[memory] Make tstate.Insert more efficient and use a sync.Pool in tokenvm.storage

[patrick-ogrady]

Allocated memory:

Showing nodes accounting for 19.24GB, 48.93% of 39.32GB total
Dropped 435 nodes (cum <= 0.20GB)
Showing top 10 nodes out of 188
      flat  flat%   sum%        cum   cum%
    2.88GB  7.33%  7.33%     2.88GB  7.33%  github.com/ava-labs/hypersdk/tstate.(*TState).Insert
    2.64GB  6.71% 14.04%     2.64GB  6.71%  github.com/ava-labs/hypersdk/examples/tokenvm/storage.PrefixBalanceKey (inline)
    2.52GB  6.40% 20.44%     2.52GB  6.40%  github.com/ava-labs/avalanchego/utils/wrappers.(*Packer).expand
    2.47GB  6.28% 26.72%     2.47GB  6.28%  github.com/ava-labs/avalanchego/x/merkledb.(*trieView).invalidateChildrenExcept
       2GB  5.09% 31.81%        2GB  5.09%  golang.org/x/exp/maps.Clone[...] (inline)
    1.42GB  3.60% 35.41%     1.42GB  3.60%  github.com/ava-labs/avalanchego/x/merkledb.newPath (inline)
    1.41GB  3.58% 39.00%     3.41GB  8.67%  github.com/ava-labs/avalanchego/x/merkledb.(*node).clone
    1.33GB  3.39% 42.39%     3.51GB  8.93%  github.com/ava-labs/avalanchego/x/merkledb.(*trieView).calculateNodeIDsHelper
    1.29GB  3.28% 45.67%     1.29GB  3.28%  google.golang.org/grpc.(*parser).recvMsg
    1.28GB  3.27% 48.93%     3.07GB  7.82%  github.com/ava-labs/hypersdk/chain.(*Processor).Prefetch.func1

[patrick-ogrady]

Good example of a byte pool: https://github.com/ava-labs/avalanchego/blob/7d73b59cb4838d304387ea680b9cc4053b72620c/database/prefixdb/db.go#L32

[github-actions]

This issue has become stale because it has been open 60 days with no activity. Adding the lifecycle/frozen label will exempt this issue from future lifecycle events.