Add EVM execution microbenchmarks via go test -bench

[pdrobnjak]

Summary

• Widen SetupWithSc, NewTestWrapperWithSc, NewGigaTestWrapper, and NewGigaTestWrapperWithRegularStore to accept testing.TB instead of *testing.T, enabling benchmarks to use the SC code path
• Add table-driven EVM microbenchmarks in occ_tests/evm_benchmark_test.go covering non-conflicting, conflicting, and mixed transfer workloads at 100-5000 tx scales
• Each benchmark has OCC and sequential sub-variants and reports txns/op, tps, gas/op, gas/sec, ns/tx

Initial Results (benchtime=3s)

Non-Conflicting EVM Transfers

Txs	Mode	ns/op	TPS	gas/sec
100	occ	14.7M	6,820	143M
100	sequential	39.9M	2,509	52.7M
500	occ	92.0M	5,433	114M
500	sequential	224.7M	2,225	46.7M
1000	occ	388.5M	2,574	54.1M
1000	sequential	525.1M	1,904	40.0M
5000	occ	6.72B	744	15.6M
5000	sequential	4.82B	1,037	21.8M

Conflicting EVM Transfers

Txs	Mode	ns/op	TPS	gas/sec
100	occ	91.2M	1,096	23.0M
100	sequential	41.1M	2,435	51.1M
1000	occ	3.46B	289	6.1M
1000	sequential	519.3M	1,926	40.4M

Mixed (50/50 conflicting + non-conflicting)

Txs	Mode	ns/op	TPS	gas/sec
1000	occ	3.96B	253	5.3M
1000	sequential	510.0M	1,961	41.2M

Key Observations

• OCC provides ~2.5-2.7x speedup for non-conflicting txs at smaller batch sizes (100-500)
• OCC degrades with conflicts due to retry overhead (4-7x slower than sequential)
• At 5000 txs, even non-conflicting OCC is slower than sequential (memory/contention overhead)
• Sequential is consistently ~1,900-2,500 TPS regardless of conflict patterns

Usage

# Run all EVM benchmarks (~30s-2min depending on scale)
go test -bench=BenchmarkEVM -benchtime=3s -count=1 ./occ_tests/

# Run specific scale
go test -bench=BenchmarkEVMTransfer/txs_1000 -benchtime=5s -count=3 ./occ_tests/

# Compare before/after with benchstat
go test -bench=BenchmarkEVM -benchtime=5s -count=5 ./occ_tests/ > old.txt
# ... make changes ...
go test -bench=BenchmarkEVM -benchtime=5s -count=5 ./occ_tests/ > new.txt
benchstat old.txt new.txt

Test plan

• go test -bench=BenchmarkEVM -benchtime=1x -count=1 ./occ_tests/ — all benchmarks run without error
• go test ./occ_tests/ -run TestParallelTransactionsEvmTransfer — existing OCC tests still pass
• go vet ./occ_tests/ — no errors
• gofmt -s -l app/test_helpers.go occ_tests/evm_benchmark_test.go — no output

🤖 Generated with Claude Code

[github-actions]

The latest Buf updates on your PR. Results from workflow Buf / buf (pull_request).

Build	Format	Lint	Breaking	Updated (UTC)
✅ passed	✅ passed	✅ passed	✅ passed	Feb 17, 2026, 1:32 PM

[codecov]

Codecov Report

❌ Patch coverage is 75.00000% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 57.25%. Comparing base (7fb1494) to head (d54ca37).
⚠️ Report is 8 commits behind head on main.

Files with missing lines	Patch %	Lines
app/test_helpers.go	75.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2901      +/-   ##
==========================================
+ Coverage   57.16%   57.25%   +0.09%     
==========================================
  Files        2091     2093       +2     
  Lines      171179   171729     +550     
==========================================
+ Hits        97853    98324     +471     
- Misses      64616    64640      +24     
- Partials     8710     8765      +55     

Flag	Coverage Δ
sei-chain	52.72% <75.00%> (+0.09%)	⬆️
sei-cosmos	48.19% <ø> (+0.05%)	⬆️
sei-db	68.94% <ø> (+0.22%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
app/test_helpers.go	40.16% <75.00%> (-0.04%)	⬇️

... and 76 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.