_original_program_size_
 BOOST_FIXTURE_TEST_CASE(evaluate_should_multiply_the_result_by_scaling_factor, ProgramBasedMetricFixture)
 {
 	double sizeRatio = static_cast<double>(m_optimisedProgram.codeSize()) / m_program.codeSize();
-	BOOST_TEST(RelativeProgramSize(m_program, 0).evaluate(m_chromosome) == round(1.0 * sizeRatio));
+	BOOST_TEST(RelativeProgramSize(m_program, nullptr, 0).evaluate(m_chromosome) == round(1.0 * sizeRatio));
-	BOOST_TEST(RelativeProgramSize(m_program, 1).evaluate(m_chromosome) == round(10.0 * sizeRatio));
+	BOOST_TEST(RelativeProgramSize(m_program, nullptr, 1).evaluate(m_chromosome) == round(10.0 * sizeRatio));
-	BOOST_TEST(RelativeProgramSize(m_program, 2).evaluate(m_chromosome) == round(100.0 * sizeRatio));
+	BOOST_TEST(RelativeProgramSize(m_program, nullptr, 2).evaluate(m_chromosome) == round(100.0 * sizeRatio));
-	BOOST_TEST(RelativeProgramSize(m_program, 3).evaluate(m_chromosome) == round(1000.0 * sizeRatio));
+	BOOST_TEST(RelativeProgramSize(m_program, nullptr, 3).evaluate(m_chromosome) == round(1000.0 * sizeRatio));
-	BOOST_TEST(RelativeProgramSize(m_program, 4).evaluate(m_chromosome) == round(10000.0 * sizeRatio));
+	BOOST_TEST(RelativeProgramSize(m_program, nullptr, 4).evaluate(m_chromosome) == round(10000.0 * sizeRatio));
 }
 
 BOOST_AUTO_TEST_SUITE_END()

test/yulPhaser/Phaser.cpp

@@ -55,7 +55,7 @@ protected:
 	};
 };
 
-class FitnessMetricFactoryFixture
+class FixtureWithPrograms
 {
 protected:
 	vector<CharStream> m_sourceStreams = {
@@ -68,6 +68,11 @@ protected:
 		get<Program>(Program::load(m_sourceStreams[1])),
 		get<Program>(Program::load(m_sourceStreams[2])),
 	};
+};
+
+class FitnessMetricFactoryFixture: public FixtureWithPrograms
+{
+protected:
 	FitnessMetricFactory::Options m_options = {
 		/* metric = */ MetricChoice::CodeSize,
 		/* metricAggregator = */ MetricAggregatorChoice::Average,
@@ -154,7 +159,7 @@ BOOST_FIXTURE_TEST_CASE(build_should_create_metric_of_the_right_type, FitnessMet
 {
 	m_options.metric = MetricChoice::RelativeCodeSize;
 	m_options.metricAggregator = MetricAggregatorChoice::Sum;
-	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, {m_programs[0]});
+	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, {m_programs[0]}, {nullptr});
 	BOOST_REQUIRE(metric != nullptr);
 
 	auto sumMetric = dynamic_cast<FitnessMetricSum*>(metric.get());
@@ -172,7 +177,7 @@ BOOST_FIXTURE_TEST_CASE(build_should_respect_chromosome_repetitions_option, Fitn
 	m_options.metric = MetricChoice::CodeSize;
 	m_options.metricAggregator = MetricAggregatorChoice::Average;
 	m_options.chromosomeRepetitions = 5;
-	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, {m_programs[0]});
+	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, {m_programs[0]}, {nullptr});
 	BOOST_REQUIRE(metric != nullptr);
 
 	auto averageMetric = dynamic_cast<FitnessMetricAverage*>(metric.get());
@@ -190,7 +195,7 @@ BOOST_FIXTURE_TEST_CASE(build_should_set_relative_metric_scale, FitnessMetricFac
 	m_options.metric = MetricChoice::RelativeCodeSize;
 	m_options.metricAggregator = MetricAggregatorChoice::Average;
 	m_options.relativeMetricScale = 10;
-	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, {m_programs[0]});
+	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, {m_programs[0]}, {nullptr});
 	BOOST_REQUIRE(metric != nullptr);
 
 	auto averageMetric = dynamic_cast<FitnessMetricAverage*>(metric.get());
@@ -205,7 +210,11 @@ BOOST_FIXTURE_TEST_CASE(build_should_set_relative_metric_scale, FitnessMetricFac
 
 BOOST_FIXTURE_TEST_CASE(build_should_create_metric_for_each_input_program, FitnessMetricFactoryFixture)
 {
-	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, m_programs);
+	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(
+		m_options,
+		m_programs,
+		vector<shared_ptr<ProgramCache>>(m_programs.size(), nullptr)
+	);
 	BOOST_REQUIRE(metric != nullptr);
 
 	auto combinedMetric = dynamic_cast<FitnessMetricCombination*>(metric.get());
@@ -213,6 +222,31 @@ BOOST_FIXTURE_TEST_CASE(build_should_create_metric_for_each_input_program, Fitne
 	BOOST_REQUIRE(combinedMetric->metrics().size() == m_programs.size());
 }
 
+BOOST_FIXTURE_TEST_CASE(build_should_pass_program_caches_to_metrics, FitnessMetricFactoryFixture)
+{
+	assert(m_programs.size() == 3);
+	vector<shared_ptr<ProgramCache>> caches = {
+		make_shared<ProgramCache>(m_programs[0]),
+		make_shared<ProgramCache>(m_programs[1]),
+		make_shared<ProgramCache>(m_programs[2]),
+	};
+
+	m_options.metric = MetricChoice::RelativeCodeSize;
+	unique_ptr<FitnessMetric> metric = FitnessMetricFactory::build(m_options, m_programs, caches);
+	BOOST_REQUIRE(metric != nullptr);
+
+	auto combinedMetric = dynamic_cast<FitnessMetricCombination*>(metric.get());
+	BOOST_REQUIRE(combinedMetric != nullptr);
+	BOOST_REQUIRE(combinedMetric->metrics().size() == caches.size());
+
+	for (size_t i = 0; i < caches.size(); ++i)
+	{
+		auto programBasedMetric = dynamic_cast<ProgramBasedMetric*>(combinedMetric->metrics()[i].get());
+		BOOST_REQUIRE(programBasedMetric != nullptr);
+		BOOST_TEST(programBasedMetric->programCache() == caches[i].get());
+	}
+}
+
 BOOST_AUTO_TEST_SUITE_END()
 BOOST_AUTO_TEST_SUITE(PopulationFactoryTest)
 
@@ -317,6 +351,35 @@ BOOST_FIXTURE_TEST_CASE(build_should_combine_populations_from_all_sources, Poula
 	BOOST_TEST(count(begin, end, Individual(Chromosome("fcL"), *m_fitnessMetric)) >= 2);
 }
 
+
+BOOST_AUTO_TEST_SUITE_END()
+BOOST_AUTO_TEST_SUITE(ProgramCacheFactoryTest)
+
+BOOST_FIXTURE_TEST_CASE(build_should_create_cache_for_each_input_program_if_cache_enabled, FixtureWithPrograms)
+{
+	ProgramCacheFactory::Options options{/* programCacheEnabled = */ true};
+	vector<shared_ptr<ProgramCache>> caches = ProgramCacheFactory::build(options, m_programs);
+	assert(m_programs.size() >= 2 && "There must be at least 2 programs for this test to be meaningful");
+
+	BOOST_TEST(caches.size() == m_programs.size());
+	for (size_t i = 0; i < m_programs.size(); ++i)
+	{
+		BOOST_REQUIRE(caches[i] != nullptr);
+		BOOST_TEST(toString(caches[i]->program()) == toString(m_programs[i]));
+	}
+}
+
+BOOST_FIXTURE_TEST_CASE(build_should_return_nullptr_for_each_input_program_if_cache_disabled, FixtureWithPrograms)
+{
+	ProgramCacheFactory::Options options{/* programCacheEnabled = */ false};
+	vector<shared_ptr<ProgramCache>> caches = ProgramCacheFactory::build(options, m_programs);
+	assert(m_programs.size() >= 2 && "There must be at least 2 programs for this test to be meaningful");
+
+	BOOST_TEST(caches.size() == m_programs.size());
+	for (size_t i = 0; i < m_programs.size(); ++i)
+		BOOST_TEST(caches[i] == nullptr);
+}
+
 BOOST_AUTO_TEST_SUITE_END()
 BOOST_AUTO_TEST_SUITE(ProgramFactoryTest)
 

test/yulPhaser/ProgramCache.cpp

@@ -0,0 +1,207 @@
+/*
+	This file is part of solidity.
+
+	solidity is free software: you can redistribute it and/or modify
+	it under the terms of the GNU General Public License as published by
+	the Free Software Foundation, either version 3 of the License, or
+	(at your option) any later version.
+
+	solidity is distributed in the hope that it will be useful,
+	but WITHOUT ANY WARRANTY; without even the implied warranty of
+	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+	GNU General Public License for more details.
+
+	You should have received a copy of the GNU General Public License
+	along with solidity.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <tools/yulPhaser/ProgramCache.h>
+#include <tools/yulPhaser/Chromosome.h>
+
+#include <liblangutil/CharStream.h>
+
+#include <libsolutil/CommonIO.h>
+
+#include <boost/test/unit_test.hpp>
+
+#include <string>
+#include <set>
+
+using namespace std;
+using namespace solidity::util;
+using namespace solidity::langutil;
+using namespace solidity::yul;
+
+namespace solidity::phaser::test
+{
+
+class ProgramCacheFixture
+{
+protected:
+	static constexpr char SampleSourceCode[] =
+		"{\n"
+		"    for { let i := 0 } not(eq(i, 15)) { i := add(i, 1) }\n"
+		"    {\n"
+		"        let x := 1\n"
+		"        mstore(i, 2)\n"
+		"    }\n"
+		"}\n";
+
+	Program optimisedProgram(Program _program, string _abbreviatedOptimisationSteps) const
+	{
+		Program result = move(_program);
+		result.optimise(Chromosome(_abbreviatedOptimisationSteps).optimisationSteps());
+		return result;
+	}
+
+	static set<string> cachedKeys(ProgramCache const& _programCache)
+	{
+		 set<string> keys;
+		for (auto pair = _programCache.entries().begin(); pair != _programCache.entries().end(); ++pair)
+			keys.insert(pair->first);
+
+		return keys;
+	}
+
+	CharStream m_sourceStream = CharStream(SampleSourceCode, "program-cache-test");
+	Program m_program = get<Program>(Program::load(m_sourceStream));
+	ProgramCache m_programCache{m_program};
+};
+
+BOOST_AUTO_TEST_SUITE(Phaser)
+BOOST_AUTO_TEST_SUITE(ProgramCacheTest)
+
+BOOST_FIXTURE_TEST_CASE(optimiseProgram_should_apply_optimisation_steps_to_program, ProgramCacheFixture)
+{
+	Program expectedProgram = optimisedProgram(m_program, "IuO");
+	assert(toString(expectedProgram) != toString(m_program));
+
+	Program cachedProgram = m_programCache.optimiseProgram("IuO");
+
+	BOOST_TEST(toString(cachedProgram) == toString(expectedProgram));
+}
+
+BOOST_FIXTURE_TEST_CASE(optimiseProgram_should_store_programs_for_all_prefixes, ProgramCacheFixture)
+{
+	Program programI = optimisedProgram(m_program, "I");
+	Program programIu = optimisedProgram(programI, "u");
+	Program programIuO = optimisedProgram(programIu, "O");
+	assert(toString(m_program) != toString(programI));
+	assert(toString(m_program) != toString(programIu));
+	assert(toString(m_program) != toString(programIuO));
+	assert(toString(programI) != toString(programIu));
+	assert(toString(programI) != toString(programIuO));
+	assert(toString(programIu) != toString(programIuO));
+
+	BOOST_REQUIRE(m_programCache.size() == 0);
+
+	Program cachedProgram = m_programCache.optimiseProgram("IuO");
+
+	BOOST_TEST(toString(cachedProgram) == toString(programIuO));
+
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"I", "Iu", "IuO"}));
+	BOOST_TEST(toString(*m_programCache.find("I")) == toString(programI));
+	BOOST_TEST(toString(*m_programCache.find("Iu")) == toString(programIu));
+	BOOST_TEST(toString(*m_programCache.find("IuO")) == toString(programIuO));
+}
+
+BOOST_FIXTURE_TEST_CASE(optimiseProgram_should_repeat_the_chromosome_requested_number_of_times, ProgramCacheFixture)
+{
+	string steps = "IuOIuO";
+
+	Program cachedProgram = m_programCache.optimiseProgram("IuO", 2);
+
+	ProgramCache cacheNoRepetitions(m_program);
+	Program cachedProgramNoRepetitions = cacheNoRepetitions.optimiseProgram("IuOIuO");
+
+	BOOST_TEST(toString(cachedProgram) == toString(cachedProgramNoRepetitions));
+
+	for (size_t size = 1; size <= 6; ++size)
+	{
+		BOOST_REQUIRE(m_programCache.contains(steps.substr(0, size)));
+		BOOST_REQUIRE(cacheNoRepetitions.contains(steps.substr(0, size)));
+		BOOST_TEST(
+			toString(*cacheNoRepetitions.find(steps.substr(0, size))) ==
+			toString(*m_programCache.find(steps.substr(0, size)))
+		);
+	}
+}
+
+BOOST_FIXTURE_TEST_CASE(optimiseProgram_should_reuse_the_longest_prefix_and_move_it_to_the_next_round, ProgramCacheFixture)
+{
+	BOOST_TEST(m_programCache.currentRound() == 0);
+
+	m_programCache.optimiseProgram("Iu");
+	m_programCache.optimiseProgram("Ia");
+	m_programCache.startRound(1);
+
+	BOOST_TEST(m_programCache.currentRound() == 1);
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"I", "Iu", "Ia"}));
+	BOOST_TEST(m_programCache.entries().find("I")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("Iu")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("Ia")->second.roundNumber == 0);
+
+	m_programCache.optimiseProgram("IuOI");
+
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"I", "Iu", "Ia", "IuO", "IuOI"}));
+	BOOST_TEST(m_programCache.entries().find("I")->second.roundNumber == 1);
+	BOOST_TEST(m_programCache.entries().find("Iu")->second.roundNumber == 1);
+	BOOST_TEST(m_programCache.entries().find("Ia")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("IuO")->second.roundNumber == 1);
+	BOOST_TEST(m_programCache.entries().find("IuOI")->second.roundNumber == 1);
+}
+
+BOOST_FIXTURE_TEST_CASE(startRound_should_remove_entries_older_than_two_rounds, ProgramCacheFixture)
+{
+	BOOST_TEST(m_programCache.currentRound() == 0);
+	BOOST_TEST(m_programCache.size() == 0);
+
+	m_programCache.optimiseProgram("Iu");
+
+	BOOST_TEST(m_programCache.currentRound() == 0);
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"I", "Iu"}));
+	BOOST_TEST(m_programCache.entries().find("I")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("Iu")->second.roundNumber == 0);
+
+	m_programCache.optimiseProgram("a");
+
+	BOOST_TEST(m_programCache.currentRound() == 0);
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"I", "Iu", "a"}));
+	BOOST_TEST(m_programCache.entries().find("I")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("Iu")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("a")->second.roundNumber == 0);
+
+	m_programCache.startRound(1);
+
+	BOOST_TEST(m_programCache.currentRound() == 1);
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"I", "Iu", "a"}));
+	BOOST_TEST(m_programCache.entries().find("I")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("Iu")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("a")->second.roundNumber == 0);
+
+	m_programCache.optimiseProgram("af");
+
+	BOOST_TEST(m_programCache.currentRound() == 1);
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"I", "Iu", "a", "af"}));
+	BOOST_TEST(m_programCache.entries().find("I")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("Iu")->second.roundNumber == 0);
+	BOOST_TEST(m_programCache.entries().find("a")->second.roundNumber == 1);
+	BOOST_TEST(m_programCache.entries().find("af")->second.roundNumber == 1);
+
+	m_programCache.startRound(2);
+
+	BOOST_TEST(m_programCache.currentRound() == 2);
+	BOOST_REQUIRE((cachedKeys(m_programCache) == set<string>{"a", "af"}));
+	BOOST_TEST(m_programCache.entries().find("a")->second.roundNumber == 1);
+	BOOST_TEST(m_programCache.entries().find("af")->second.roundNumber == 1);
+
+	m_programCache.startRound(3);
+
+	BOOST_TEST(m_programCache.currentRound() == 3);
+	BOOST_TEST(m_programCache.size() == 0);
+}
+
+BOOST_AUTO_TEST_SUITE_END()
+BOOST_AUTO_TEST_SUITE_END()
+
+}

tools/CMakeLists.txt

@@ -35,6 +35,8 @@ add_executable(yul-phaser
 	yulPhaser/PairSelections.cpp
 	yulPhaser/Selections.h
 	yulPhaser/Selections.cpp
+	yulPhaser/ProgramCache.h
+	yulPhaser/ProgramCache.cpp
 	yulPhaser/Program.h
 	yulPhaser/Program.cpp
 	yulPhaser/SimulationRNG.h

tools/yulPhaser/AlgorithmRunner.cpp

@@ -31,9 +31,12 @@ using namespace solidity::phaser;
 void AlgorithmRunner::run(GeneticAlgorithm& _algorithm)
 {
 	populationAutosave();
+	cacheClear();
 
 	for (size_t round = 0; !m_options.maxRounds.has_value() || round < m_options.maxRounds.value(); ++round)
 	{
+		cacheStartRound(round + 1);
+
 		m_population = _algorithm.runNextRound(m_population);
 		randomiseDuplicates();
 
@@ -66,6 +69,20 @@ void AlgorithmRunner::populationAutosave() const
 	);
 }
 
+void AlgorithmRunner::cacheClear()
+{
+	for (auto& cache: m_programCaches)
+		if (cache != nullptr)
+			cache->clear();
+}
+
+void AlgorithmRunner::cacheStartRound(size_t _roundNumber)
+{
+	for (auto& cache: m_programCaches)
+		if (cache != nullptr)
+			cache->startRound(_roundNumber);
+}
+
 void AlgorithmRunner::randomiseDuplicates()
 {
 	if (m_options.randomiseDuplicates)

tools/yulPhaser/AlgorithmRunner.h

@@ -22,6 +22,7 @@
 
 #include <tools/yulPhaser/GeneticAlgorithms.h>
 #include <tools/yulPhaser/Population.h>
+#include <tools/yulPhaser/ProgramCache.h>
 
 #include <optional>
 #include <ostream>
@@ -50,10 +51,12 @@ public:
 
 	AlgorithmRunner(
 		Population _initialPopulation,
+		std::vector<std::shared_ptr<ProgramCache>> _programCaches,
 		Options _options,
 		std::ostream& _outputStream
 	):
 		m_population(std::move(_initialPopulation)),
+		m_programCaches(std::move(_programCaches)),
 		m_options(std::move(_options)),
 		m_outputStream(_outputStream) {}
 
@@ -65,6 +68,9 @@ public:
 private:
 	void populationAutosave() const;
 	void randomiseDuplicates();
+	void cacheClear();
+	void cacheStartRound(size_t _roundNumber);
+
 	static Population randomiseDuplicates(
 		Population _population,
 		size_t _minChromosomeLength,
@@ -72,6 +78,7 @@ private:
 	);
 
 	Population m_population;
+	std::vector<std::shared_ptr<ProgramCache>> m_programCaches;
 	Options m_options;
 	std::ostream& m_outputStream;
 };

tools/yulPhaser/FitnessMetrics.cpp

@@ -17,14 +17,36 @@
 
 #include <tools/yulPhaser/FitnessMetrics.h>
 
+#include <libsolutil/CommonIO.h>
+
 #include <cmath>
 
 using namespace std;
+using namespace solidity::util;
 using namespace solidity::phaser;
 
-Program ProgramBasedMetric::optimisedProgram(Chromosome const& _chromosome) const
+Program const& ProgramBasedMetric::program() const
 {
-	Program programCopy = m_program;
+	if (m_programCache == nullptr)
+		return m_program.value();
+	else
+		return m_programCache->program();
+}
+
+Program ProgramBasedMetric::optimisedProgram(Chromosome const& _chromosome)
+{
+	if (m_programCache == nullptr)
+		return optimisedProgramNoCache(_chromosome);
+
+	return m_programCache->optimiseProgram(
+		toString(_chromosome),
+		m_repetitionCount
+	);
+}
+
+Program ProgramBasedMetric::optimisedProgramNoCache(Chromosome const& _chromosome) const
+{
+	Program programCopy = program();
 	for (size_t i = 0; i < m_repetitionCount; ++i)
 		programCopy.optimise(_chromosome.optimisationSteps());
 

tools/yulPhaser/FitnessMetrics.h

@@ -22,8 +22,10 @@
 
 #include <tools/yulPhaser/Chromosome.h>
 #include <tools/yulPhaser/Program.h>
+#include <tools/yulPhaser/ProgramCache.h>
 
 #include <cstddef>
+#include <optional>
 
 namespace solidity::phaser
 {
@@ -50,7 +52,7 @@ public:
  * Abstract base class for fitness metrics that return values based on program size.
  *
  * The class provides utilities for optimising programs according to the information stored in
- * chromosomes.
+ * chromosomes. Allows using @a ProgramCache.
  *
  * It can also store weights for the @a CodeSize metric. It does not do anything with
  * them because it does not actually compute the code size but they are readily available for use
@@ -60,19 +62,27 @@ class ProgramBasedMetric: public FitnessMetric
 {
 public:
 	explicit ProgramBasedMetric(
-		Program _program,
+		std::optional<Program> _program,
+		std::shared_ptr<ProgramCache> _programCache,
 		size_t _repetitionCount = 1
 	):
 		m_program(std::move(_program)),
-		m_repetitionCount(_repetitionCount) {}
+		m_programCache(std::move(_programCache)),
+		m_repetitionCount(_repetitionCount)
+	{
+		assert(m_program.has_value() == (m_programCache == nullptr));
+	}
 
-	Program const& program() const { return m_program; }
+	Program const& program() const;
+	ProgramCache const* programCache() const { return m_programCache.get(); }
 	size_t repetitionCount() const { return m_repetitionCount; }
 
-	Program optimisedProgram(Chromosome const& _chromosome) const;
+	Program optimisedProgram(Chromosome const& _chromosome);
+	Program optimisedProgramNoCache(Chromosome const& _chromosome) const;
 
 private:
-	Program m_program;
+	std::optional<Program> m_program;
+	std::shared_ptr<ProgramCache> m_programCache;
 	size_t m_repetitionCount;
 };
 
@@ -98,11 +108,12 @@ class RelativeProgramSize: public ProgramBasedMetric
 {
 public:
 	explicit RelativeProgramSize(
-		Program _program,
+		std::optional<Program> _program,
+		std::shared_ptr<ProgramCache> _programCache,
 		size_t _fixedPointPrecision,
 		size_t _repetitionCount = 1
 	):
-		ProgramBasedMetric(std::move(_program), _repetitionCount),
+		ProgramBasedMetric(std::move(_program), std::move(_programCache), _repetitionCount),
 		m_fixedPointPrecision(_fixedPointPrecision) {}
 
 	size_t fixedPointPrecision() const { return m_fixedPointPrecision; }

tools/yulPhaser/Phaser.cpp

@@ -158,9 +158,11 @@ FitnessMetricFactory::Options FitnessMetricFactory::Options::fromCommandLine(po:
 
 unique_ptr<FitnessMetric> FitnessMetricFactory::build(
 	Options const& _options,
-	vector<Program> _programs
+	vector<Program> _programs,
+	vector<shared_ptr<ProgramCache>> _programCaches
 )
 {
+	assert(_programCaches.size() == _programs.size());
 	assert(_programs.size() > 0 && "Validations should prevent this from being executed with zero files.");
 
 	vector<shared_ptr<FitnessMetric>> metrics;
@@ -168,9 +170,10 @@ unique_ptr<FitnessMetric> FitnessMetricFactory::build(
 	{
 		case MetricChoice::CodeSize:
 		{
-			for (Program& program: _programs)
+			for (size_t i = 0; i < _programs.size(); ++i)
 				metrics.push_back(make_unique<ProgramSize>(
-					move(program),
+					_programCaches[i] != nullptr ? optional<Program>{} : move(_programs[i]),
+					move(_programCaches[i]),
 					_options.chromosomeRepetitions
 				));
 
@@ -178,9 +181,10 @@ unique_ptr<FitnessMetric> FitnessMetricFactory::build(
 		}
 		case MetricChoice::RelativeCodeSize:
 		{
-			for (Program& program: _programs)
+			for (size_t i = 0; i < _programs.size(); ++i)
 				metrics.push_back(make_unique<RelativeProgramSize>(
-					move(program),
+					_programCaches[i] != nullptr ? optional<Program>{} : move(_programs[i]),
+					move(_programCaches[i]),
 					_options.relativeMetricScale,
 					_options.chromosomeRepetitions
 				));
@@ -281,6 +285,25 @@ Population PopulationFactory::buildFromFile(
 	return buildFromStrings(readLinesFromFile(_filePath), move(_fitnessMetric));
 }
 
+ProgramCacheFactory::Options ProgramCacheFactory::Options::fromCommandLine(po::variables_map const& _arguments)
+{
+	return {
+		_arguments["program-cache"].as<bool>(),
+	};
+}
+
+vector<shared_ptr<ProgramCache>> ProgramCacheFactory::build(
+	Options const& _options,
+	vector<Program> _programs
+)
+{
+	vector<shared_ptr<ProgramCache>> programCaches;
+	for (Program& program: _programs)
+		programCaches.push_back(_options.programCacheEnabled ? make_shared<ProgramCache>(move(program)) : nullptr);
+
+	return programCaches;
+}
+
 ProgramFactory::Options ProgramFactory::Options::fromCommandLine(po::variables_map const& _arguments)
 {
 	return {
@@ -507,6 +530,19 @@ Phaser::CommandLineDescription Phaser::buildCommandLineDescription()
 	;
 	keywordDescription.add(metricsDescription);
 
+	po::options_description cacheDescription("CACHE", lineLength, minDescriptionLength);
+	cacheDescription.add_options()
+		(
+			"program-cache",
+			po::bool_switch(),
+			"Enables caching of intermediate programs corresponding to chromosome prefixes.\n"
+			"This speeds up fitness evaluation by a lot but eats tons of memory if the chromosomes are long. "
+			"Disabled by default since there's currently no way to set an upper limit on memory usage but "
+			"highly recommended if your computer has enough RAM."
+		)
+	;
+	keywordDescription.add(cacheDescription);
+
 	po::positional_options_description positionalDescription;
 	positionalDescription.add("input-files", -1);
 
@@ -562,12 +598,15 @@ AlgorithmRunner::Options Phaser::buildAlgorithmRunnerOptions(po::variables_map c
 void Phaser::runAlgorithm(po::variables_map const& _arguments)
 {
 	auto programOptions = ProgramFactory::Options::fromCommandLine(_arguments);
+	auto cacheOptions = ProgramCacheFactory::Options::fromCommandLine(_arguments);
 	auto metricOptions = FitnessMetricFactory::Options::fromCommandLine(_arguments);
 	auto populationOptions = PopulationFactory::Options::fromCommandLine(_arguments);
 	auto algorithmOptions = GeneticAlgorithmFactory::Options::fromCommandLine(_arguments);
 
 	vector<Program> programs = ProgramFactory::build(programOptions);
-	unique_ptr<FitnessMetric> fitnessMetric = FitnessMetricFactory::build(metricOptions, move(programs));
+	vector<shared_ptr<ProgramCache>> programCaches = ProgramCacheFactory::build(cacheOptions, programs);
+
+	unique_ptr<FitnessMetric> fitnessMetric = FitnessMetricFactory::build(metricOptions, move(programs), programCaches);
 	Population population = PopulationFactory::build(populationOptions, move(fitnessMetric));
 
 	unique_ptr<GeneticAlgorithm> geneticAlgorithm = GeneticAlgorithmFactory::build(
@@ -575,6 +614,6 @@ void Phaser::runAlgorithm(po::variables_map const& _arguments)
 		population.individuals().size()
 	);
 
-	AlgorithmRunner algorithmRunner(population, buildAlgorithmRunnerOptions(_arguments), cout);
+	AlgorithmRunner algorithmRunner(population, move(programCaches), buildAlgorithmRunnerOptions(_arguments), cout);
 	algorithmRunner.run(*geneticAlgorithm);
 }

tools/yulPhaser/Phaser.h

@@ -45,6 +45,7 @@ class FitnessMetric;
 class GeneticAlgorithm;
 class Population;
 class Program;
+class ProgramCache;
 
 enum class Algorithm
 {
@@ -119,7 +120,8 @@ public:
 
 	static std::unique_ptr<FitnessMetric> build(
 		Options const& _options,
-		std::vector<Program> _programs
+		std::vector<Program> _programs,
+		std::vector<std::shared_ptr<ProgramCache>> _programCaches
 	);
 };
 
@@ -160,6 +162,25 @@ public:
 	);
 };
 
+/**
+ * Builds and validates instances of @a ProgramCache.
+ */
+class ProgramCacheFactory
+{
+public:
+	struct Options
+	{
+		bool programCacheEnabled;
+
+		static Options fromCommandLine(boost::program_options::variables_map const& _arguments);
+	};
+
+	static std::vector<std::shared_ptr<ProgramCache>> build(
+		Options const& _options,
+		std::vector<Program> _programs
+	);
+};
+
 /**
  * Builds and validates instances of @a Program.
  */

tools/yulPhaser/ProgramCache.cpp

@@ -0,0 +1,94 @@
+/*
+	This file is part of solidity.
+
+	solidity is free software: you can redistribute it and/or modify
+	it under the terms of the GNU General Public License as published by
+	the Free Software Foundation, either version 3 of the License, or
+	(at your option) any later version.
+
+	solidity is distributed in the hope that it will be useful,
+	but WITHOUT ANY WARRANTY; without even the implied warranty of
+	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+	GNU General Public License for more details.
+
+	You should have received a copy of the GNU General Public License
+	along with solidity.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <tools/yulPhaser/ProgramCache.h>
+
+#include <libyul/optimiser/Suite.h>
+
+using namespace std;
+using namespace solidity::yul;
+using namespace solidity::phaser;
+
+Program ProgramCache::optimiseProgram(
+	string const& _abbreviatedOptimisationSteps,
+	size_t _repetitionCount
+)
+{
+	string targetOptimisations = _abbreviatedOptimisationSteps;
+	for (size_t i = 1; i < _repetitionCount; ++i)
+		targetOptimisations += _abbreviatedOptimisationSteps;
+
+	size_t prefixSize = 0;
+	for (size_t i = 1; i <= targetOptimisations.size(); ++i)
+	{
+		auto const& pair = m_entries.find(targetOptimisations.substr(0, i));
+		if (pair != m_entries.end())
+		{
+			pair->second.roundNumber = m_currentRound;
+			++prefixSize;
+		}
+		else
+			break;
+	}
+
+	Program intermediateProgram = (
+		prefixSize == 0 ?
+		m_program :
+		m_entries.at(targetOptimisations.substr(0, prefixSize)).program
+	);
+
+	for (size_t i = prefixSize + 1; i <= targetOptimisations.size(); ++i)
+	{
+		string stepName = OptimiserSuite::stepAbbreviationToNameMap().at(targetOptimisations[i - 1]);
+		intermediateProgram.optimise({stepName});
+
+		m_entries.insert({targetOptimisations.substr(0, i), {intermediateProgram, m_currentRound}});
+	}
+
+	return intermediateProgram;
+}
+
+void ProgramCache::startRound(size_t _roundNumber)
+{
+	assert(_roundNumber > m_currentRound);
+	m_currentRound = _roundNumber;
+
+	for (auto pair = m_entries.begin(); pair != m_entries.end();)
+	{
+		assert(pair->second.roundNumber < m_currentRound);
+
+		if (pair->second.roundNumber < m_currentRound - 1)
+			m_entries.erase(pair++);
+		else
+			++pair;
+	}
+}
+
+void ProgramCache::clear()
+{
+	m_entries.clear();
+	m_currentRound = 0;
+}
+
+Program const* ProgramCache::find(string const& _abbreviatedOptimisationSteps) const
+{
+	auto const& pair = m_entries.find(_abbreviatedOptimisationSteps);
+	if (pair == m_entries.end())
+		return nullptr;
+
+	return &(pair->second.program);
+}

tools/yulPhaser/ProgramCache.h

@@ -0,0 +1,91 @@
+/*
+	This file is part of solidity.
+
+	solidity is free software: you can redistribute it and/or modify
+	it under the terms of the GNU General Public License as published by
+	the Free Software Foundation, either version 3 of the License, or
+	(at your option) any later version.
+
+	solidity is distributed in the hope that it will be useful,
+	but WITHOUT ANY WARRANTY; without even the implied warranty of
+	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+	GNU General Public License for more details.
+
+	You should have received a copy of the GNU General Public License
+	along with solidity.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#pragma once
+
+#include <tools/yulPhaser/Program.h>
+
+#include <map>
+#include <string>
+
+namespace solidity::phaser
+{
+
+/**
+ * Structure used by @a ProgramCache to store intermediate programs and metadata associated
+ * with them.
+ */
+struct CacheEntry
+{
+	Program program;
+	size_t roundNumber;
+
+	CacheEntry(Program _program, size_t _roundNumber):
+		program(std::move(_program)),
+		roundNumber(_roundNumber) {}
+};
+
+/**
+ * Class that optimises programs one step at a time which allows it to store and later reuse the
+ * results of the intermediate steps.
+ *
+ * The cache keeps track of the current round number and associates newly created entries with it.
+ * @a startRound() must be called at the beginning of a round so that entries that are too old
+ * can be purged. The current strategy is to store programs corresponding to all possible prefixes
+ * encountered in the current and the previous rounds. Entries older than that get removed to
+ * conserve memory.
+ *
+ * The current strategy does speed things up (about 4:1 hit:miss ratio observed in my limited
+ * experiments) but there's room for improvement. We could fit more useful programs in
+ * the cache by being more picky about which ones we choose.
+ *
+ * There is currently no way to purge entries without starting a new round. Since the programs
+ * take a lot of memory, this may lead to the cache eating up all the available RAM if sequences are
+ * long and programs large. A limiter based on entry count or total program size would be useful.
+ */
+class ProgramCache
+{
+public:
+	explicit ProgramCache(Program _program):
+		m_program(std::move(_program)) {}
+
+	Program optimiseProgram(
+		std::string const& _abbreviatedOptimisationSteps,
+		size_t _repetitionCount = 1
+	);
+	void startRound(size_t _nextRoundNumber);
+	void clear();
+
+	size_t size() const { return m_entries.size(); }
+	Program const* find(std::string const& _abbreviatedOptimisationSteps) const;
+	bool contains(std::string const& _abbreviatedOptimisationSteps) const { return find(_abbreviatedOptimisationSteps) != nullptr; }
+
+	std::map<std::string, CacheEntry> const& entries() const { return m_entries; };
+	Program const& program() const { return m_program; }
+	size_t currentRound() const { return m_currentRound; }
+
+private:
+	// The best matching data structure here would be a trie of chromosome prefixes but since
+	// the programs are orders of magnitude larger than the prefixes, it does not really matter.
+	// A map should be good enough.
+	std::map<std::string, CacheEntry> m_entries;
+
+	Program m_program;
+	size_t m_currentRound = 0;
+};
+
+}