Add --attempt-cache-bust to evmtool benchmark subcommand by siladu · Pull Request #8985 · besu-eth/besu

siladu · 2025-07-21T06:36:05Z

Run each test case within each warmup and exec iteration. This attempts to warm the code without warming the data, i.e. avoid warming CPU caches. Benchmark must have sufficient number and variety of test cases to be effective.

Due to how --attempt-cache-bust works by having the test cases in the inner loop:

--warm-time and --exec-time are ignored
async-profiler is not supported

This PR also includes some minor changes:

Three extra EcAdd test cases.
Filter testCases in advance rather than inline (to help with cache busting loop logic)
Use SequencedMap to indicate test case ordering matters
Lower default execution iterations to 1000 since this appears to be sufficient to average out randomness.

Note, this is an extracted piece of the exploration from #8942 but does not include the extra EcAdd test cases copied from besu-native and so support for displaying invalid cases is removed.
This PR also does not include the error confidence or statistics.
Finally, "warm-invert" was renamed to "attempt-cache-bust".

Testing

--attempt-cache-bust=true has a similar overall runtime (for all benchmarks) to --attempt-cache-bust=false

--attempt-cache-bust=false

~/besu-attempt-cache-bust$ ./build/install/besu/bin/evmtool benchmark --native --warm-iterations=10000 --exec-iterations=1000
...
real	9m21.896s
user	12m27.294s
sys	0m15.877s

--attempt-cache-bust=true

~/besu-attempt-cache-bust$ time ./build/install/besu/bin/evmtool benchmark --native --warm-iterations=10000 --exec-iterations=1000 --attempt-cache-bust=true
...
real	9m30.421s
user	12m40.148s
sys	0m16.783s

and generally the throughput is slightly lower with cache busting, as expected.

Attaching an example run with cache busting disabled and enabled. Note, --native is set and --warm-iterations is set lower than default to 10,000 which might create a bit a variance especially in the Java-only benchmarks.

benchmark_m6a.2xlarge_warm-10000-exec-1000_attempt-cache-bust-false.log
benchmark_m6a.2xlarge_warm-10000-exec-1000_attempt-cache-bust-true.log

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

This means running all the test cases within one warm up iteration in order to invalidate CPU caches. Execution iterations are unaffected so only makes sense to use with --exec-iterations=1 Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

For each exec iteration, run through each test case in serial to attempt to bust the CPU caches. Requires tracking exec timings in a map and calculating average at the end. Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Extracted for reuse with warmup invert but didn't reuse in the end Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Warmup iterations remain at 100_000 Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

If run is invalid, throw runtime error, as before Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

siladu · 2025-07-21T10:22:59Z

Attaching results for 100_000 warmup which should be more accurate. Took ~85mins to complete each run...

benchmark_m6a.2xlarge_warm-100_000-exec-1000_attempt-cache-bust-false.log

benchmark_m6a.2xlarge_warm-100_000-exec-1000_attempt-cache-bust-true.log

ahamlat · 2025-07-21T13:47:31Z

With --attempt-cache-bust=false, EcAdd is much more slower. I think it is related to the execution and some overhead on the machine. Could you re-execute it ?

EcAdd                          |      150 gas |      896 gas |      8,963.6 ns |      16.73 MGps

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

ahamlat · 2025-07-21T13:49:53Z

ethereum/evmtool/src/main/java/org/hyperledger/besu/evmtool/BenchmarkSubCommand.java

  Optional<Integer> warmTime = Optional.empty();

+  @Option(
+      names = {"--attempt-cache-bust"},


I find the name not very clear, but it can be subjective. I suggest using --warm-code-only if that works for you.

As discussed will leave as is, but can discuss further next week.

ethereum/evmtool/src/main/java/org/hyperledger/besu/evmtool/benchmarks/BenchmarkExecutor.java

ahamlat

Few comments but nothing blocking.

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

siladu · 2025-07-25T19:51:21Z

More results, from m6a.2xlarge:

attempt-cache-bust=false with three different warmup/iteration numbers, 3 times each config...

Native AltBN128
--warm-iterations=20000
--exec-iterations=1000
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      165 gas |      1,645.3 ns |      91.17 MGps
EcAddMarius                    |      150 gas |      349 gas |      3,485.9 ns |      43.03 MGps
EcAddAmez1                     |      150 gas |      355 gas |      3,548.2 ns |      42.27 MGps
EcAddAmez2                     |      150 gas |      394 gas |      3,936.9 ns |      38.10 MGps
EcAddAmez3                     |      150 gas |      315 gas |      3,151.8 ns |      47.59 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      163 gas |      1,628.9 ns |      92.09 MGps
EcAddMarius                    |      150 gas |      341 gas |      3,410.2 ns |      43.99 MGps
EcAddAmez1                     |      150 gas |      361 gas |      3,614.8 ns |      41.50 MGps
EcAddAmez2                     |      150 gas |      389 gas |      3,891.0 ns |      38.55 MGps
EcAddAmez3                     |      150 gas |      327 gas |      3,269.7 ns |      45.88 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      144 gas |      1,442.9 ns |     103.96 MGps
EcAddMarius                    |      150 gas |      338 gas |      3,378.8 ns |      44.39 MGps
EcAddAmez1                     |      150 gas |      346 gas |      3,456.5 ns |      43.40 MGps
EcAddAmez2                     |      150 gas |      378 gas |      3,783.1 ns |      39.65 MGps
EcAddAmez3                     |      150 gas |      313 gas |      3,132.4 ns |      47.89 MGps
--


Native AltBN128
--warm-iterations=100000
--exec-iterations=1000
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      147 gas |      1,465.7 ns |     102.34 MGps
EcAddMarius                    |      150 gas |      343 gas |      3,433.1 ns |      43.69 MGps
EcAddAmez1                     |      150 gas |      350 gas |      3,499.7 ns |      42.86 MGps
EcAddAmez2                     |      150 gas |      401 gas |      4,009.9 ns |      37.41 MGps
EcAddAmez3                     |      150 gas |      293 gas |      2,932.6 ns |      51.15 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      166 gas |      1,660.0 ns |      90.36 MGps
EcAddMarius                    |      150 gas |    1,848 gas |     18,483.1 ns |       8.12 MGps
EcAddAmez1                     |      150 gas |      307 gas |      3,068.2 ns |      48.89 MGps
EcAddAmez2                     |      150 gas |      358 gas |      3,584.9 ns |      41.84 MGps
EcAddAmez3                     |      150 gas |      303 gas |      3,028.4 ns |      49.53 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      160 gas |      1,599.6 ns |      93.77 MGps
EcAddMarius                    |      150 gas |      335 gas |      3,348.0 ns |      44.80 MGps
EcAddAmez1                     |      150 gas |      365 gas |      3,646.7 ns |      41.13 MGps
EcAddAmez2                     |      150 gas |      329 gas |      3,287.1 ns |      45.63 MGps
EcAddAmez3                     |      150 gas |      307 gas |      3,070.9 ns |      48.85 MGps
--

Native AltBN128
--warm-iterations=100000
--exec-iterations=100000
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      225 gas |      2,246.4 ns |      66.77 MGps
EcAddMarius                    |      150 gas |      363 gas |      3,634.2 ns |      41.27 MGps
EcAddAmez1                     |      150 gas |      354 gas |      3,536.4 ns |      42.42 MGps
EcAddAmez2                     |      150 gas |      311 gas |      3,112.5 ns |      48.19 MGps
EcAddAmez3                     |      150 gas |      430 gas |      4,304.7 ns |      34.85 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      219 gas |      2,188.9 ns |      68.53 MGps
EcAddMarius                    |      150 gas |      388 gas |      3,879.9 ns |      38.66 MGps
EcAddAmez1                     |      150 gas |      346 gas |      3,457.9 ns |      43.38 MGps
EcAddAmez2                     |      150 gas |      383 gas |      3,827.9 ns |      39.19 MGps
EcAddAmez3                     |      150 gas |      344 gas |      3,443.2 ns |      43.56 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      207 gas |      2,074.4 ns |      72.31 MGps
EcAddMarius                    |      150 gas |      343 gas |      3,427.3 ns |      43.77 MGps
EcAddAmez1                     |      150 gas |      338 gas |      3,382.9 ns |      44.34 MGps
EcAddAmez2                     |      150 gas |      362 gas |      3,619.1 ns |      41.45 MGps
EcAddAmez3                     |      150 gas |      400 gas |      4,000.1 ns |      37.50 MGps

attempt-cache-bust=true...

Native AltBN128
--attempt-cache-bust=true (--warm-time and --exec-time ignored)
--warm-iterations=20000
--exec-iterations=1000
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      183 gas |      1,832.2 ns |      81.87 MGps
EcAddMarius                    |      150 gas |      367 gas |      3,670.6 ns |      40.86 MGps
EcAddAmez1                     |      150 gas |    1,028 gas |     10,284.8 ns |      14.58 MGps
EcAddAmez2                     |      150 gas |      353 gas |      3,532.6 ns |      42.46 MGps
EcAddAmez3                     |      150 gas |      345 gas |      3,453.4 ns |      43.44 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      183 gas |      1,826.7 ns |      82.12 MGps
EcAddMarius                    |      150 gas |      359 gas |      3,594.9 ns |      41.73 MGps
EcAddAmez1                     |      150 gas |      347 gas |      3,465.8 ns |      43.28 MGps
EcAddAmez2                     |      150 gas |    1,025 gas |     10,251.0 ns |      14.63 MGps
EcAddAmez3                     |      150 gas |      347 gas |      3,467.2 ns |      43.26 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      203 gas |      2,032.4 ns |      73.81 MGps
EcAddMarius                    |      150 gas |      368 gas |      3,675.5 ns |      40.81 MGps
EcAddAmez1                     |      150 gas |      364 gas |      3,637.7 ns |      41.23 MGps
EcAddAmez2                     |      150 gas |      358 gas |      3,580.5 ns |      41.89 MGps
EcAddAmez3                     |      150 gas |      356 gas |      3,557.9 ns |      42.16 MGps
--


Native AltBN128
--attempt-cache-bust=true (--warm-time and --exec-time ignored)
--warm-iterations=100000
--exec-iterations=1000
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      165 gas |      1,653.9 ns |      90.69 MGps
EcAddMarius                    |      150 gas |      321 gas |      3,211.3 ns |      46.71 MGps
EcAddAmez1                     |      150 gas |      321 gas |      3,212.5 ns |      46.69 MGps
EcAddAmez2                     |      150 gas |      316 gas |      3,160.5 ns |      47.46 MGps
EcAddAmez3                     |      150 gas |      319 gas |      3,194.9 ns |      46.95 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      157 gas |      1,566.6 ns |      95.75 MGps
EcAddMarius                    |      150 gas |      324 gas |      3,239.9 ns |      46.30 MGps
EcAddAmez1                     |      150 gas |      324 gas |      3,237.7 ns |      46.33 MGps
EcAddAmez2                     |      150 gas |      328 gas |      3,282.0 ns |      45.70 MGps
EcAddAmez3                     |      150 gas |      322 gas |      3,222.0 ns |      46.55 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      168 gas |      1,675.2 ns |      89.54 MGps
EcAddMarius                    |      150 gas |      347 gas |      3,471.8 ns |      43.20 MGps
EcAddAmez1                     |      150 gas |      344 gas |      3,443.9 ns |      43.56 MGps
EcAddAmez2                     |      150 gas |      354 gas |      3,536.4 ns |      42.42 MGps
EcAddAmez3                     |      150 gas |      339 gas |      3,389.7 ns |      44.25 MGps
--


Native AltBN128
--attempt-cache-bust=true (--warm-time and --exec-time ignored)
--warm-iterations=100000
--exec-iterations=100000
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      183 gas |      1,834.7 ns |      81.76 MGps
EcAddMarius                    |      150 gas |      413 gas |      4,126.3 ns |      36.35 MGps
EcAddAmez1                     |      150 gas |      344 gas |      3,436.0 ns |      43.66 MGps
EcAddAmez2                     |      150 gas |      444 gas |      4,442.6 ns |      33.76 MGps
EcAddAmez3                     |      150 gas |      419 gas |      4,187.1 ns |      35.82 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      179 gas |      1,795.0 ns |      83.57 MGps
EcAddMarius                    |      150 gas |      510 gas |      5,100.7 ns |      29.41 MGps
EcAddAmez1                     |      150 gas |      365 gas |      3,645.5 ns |      41.15 MGps
EcAddAmez2                     |      150 gas |      337 gas |      3,367.6 ns |      44.54 MGps
EcAddAmez3                     |      150 gas |      335 gas |      3,346.3 ns |      44.83 MGps
--
                               |  Actual cost | Derived Cost |  Iteration time |      Throughput
EcAdd                          |      150 gas |      265 gas |      2,653.3 ns |      56.53 MGps
EcAddMarius                    |      150 gas |      347 gas |      3,465.8 ns |      43.28 MGps
EcAddAmez1                     |      150 gas |      380 gas |      3,802.6 ns |      39.45 MGps
EcAddAmez2                     |      150 gas |      342 gas |      3,415.9 ns |      43.91 MGps
EcAddAmez3                     |      150 gas |      372 gas |      3,717.5 ns |      40.35 MGps

* docs: edited the badge that was displayed incorrectly (#8623) * Update README.md Signed-off-by: Olexandr88 <radole1203@gmail.com> * Update README.md Signed-off-by: Jason Frame <jasonwframe@gmail.com> --------- Signed-off-by: Olexandr88 <radole1203@gmail.com> Signed-off-by: Jason Frame <jasonwframe@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> Co-authored-by: Jason Frame <jasonwframe@gmail.com> Co-authored-by: Jason Frame <jason.frame@consensys.net> * getBlobsV2 prohibits partial responses (#8967) * getBlobsV2 to not return partial responses Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Change precompile name in eth_config (#8986) * Change precompile name in eth_config The precompile names for bn254 were encoded as bn256, which is a widely used but incorrect name. The EIP specified alt_bn128, for 128 bits of security, but cryptographic analysis has shown the security to be less than 128 bits, so the name referring to the field size is preferred. Signed-off-by: Danno Ferrin <danno@numisight.com> * fixed test Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Danno Ferrin <danno@numisight.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * remove separate qbft ATs gradle task and workflow (#8962) Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * remove deprecated method Quantity.getValue (#8968) * update tests to use non-deprecated method * remove deprecated method * remove getValue methods that overrode the removed method Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Clean up production of the fastSyncFactory in DefaultSynchronizer (#8975) * Clean up production of the fastSyncFactory in DefaultSynchronizer Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix NPE caused by using a mock SynchronizerConfiguration Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix NPE caused by using a mock SynchronizerConfiguration Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * spotless Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Avoid using syncMode directly Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> --------- Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * remove deprecated PoAMetricsService and IbftQueryService (#8940) * remove deprecated interfaces * update plugin api hash Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * refactor: central control point for genesis used in tests (#8819) * made a constant Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Rename block hash processor (#8993) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Generate distribution dependencies catalog (#8987) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * feat: add world view to traceStartBlock hook (#8981) * Add world view to trace start block Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * change known hash Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * fix unit tests Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> --------- Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Do not account for system calls in the tracer (#8992) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Remove PoW ATs which were replaced by QBFT miner based ATs (#8949) * Delete PoW ATs which were replaced by QBFT miner based ATs * Change another AT to used QBFT genesis * update node config to fix couple of AT failures * Update transaction validator AT to use 4 validators in QBFT cluster * Update Besueventsplugintest * Revert an accidentally deleted AT and now modified to use QBFT Signed-off-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> --------- Signed-off-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Co-authored-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Improve BENCHMARKING.md documentation (#9001) Signed-off-by: Ameziane H. <ameziane.hamlat@consensys.net> * Docker labels update (#9002) * fix docker build Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> * update docker labels to the new oci standard Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> --------- Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> * No RLP decoding of Bodies during the sync (fix #8242) (#8959) * By not decoding the bodies we use less memory and CPU, which allows us to increase the chain download parallelism. Signed-off-by: stefan.pingel@consensys.net <stefan.pingel@consensys.net> Signed-off-by: Stefan Pingel <16143240+pinges@users.noreply.github.com> Co-authored-by: Simon Dudley <simon.dudley@consensys.net> * Enable decoding for large RPC requests (#8877) Use maxRequestContentLength for EngineJsonRpcService bodyLimit and increase default value from 5MB to 128MB. This has a side effect of increasing the JsonRpcService limit. Use maxRequestContentLength to globally set maxStringLength for Vertx ObjectMapper when engine API enabled. Reusing maxRequestContentLength to keep aligned with HTTP body limit. String length should be comparable with byte size assuming UTF-8 encoding. Also remove redundant serialization of params for optionalList Similar to #8807 Signed-off-by: Simon Dudley <simon.dudley@consensys.net> * Add --attempt-cache-bust to evmtool benchmark subcommand (#8985) This means running all the test cases within one warmup or exec iteration in an attempt to invalidate CPU caches. Requires tracking exec timings in a map and calculating average at the end. Reduce warmup iteration to 20_000 and execution iterations to 1000, which appears to be as stable as 100_000. This helps reduces overall benchmark execution time (previously warmTime and execTime were prematurely reducing overall execution time). Use SequencedMap to indicate test case ordering matters. Signed-off-by: Simon Dudley <simon.dudley@consensys.net> * TransactionDecoder: handle empty bytes (#8997) * handle empty bytes * check for empty before getting tx type Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Add a message for disconnect message reasons (#9014) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Remove the possibility to create block before Byzantium hard-fork (#9005) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * chore: remove support for eth/67 (#9008) * remove eth/67 capability Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * fix naming Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * changelog Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Fix mock test Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Fix tests Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * fix tests Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Update CHANGELOG.md Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> --------- Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Improve info logs during block creation (#9020) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * Chore: Prevent repeated cli boolean flags for being accepted (#8898) * Chore: Prevent repeated cli boolean flags for being accepted Signed-off-by: Preeti <35308865+pr9t@users.noreply.github.com> * allow unmatched options Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Preeti <35308865+pr9t@users.noreply.github.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> * Eliminate AbstractRLPInput abstraction to reduce technical debt (#8885) Merge AbstractRLPInput (607 lines) and BytesValueRLPInput (84 lines) into a single concrete class implementing RLPInput directly. This eliminates the unnecessary abstraction layer where the abstract class had only one concrete implementation. Changes: - Merged all AbstractRLPInput logic into BytesValueRLPInput - Made previously protected methods private - Removed AbstractRLPInput.java file - All public APIs remain unchanged (zero breaking changes) - All tests pass unchanged Benefits: - Eliminates 15-30 minutes investigation time per RLP modification - Reduces cognitive load when working with RLP functionality - Simplifies maintenance with single concrete implementation - No performance or functionality impact Fixes #8884 🤖 Generated with [Claude Code](https://claude.ai/code) Signed-off-by: jflo <justin+github@florentine.us> Co-authored-by: Claude <noreply@anthropic.com> * Add WorldStateService to the plugin API (#9024) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * Change eth_config hashing and config members (#9015) * Change eth_config hashing and config members - Remove ConfigHash - move ForkID into config - invert key/value of precompiles Signed-off-by: Danno Ferrin <danno@numisight.com> * linting Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Danno Ferrin <danno@numisight.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Era1 import prepipeline qol improvements (#8964) * Add --era1-import-prepipeline-enabled and --era1-data-path flags Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Implement ERA1 file import prepipeline for full sync Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Filter out block 0 as this is included by default Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix SynchronizerOptionsTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * spotless Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix BesuCommandTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Add ability to load ERA1 files direct from HTTP source, fix bug in Era1Reader exposed by using HTTP source Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix Era1ReaderTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Add javadoc to InputStreamFactory Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Add default constructor to InputStreamFactory Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix Era1FileSourceTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Only produce era1 import prepipeline if sync mode is FULL Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Make era1 prepipeline concurrency configurable Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Use Integer instead of int in SynchronizerOptions Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Update era1 import prepipeline concurrency to 1 Since the bottleneck for this pipeline is typically the import block step, there is no need for more concurrency than a single pipeline gives. Signed-off-by: Matilda-Clerke <matilda.clerke@consensys.net> * Inline starting full sync chain downloader Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Simplify comment Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Use constant for era1 block count Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Use Era1FileReader as a launching point for more worker threads to achieve better parallelism Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> --------- Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> Signed-off-by: Matilda-Clerke <matilda.clerke@consensys.net> * Eliminate AbstractRLPOutput abstraction to reduce technical debt (#8890) - Consolidate AbstractRLPOutput logic into BytesValueRLPOutput - Remove unnecessary inheritance layer with single implementation - Maintain identical public API and functionality - Zero breaking changes to 150+ usage sites across codebase This follows the same pattern as recent AbstractRLPInput elimination, completing the removal of premature abstractions in the RLP module. Fixes #8889 🤖 Generated with [Claude Code](https://claude.ai/code) Signed-off-by: jflo <justin+github@florentine.us> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Olexandr88 <radole1203@gmail.com> Signed-off-by: Jason Frame <jasonwframe@gmail.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Signed-off-by: Danno Ferrin <danno@numisight.com> Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> Signed-off-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Signed-off-by: Ameziane H. <ameziane.hamlat@consensys.net> Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> Signed-off-by: stefan.pingel@consensys.net <stefan.pingel@consensys.net> Signed-off-by: Stefan Pingel <16143240+pinges@users.noreply.github.com> Signed-off-by: Simon Dudley <simon.dudley@consensys.net> Signed-off-by: Preeti <35308865+pr9t@users.noreply.github.com> Signed-off-by: jflo <justin+github@florentine.us> Signed-off-by: Matilda-Clerke <matilda.clerke@consensys.net> Co-authored-by: Olexandr88 <radole1203@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> Co-authored-by: Jason Frame <jasonwframe@gmail.com> Co-authored-by: Jason Frame <jason.frame@consensys.net> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Danno Ferrin <danno@numisight.com> Co-authored-by: Matilda-Clerke <matilda.clerke@consensys.net> Co-authored-by: Fabio Di Fabio <fabio.difabio@consensys.net> Co-authored-by: Bhanu Pulluri <59369753+pullurib@users.noreply.github.com> Co-authored-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Co-authored-by: ahamlat <ameziane.hamlat@consensys.net> Co-authored-by: Joshua Fernandes <joshua.fernandes@consensys.net> Co-authored-by: Stefan Pingel <16143240+pinges@users.noreply.github.com> Co-authored-by: Simon Dudley <simon.dudley@consensys.net> Co-authored-by: Preeti <35308865+pr9t@users.noreply.github.com> Co-authored-by: Claude <noreply@anthropic.com>

* docs: edited the badge that was displayed incorrectly (#8623) * Update README.md Signed-off-by: Olexandr88 <radole1203@gmail.com> * Update README.md Signed-off-by: Jason Frame <jasonwframe@gmail.com> --------- Signed-off-by: Olexandr88 <radole1203@gmail.com> Signed-off-by: Jason Frame <jasonwframe@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> Co-authored-by: Jason Frame <jasonwframe@gmail.com> Co-authored-by: Jason Frame <jason.frame@consensys.net> * getBlobsV2 prohibits partial responses (#8967) * getBlobsV2 to not return partial responses Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Change precompile name in eth_config (#8986) * Change precompile name in eth_config The precompile names for bn254 were encoded as bn256, which is a widely used but incorrect name. The EIP specified alt_bn128, for 128 bits of security, but cryptographic analysis has shown the security to be less than 128 bits, so the name referring to the field size is preferred. Signed-off-by: Danno Ferrin <danno@numisight.com> * fixed test Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Danno Ferrin <danno@numisight.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * remove separate qbft ATs gradle task and workflow (#8962) Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * remove deprecated method Quantity.getValue (#8968) * update tests to use non-deprecated method * remove deprecated method * remove getValue methods that overrode the removed method Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Clean up production of the fastSyncFactory in DefaultSynchronizer (#8975) * Clean up production of the fastSyncFactory in DefaultSynchronizer Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix NPE caused by using a mock SynchronizerConfiguration Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix NPE caused by using a mock SynchronizerConfiguration Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * spotless Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Avoid using syncMode directly Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> --------- Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * remove deprecated PoAMetricsService and IbftQueryService (#8940) * remove deprecated interfaces * update plugin api hash Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * refactor: central control point for genesis used in tests (#8819) * made a constant Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Rename block hash processor (#8993) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Generate distribution dependencies catalog (#8987) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * feat: add world view to traceStartBlock hook (#8981) * Add world view to trace start block Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * change known hash Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * fix unit tests Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> --------- Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Do not account for system calls in the tracer (#8992) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Remove PoW ATs which were replaced by QBFT miner based ATs (#8949) * Delete PoW ATs which were replaced by QBFT miner based ATs * Change another AT to used QBFT genesis * update node config to fix couple of AT failures * Update transaction validator AT to use 4 validators in QBFT cluster * Update Besueventsplugintest * Revert an accidentally deleted AT and now modified to use QBFT Signed-off-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> --------- Signed-off-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Co-authored-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Improve BENCHMARKING.md documentation (#9001) Signed-off-by: Ameziane H. <ameziane.hamlat@consensys.net> * Docker labels update (#9002) * fix docker build Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> * update docker labels to the new oci standard Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> --------- Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> * No RLP decoding of Bodies during the sync (fix #8242) (#8959) * By not decoding the bodies we use less memory and CPU, which allows us to increase the chain download parallelism. Signed-off-by: stefan.pingel@consensys.net <stefan.pingel@consensys.net> Signed-off-by: Stefan Pingel <16143240+pinges@users.noreply.github.com> Co-authored-by: Simon Dudley <simon.dudley@consensys.net> * Enable decoding for large RPC requests (#8877) Use maxRequestContentLength for EngineJsonRpcService bodyLimit and increase default value from 5MB to 128MB. This has a side effect of increasing the JsonRpcService limit. Use maxRequestContentLength to globally set maxStringLength for Vertx ObjectMapper when engine API enabled. Reusing maxRequestContentLength to keep aligned with HTTP body limit. String length should be comparable with byte size assuming UTF-8 encoding. Also remove redundant serialization of params for optionalList Similar to #8807 Signed-off-by: Simon Dudley <simon.dudley@consensys.net> * Add --attempt-cache-bust to evmtool benchmark subcommand (#8985) This means running all the test cases within one warmup or exec iteration in an attempt to invalidate CPU caches. Requires tracking exec timings in a map and calculating average at the end. Reduce warmup iteration to 20_000 and execution iterations to 1000, which appears to be as stable as 100_000. This helps reduces overall benchmark execution time (previously warmTime and execTime were prematurely reducing overall execution time). Use SequencedMap to indicate test case ordering matters. Signed-off-by: Simon Dudley <simon.dudley@consensys.net> * TransactionDecoder: handle empty bytes (#8997) * handle empty bytes * check for empty before getting tx type Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Add a message for disconnect message reasons (#9014) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Remove the possibility to create block before Byzantium hard-fork (#9005) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * chore: remove support for eth/67 (#9008) * remove eth/67 capability Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * fix naming Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * changelog Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Fix mock test Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Fix tests Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * fix tests Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Update CHANGELOG.md Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> --------- Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Improve info logs during block creation (#9020) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * Chore: Prevent repeated cli boolean flags for being accepted (#8898) * Chore: Prevent repeated cli boolean flags for being accepted Signed-off-by: Preeti <35308865+pr9t@users.noreply.github.com> * allow unmatched options Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Preeti <35308865+pr9t@users.noreply.github.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> * Eliminate AbstractRLPInput abstraction to reduce technical debt (#8885) Merge AbstractRLPInput (607 lines) and BytesValueRLPInput (84 lines) into a single concrete class implementing RLPInput directly. This eliminates the unnecessary abstraction layer where the abstract class had only one concrete implementation. Changes: - Merged all AbstractRLPInput logic into BytesValueRLPInput - Made previously protected methods private - Removed AbstractRLPInput.java file - All public APIs remain unchanged (zero breaking changes) - All tests pass unchanged Benefits: - Eliminates 15-30 minutes investigation time per RLP modification - Reduces cognitive load when working with RLP functionality - Simplifies maintenance with single concrete implementation - No performance or functionality impact Fixes #8884 🤖 Generated with [Claude Code](https://claude.ai/code) Signed-off-by: jflo <justin+github@florentine.us> Co-authored-by: Claude <noreply@anthropic.com> * Add WorldStateService to the plugin API (#9024) Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> * Change eth_config hashing and config members (#9015) * Change eth_config hashing and config members - Remove ConfigHash - move ForkID into config - invert key/value of precompiles Signed-off-by: Danno Ferrin <danno@numisight.com> * linting Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Danno Ferrin <danno@numisight.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Era1 import prepipeline qol improvements (#8964) * Add --era1-import-prepipeline-enabled and --era1-data-path flags Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Implement ERA1 file import prepipeline for full sync Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Filter out block 0 as this is included by default Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix SynchronizerOptionsTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * spotless Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix BesuCommandTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Add ability to load ERA1 files direct from HTTP source, fix bug in Era1Reader exposed by using HTTP source Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix Era1ReaderTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Add javadoc to InputStreamFactory Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Add default constructor to InputStreamFactory Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Fix Era1FileSourceTest Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Only produce era1 import prepipeline if sync mode is FULL Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Make era1 prepipeline concurrency configurable Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Use Integer instead of int in SynchronizerOptions Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Update era1 import prepipeline concurrency to 1 Since the bottleneck for this pipeline is typically the import block step, there is no need for more concurrency than a single pipeline gives. Signed-off-by: Matilda-Clerke <matilda.clerke@consensys.net> * Inline starting full sync chain downloader Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Simplify comment Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Use constant for era1 block count Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> * Use Era1FileReader as a launching point for more worker threads to achieve better parallelism Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> --------- Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> Signed-off-by: Matilda-Clerke <matilda.clerke@consensys.net> * Eliminate AbstractRLPOutput abstraction to reduce technical debt (#8890) - Consolidate AbstractRLPOutput logic into BytesValueRLPOutput - Remove unnecessary inheritance layer with single implementation - Maintain identical public API and functionality - Zero breaking changes to 150+ usage sites across codebase This follows the same pattern as recent AbstractRLPInput elimination, completing the removal of premature abstractions in the RLP module. Fixes #8889 🤖 Generated with [Claude Code](https://claude.ai/code) Signed-off-by: jflo <justin+github@florentine.us> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * [BESU-8973] Display configured target-gas-limit in startup logs (#8998) * [BESU-8973] Fix issue #8973 * Added `targetGasLimit` property and related setter to `ConfigurationOverviewBuilder` class * Added line to `build` function for `targetGasLimit` in `ConfigurationOverviewBuilder` class * Added set of `targetGasLimit` for `builder` object in `generateConfigurationOverview` of `BesuCommand` class * Set default value for `targetGasLimit` in `BesuCommand` class * Changed visibility of `normalizeGas` in `ConfigurationOverviewBuilder` class for usage in tests * Added `getDefaultGasLimitByChainId` method in `MergeCoordinator` * Added tests for `MergeCoordinator`, `ConfigurationOverviewBuilder` and `BesuCommandTest` * Tests for mainnet, testnet and dev with default and custom targetGasLimit * Replaced Sepolia with Hoodi for MergeCoordinatorTest Signed-off-by: Gabriele Del Monte <gabriele.delmonte@proton.me> * spotless Signed-off-by: Simon Dudley <simon.dudley@consensys.net> --------- Signed-off-by: Gabriele Del Monte <gabriele.delmonte@proton.me> Signed-off-by: Gabriele Del Monte <37625739+gabrieledm@users.noreply.github.com> Signed-off-by: Simon Dudley <simon.dudley@consensys.net> Co-authored-by: Simon Dudley <simon.dudley@consensys.net> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * [CHANGELOG] consolidate upcoming breaking changes (#9030) * remove unnecessary use of deprecated options * changelog Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * use dummy coinbase in simulation if not defined (#9031) Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Use system call for beacon root (#8982) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Backward sync - check we have peers before starting (#9003) * check if ready before first step * add peer count > 0 check to isReady * added a test for zero peers Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * Fix mixHash / prevRandao deserialization in BlockOverrides (#8983) * Fix mixHash / prevRandao deserialization in BlockOverrides Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Update plugin-api checksum Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Add deserialization test failing on main and passing on this branch Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Update ethereum/api/src/test/java/org/hyperledger/besu/ethereum/api/jsonrpc/internal/parameters/BlockOverridesParameterTest.java Co-authored-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Apply formatter Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> --------- Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> Co-authored-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Feat: add gas usage metric on eth_call (#9019) * WIP: update EthCall.java to capture gas usage metric * update EthJsonRpcMethods.java and JsonRpcMethodsFactory.java constructors to include metricsSystem * Update EthCallTest.java to mock gas usage metric * Update transactionSimulator error label to internal_error * Update CHANGELOG.md Signed-off-by: Reuben George <reubengeorge101@gmail.com> --------- Signed-off-by: Reuben George <reubengeorge101@gmail.com> Signed-off-by: Reuben George <69674713+georgereuben@users.noreply.github.com> Co-authored-by: Karim Taam <karim.t2am@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Remove CLI options for JMH and adapt benchmarks (#9025) PR allows for changing JMH settings in benchmarks themselves instead of gradle plugin but fixes forks to 3 for all benchamrks. Also done some tuning of CLZ and Keccak benchmarks. * [CHARTER] make explicit the voting for emeritus process (#8937) * make explicit the voting for emeritus process Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * headings formatting Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> * Bump prioritized transactions for blobs (#9037) Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Removed BlockExplorer and AbstractBlockScheduler due single implementation. (#9026) Signed-off-by: Marcos Serradilla Diez <marcos@io.builders> * sonarcloud failing due to deprecated actions/cache version (#9036) * sonarcloud failing due to deprecated actions/cache version Signed-off-by: jflo <justin+github@florentine.us> * new release just dropped Signed-off-by: jflo <justin+github@florentine.us> --------- Signed-off-by: jflo <justin+github@florentine.us> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * Include authorization list in the transaction call object and use it when building transaction during simulation (#8984) * Include code delegation authorizations in the call parameter and use when building transaction during simulation Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Add authorizationList deserialization test Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Test that transaction type is correct if authorizations present in call param Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Add javadoc Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Extract TransactionCallParameterBuilder creation to a separate method for consistency Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> * Include codeDelegationAuthorizations when building Transaction from CallParameter Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> --------- Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> Co-authored-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> * Improve P256Verify precompile performance (#9035) - Upgrade besu-native to 1.4.0. - Use BoringSSL by default and fallback to SECP256R1 signature algorithm which has its own native library (as well as Java fallback). - Add --Xp256verify-native-enabled to disable BoringSSL. - P256VerifyBenchmark switches between BoringSSL and Java. - Fixup P256VerifyBenchmark testcase names so they don't override each other in the Map. Signed-off-by: Simon Dudley <simon.dudley@consensys.net> Co-authored-by: garyschulte <garyschulte@gmail.com> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> * update checkout action to 4.2.2 (#9052) Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> * EngineNewPayload - respond with error if params invalid (#8729) * added test for duplicate requests * most errors don't result in INVALID status * return invalid for invalid block * new error code * changelog entry Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> --------- Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> * Fix merge main Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> --------- Signed-off-by: Olexandr88 <radole1203@gmail.com> Signed-off-by: Jason Frame <jasonwframe@gmail.com> Signed-off-by: Sally MacFarlane <macfarla.github@gmail.com> Signed-off-by: Danno Ferrin <danno@numisight.com> Signed-off-by: Matilda Clerke <matilda.clerke@consensys.net> Signed-off-by: Gabriel-Trintinalia <gabriel.trintinalia@consensys.net> Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net> Signed-off-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Signed-off-by: Ameziane H. <ameziane.hamlat@consensys.net> Signed-off-by: Joshua Fernandes <joshua.fernandes@consensys.net> Signed-off-by: stefan.pingel@consensys.net <stefan.pingel@consensys.net> Signed-off-by: Stefan Pingel <16143240+pinges@users.noreply.github.com> Signed-off-by: Simon Dudley <simon.dudley@consensys.net> Signed-off-by: Preeti <35308865+pr9t@users.noreply.github.com> Signed-off-by: jflo <justin+github@florentine.us> Signed-off-by: Matilda-Clerke <matilda.clerke@consensys.net> Signed-off-by: Gabriele Del Monte <gabriele.delmonte@proton.me> Signed-off-by: Gabriele Del Monte <37625739+gabrieledm@users.noreply.github.com> Signed-off-by: Miroslav Kovar <miroslavkovar@protonmail.com> Signed-off-by: Reuben George <reubengeorge101@gmail.com> Signed-off-by: Reuben George <69674713+georgereuben@users.noreply.github.com> Signed-off-by: Marcos Serradilla Diez <marcos@io.builders> Co-authored-by: Olexandr88 <radole1203@gmail.com> Co-authored-by: Justin Florentine <justin+github@florentine.us> Co-authored-by: Jason Frame <jasonwframe@gmail.com> Co-authored-by: Jason Frame <jason.frame@consensys.net> Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com> Co-authored-by: Danno Ferrin <danno@numisight.com> Co-authored-by: Matilda-Clerke <matilda.clerke@consensys.net> Co-authored-by: Fabio Di Fabio <fabio.difabio@consensys.net> Co-authored-by: Bhanu Pulluri <59369753+pullurib@users.noreply.github.com> Co-authored-by: Bhanu Pulluri <bhanu.pulluri@kaleido.io> Co-authored-by: ahamlat <ameziane.hamlat@consensys.net> Co-authored-by: Joshua Fernandes <joshua.fernandes@consensys.net> Co-authored-by: Stefan Pingel <16143240+pinges@users.noreply.github.com> Co-authored-by: Simon Dudley <simon.dudley@consensys.net> Co-authored-by: Preeti <35308865+pr9t@users.noreply.github.com> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Gabriele Del Monte <37625739+gabrieledm@users.noreply.github.com> Co-authored-by: Miroslav Kovar <miroslavkovar@protonmail.com> Co-authored-by: Reuben George <69674713+georgereuben@users.noreply.github.com> Co-authored-by: Karim Taam <karim.t2am@gmail.com> Co-authored-by: Luis Pinto <luis.pinto@consensys.net> Co-authored-by: Marcos Serradilla Diez <marcos@io.builders> Co-authored-by: garyschulte <garyschulte@gmail.com>

This means running all the test cases within one warmup or exec iteration in an attempt to invalidate CPU caches. Requires tracking exec timings in a map and calculating average at the end. Reduce warmup iteration to 20_000 and execution iterations to 1000, which appears to be as stable as 100_000. This helps reduces overall benchmark execution time (previously warmTime and execTime were prematurely reducing overall execution time). Use SequencedMap to indicate test case ordering matters. Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

This means running all the test cases within one warmup or exec iteration in an attempt to invalidate CPU caches. Requires tracking exec timings in a map and calculating average at the end. Reduce warmup iteration to 20_000 and execution iterations to 1000, which appears to be as stable as 100_000. This helps reduces overall benchmark execution time (previously warmTime and execTime were prematurely reducing overall execution time). Use SequencedMap to indicate test case ordering matters. Signed-off-by: Simon Dudley <simon.dudley@consensys.net> Signed-off-by: jflo <justin+github@florentine.us>

siladu added 11 commits July 9, 2025 20:24

Copy eip196_g1_add cases and run with benchmark subcommand

293b0f8

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Add option to invert benchmark warm up

cf90f8c

This means running all the test cases within one warm up iteration in order to invalidate CPU caches. Execution iterations are unaffected so only makes sense to use with --exec-iterations=1 Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Also invert exec iterations

4fdc24c

For each exec iteration, run through each test case in serial to attempt to bust the CPU caches. Requires tracking exec timings in a map and calculating average at the end. Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Inline executeIterations method again

b4c8a76

Extracted for reuse with warmup invert but didn't reuse in the end Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Lower default execution iterations to 1000

e12271f

Warmup iterations remain at 100_000 Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Rename warm-invert to attempt-cache-bust

8ef69da

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Use SequencedMap to indicate test case ordering matters

7a518ab

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Remove csv test cases and with it invalid test case support

73d6034

If run is invalid, throw runtime error, as before Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Reorder method

ec4b1e9

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

log about ignoring warmTime and execTime when attempt-cache-bust set

8bd9531

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Fix javadoc and tweak help test

9130042

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

siladu requested a review from ahamlat July 21, 2025 07:00

This comment was marked as resolved.

Sign in to view

Merge remote-tracking branch 'upstream/main' into attempt-cache-bust

1b25cd1

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

ahamlat reviewed Jul 25, 2025

View reviewed changes

ahamlat approved these changes Jul 25, 2025

View reviewed changes

siladu added 3 commits July 26, 2025 05:39

Merge remote-tracking branch 'upstream/main' into attempt-cache-bust

cdbd633

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

Reduce warmup iteration to 20_000 which should be sufficient

5cb7faa

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

changelog

cbebcdb

Signed-off-by: Simon Dudley <simon.dudley@consensys.net>

siladu enabled auto-merge (squash) July 25, 2025 19:56

siladu merged commit 6aef9cd into besu-eth:main Jul 25, 2025
46 checks passed

siladu deleted the attempt-cache-bust branch July 25, 2025 20:11

macfarla added the performance label Sep 22, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add --attempt-cache-bust to evmtool benchmark subcommand#8985

Add --attempt-cache-bust to evmtool benchmark subcommand#8985
siladu merged 15 commits intobesu-eth:mainfrom
siladu:attempt-cache-bust

siladu commented Jul 21, 2025 •

edited

Loading

Uh oh!

siladu commented Jul 21, 2025

Uh oh!

ahamlat commented Jul 21, 2025

Uh oh!

This comment was marked as resolved.

Uh oh!

ahamlat Jul 21, 2025

Uh oh!

siladu Jul 25, 2025

Uh oh!

Uh oh!

ahamlat left a comment

Uh oh!

siladu commented Jul 25, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

siladu commented Jul 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Testing

Uh oh!

siladu commented Jul 21, 2025

Uh oh!

ahamlat commented Jul 21, 2025

Uh oh!

This comment was marked as resolved.

Uh oh!

ahamlat Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

siladu Jul 25, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ahamlat left a comment

Choose a reason for hiding this comment

Uh oh!

siladu commented Jul 25, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

siladu commented Jul 21, 2025 •

edited

Loading