cc741ed8af
* begin state service * begin on the state trie idea * created beacon state structure * add in the full clone getter * return by value instead * add all setters * new state setters are being completed * arrays roots exposed * close to finishing all these headerssss * functionality complete * added in proto benchmark test * test for compatibility * add test for compat * comments fixed * add clone * add clone * remove underlying copies * make it immutable * integrate it into chainservice * revert * wrap up comments for package * address all comments and godocs * address all comments * clone the pending attestation properly * properly clone remaining items * tests pass fixed bug * begin using it instead of head state * prevent nil pointer exceptions * begin using new struct in db * integrated new type into db package * add proper nil checks * using new state in archiver * refactored much of core * editing all the precompute functions * done with most core refactor * fixed up some bugs in the clone comparisons * append current epoch atts * add missing setters * add new setters * fix other core methods * fix up transition * main service and forkchoice * fix rpc * integrated to powchain * some more changes * fix build * improve processing of deposits * fix error * prevent panic * comment * fix process att * gaz * fix up att process * resolve existing review comments * resolve another batch of gh comments * resolve broken cpt state * revise testutil to use the new state * begin updating the state transition func to pass in more compartmentalized args * finish editing transition function to return errors * block operations pretty much done with refactor * state transition fully refactored * got epoch processing completed * fix build in fork choice * fixing more of the build * fix up broken sync package * it builds nowww it buildssss * revert registry changes * Recompute on Read (#4627) * compute on read * fix up eth1 data votes * looking into slashings bug introduced in core/ * able to advance more slots * add logging * can now sync with testnet yay * remove the leaves algorithm and other merkle imports * expose initialize unsafe funcs * Update beacon-chain/db/kv/state.go * lint Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * More Optimizations for New State (#4641) * map optimization * more optimizations * use a custom hasher * comment * block operations optimizations * Update beacon-chain/state/types.go Co-Authored-By: Raul Jordan <raul@prysmaticlabs.com> * fixed up various operations to use the validator index map access Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * archiver tests pass * fixing cache tests * cache tests passing * edited validator tests * powchain tests passing * halfway thru sync tests * more sync test fixes * add in tests for state/ * working through rpc tests * assignments tests passed * almost done with rpc/beacon tests * resolved painful validator test * fixed up even more tests * resolve tests * fix build * reduce a randao mixes copy * fixes under //beacon-chain/blockchain/... * build //beacon-chain/core/... * fixes * Runtime Optimizations (#4648) * parallelize shuffling * clean up * lint * fix build * use callback to read from registry * fix array roots and size map * new improvements * reduce hash allocs * improved shuffling * terence's review * use different method * raul's comment * new array roots * remove clone in pre-compute * Update beacon-chain/state/types.go Co-Authored-By: Raul Jordan <raul@prysmaticlabs.com> * raul's review * lint * fix build issues * fix visibility Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * fix visibility * build works for all * fix blockchain test * fix a few tests * fix more tests * update validator in slashing * archiver passing * fixed rpc/validator * progress on core tests * resolve broken rpc tests * blockchain tests passed * fix up some tests in core * fix message diff * remove unnecessary save * Save validator after slashing * Update validators one by one * another update * fix everything * fix more precompute tests * fix blocks tests * more elegant fix * more helper fixes * change back ? * fix test * fix skip slot * fix test * reset caches * fix testutil * raceoff fixed * passing * Retrieve cached state in the beginning * lint * Fixed tests part 1 * Fixed rest of the tests * Minor changes to avoid copying, small refactor to reduce deplicated code * Handle att req for slot 0 * New beacon state: Only populate merkle layers as needed, copy merkle layers on copy/clone. (#4689) * Only populate merkle layers as needed, copy merkle layers on copy/clone. * use custom copy * Make maps of correct size * slightly fast, doesn't wait for lock Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com> * Target root can't be 0x00 * Don't use cache for current slot (may not be the right fix) * fixed up tests * Remove some copy for init sync. Not sure if it is safe enough for runtime though... testing... * Align with prev logic for process slots cachedState.Slot() < slot * Fix Initial Sync Flag (#4692) * fixes * fix up some test failures due to lack of nil checks * fix up some test failures due to lack of nil checks * fix up imports * revert some changes * imports Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * resolving further conflicts * Better skip slot cache (#4694) * Return copy of skip slot cache state, disable skip slot cache on sync * fix * Fix pruning * fix up issues with broken tests Co-authored-by: Nishant Das <nish1993@hotmail.com> Co-authored-by: Preston Van Loon <preston@prysmaticlabs.com> Co-authored-by: shayzluf <thezluf@gmail.com> Co-authored-by: terence tsao <terence@prysmaticlabs.com> Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com> |
||
|---|---|---|
| .. | ||
| benchmark_files | ||
| BUILD.bazel | ||
| pregen_test.go | ||
| pregen.go | ||
| README.md | ||
Benchmarks for Prysm State Transition
This package contains the functionality needed for benchmarking Prysms state transitions, this includes its block processing (with and without caching) and epoch processing functions. There is also a benchmark for HashTreeRoot on a large beacon state.
Benchmark Configuration
The following configs are in config.go:
ValidatorCount: Sets the amount of active validators to perform the benchmarks with. Default is 16384.AttestationsPerEpoch: Sets the amount of attestations per epoch for the benchmarks to perform with, this affects the amount of attestations in a full block and the amount of attestations per epoch in the state for theProcessEpochandHashTreeRootbenchmark. Default is 128.
Generating new SSZ files
Due to the sheer size of the benchmarking configurations (16384 validators), the files used for benchmarking are pregenerated so there's no wasted computations on generating a genesis state with 16384 validators. This should only be needed if there is a breaking spec change and the tests fail from SSZ issues.
To generate new files to use for benchmarking, run the below command in the root of Prysm.
bazel run //tools/benchmark-files-gen -- --output-dir $PRYSMPATH/shared/benchutil/benchmark_files/ --overwrite
Running the benchmarks
To run the ExecuteStateTransition benchmark:
bazel test //beacon-chain/core/state:go_default_test --test_filter=BenchmarkExecuteStateTransition_FullBlock --test_arg=-test.bench=BenchmarkExecuteStateTransition_FullBlock
To run the ExecuteStateTransition (with cache) benchmark:
bazel test //beacon-chain/core/state:go_default_test --test_filter=BenchmarkExecuteStateTransition_WithCache --test_arg=-test.bench=BenchmarkExecuteStateTransition_WithCache
To run the ProcessEpoch benchmark:
bazel test //beacon-chain/core/state:go_default_test --test_filter=BenchmarkProcessEpoch_2FullEpochs --test_arg=-test.bench=BenchmarkProcessEpoch_2FullEpochs
To run the HashTreeRoot benchmark:
bazel test //beacon-chain/core/state:go_default_test --test_filter=BenchmarkHashTreeRoot_FullState --test_arg=-test.bench=BenchmarkHashTreeRoot_FullState
To run the HashTreeRootState benchmark:
bazel test //beacon-chain/core/state:go_default_test --test_filter=BenchmarkHashTreeRootState_FullState --test_arg=-test.bench=BenchmarkHashTreeRootState_FullState
Extra flags needed to benchmark properly:
--nocache_test_results --test_arg=-test.v --test_timeout=2000 --test_arg=-test.cpuprofile=/tmp/cpu.profile --test_arg=-test.memprofile=/tmp/mem.profile --test_output=streamed
Current Results as of January 2020
BenchmarkExecuteStateTransition_FullBlock-4 20 2031438030 ns/op
BenchmarkExecuteStateTransition_WithCache-4 20 1857290454 ns/op
BenchmarkHashTreeRoot_FullState-4 50 297655834 ns/op
BenchmarkHashTreeRootState_FullState-4 50 155535883 ns/op