Crates.io | rdms |
lib.rs | rdms |
version | 0.0.1 |
source | src |
created_at | 2019-10-09 12:39:48.814143 |
updated_at | 2019-10-09 12:39:48.814143 |
description | Composable storage for key,value index |
homepage | https://github.com/bnclabs/rdms |
repository | https://github.com/bnclabs/rdms |
max_upload_size | |
id | 171196 |
size | 675,191 |
Given Rust's threading model, is it more efficient to en-force, using atomic primitives, that all write operations on MVCC index is deligated to single thread or is it more efficient to serialize write operations from multiple threads.
Is it enough to use Relaxed
memory-ordering for
AtomicPtr operations to manage MVCC Snapshots ?
Add a feature for using jemalloc instead of using system allocator.