Andrew Toth
Andrew Toth
I got some stats for running in steady state for this branch and master. I ran 4 nodes for 11 days and 18 hours using identical aws t2-small instances backed...
@gmaxwell @l0rinc @mzumsande @sipa thank you all for your reviews! I've updated the PR to have a random window for periodic flushes of between 50 and 70 minutes. I've also...
> add a test as a very first commit, which reproduces the old behavior first? Yeah, I realize the test commit can't be cherry-picked cleanly to master. Updated to have...
Rebased to benchmark `reindex-chainstate` with #28358 and a cache size holding the entire utxo set. > the full pre-assumevalid IBD consistently takes ~45 minutes (8%) longer after the change I...
> I did a rebased -reindex-chainstate as well, this seems to be basically the same as before: Nice, so there's no speed regression but crash resistance gained! Also, I benchmarked...
> What kind of randomness are you referring to? The random window for when this PR will flush introduced in the last commit. That is good for steady state operation...
Rebased for renaming `TRACE5` -> `TRACEPOINT`.
Rebased to test performance with #30039.
> we need to decide if the 4% IBD slowdown is worth the improved consistency (or if the performance hit can be mitigated somehow). I think it's definitely worth it....
Don't we also need to add move constructor to `Coin`? Its prevector is the only thing that really benefits from being moved I believe.