Closed torao closed 2 years ago
I made this title a separate issue because the approach for performance was different from that of Caching and Prefetching, but the only relevant code is only the following: https://github.com/line/lbm-sdk/blob/2e12c78c0aa6f9a3f38407a4020fe331af5f367a/store/cachemulti/store.go#L141-L161
Also, the origin of this change has been added with the commit for Prefetching: https://github.com/line/lbm-sdk/pull/423/commits/c3e3f47972593353cc3f072cbebdc0306f05a02d#diff-3726a67741e4ddab06683dd6e86fe5f23be2237c6c766124e32db098af8e9e74R114-R122
The work is ineffective to investigate only this small issue, and also it seems unlikely to cause conflicts with the cosmos-sdk. For this reason, we'll delegate this issue to #582, and if we revert Prefetching, we'll revert this fix as well.
Summary
Reconsideration of the effect of writing by multi-thread.
This task is to sort out useful features in order to reduce the differences between lbm-sdk and cosmos-sdk.
Parent: #549
For Admin Use