perf(store): remove dead sortedStore field from VersionIndexedStore

[pdrobnjak]

Summary

• Remove the unused sortedStore *dbm.MemDB field from VersionIndexedStore
• The field was allocated in the constructor but never read or written anywhere in the codebase
• Eliminates ~4.2M dead MemDB + btree allocations per 30s benchmark window

Problem

VersionIndexedStore allocated a dbm.NewMemDB() in its constructor for a sortedStore field intended for future iterator support (per TODO comments). This field was never accessed — only declared and initialized. Each of the ~21 module stores per transaction created a VIS with this dead allocation.

Profiling after #2819 (30s, M4 Max, 1000 EVM transfers/block):

Metric	Value
NewMemDB (via VIS) alloc_space (cum)	14.9 GB
btree.NewFreeListG alloc_space	15.3 GB (mostly from VIS)
VIS instances created per 30s	~4.2M (21 stores × 200K txs)

Changes

sei-cosmos/store/multiversion/mvkv.go — 3 lines removed:

1. Field declaration: sortedStore *dbm.MemDB
2. Constructor allocation: sortedStore: dbm.NewMemDB()
3. Associated TODO comments

Benchmark Results (M4 Max, 1000 EVM transfers/block, 30s profile)

Metric	Before (after #2819)	After	Delta
TPS (steady-state range)	7,200–8,600	8,400–9,400	median ~8,000 → ~9,000
NewMemDB (via VIS) alloc	14.9 GB cum	13.5 MB cum	-14.9 GB (eliminated)
btree.NewFreeListG alloc	15.3 GB	~4.3 GB	-11 GB
memclrNoHeapPointers CPU	10.71s	5.45s	-49%
Idle CPU (usleep+kevent+pthread)	30.1s	41.6s	+38% (workers finish faster)
Total CPU samples	132.26s	124.13s	-6% (less total work needed)

pprof -alloc_space verification

NewMemDB (via VIS):  14,900 MB → 13.5 MB   (1100x reduction)
btree.NewFreeListG:  15,300 MB → 4,300 MB   (-11 GB)

CPU shift

 -5.26s  runtime.memclrNoHeapPointers  (fewer allocs to zero)
 +6.70s  runtime.usleep                (more idle — workers finish faster)
 +3.37s  runtime.kevent                (more idle)
 -1.32s  runtime.heapBitsSmallForAddr  (less heap metadata)

Test plan

• go test ./sei-cosmos/store/multiversion/... -count=1 — pass
• go test ./giga/tests/... -count=1 — pass (all 14 giga tests)
• gofmt -s -w clean
• Benchmark 2000+ blocks with heap + CPU profiling

🤖 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 6, 2026, 6:42 PM

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 56.67%. Comparing base (ae6bbe8) to head (f2970af).

Additional details and impacted files

@@                      Coverage Diff                      @@
##           perf/skip-redundant-snapshot    #2820   +/-   ##
=============================================================
  Coverage                         56.67%   56.67%           
=============================================================
  Files                              2031     2031           
  Lines                            165979   165978    -1     
=============================================================
+ Hits                              94061    94074   +13     
+ Misses                            63667    63660    -7     
+ Partials                           8251     8244    -7     

Flag	Coverage Δ
sei-chain	41.54% <ø> (+0.01%)	⬆️
sei-cosmos	48.15% <ø> (-0.02%)	⬇️
sei-db	68.72% <ø> (ø)
sei-tendermint	58.08% <ø> (ø)

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

Files with missing lines	Coverage Δ
sei-cosmos/store/multiversion/mvkv.go	90.79% <ø> (-0.06%)	⬇️

... and 11 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.