ord wallet balance is extremely slow

[lifofifoX]

Before:

$ time ./target/release/ord wallet balance
{
  "cardinal": 7315125,
  "ordinal": 676452,
  "total": 7991577
}

real    0m0.286s
user    0m0.019s
sys     0m0.033s

After:

$ time ./target/release/ord wallet --server-url http://0.0.0.0:4200 balance
{
  "cardinal": 7315125,
  "ordinal": 676452,
  "total": 7991577
}

real    0m23.050s
user    0m1.075s
sys     0m0.646s

[raphjaph]

I expected the new wallet to be a bit slower since it gets it's info over RPC but that seems way too slow.

Can you try running with ord wallet --server-url http://0.0.0.0:4200 --no-sync balance?

Also which version is the ord server that runs in the background?

I could add async or multithreading to the wallet but didn't do that because that seems like overkill.

[lifofifoX]

@raphjaph Server is running the latest master 13c0fa1. Same time with --no-sync as well.

[raphjaph]

I assume it's a pretty large wallet with many inscriptions?

I just ran it on my mainnet wallet with around 100 inscriptions and it went fast.

[lifofifoX]

@raphjaph Yup, almost 1k inscriptions.

[raphjaph]

Before the next release I'll have to work on the wallet performance. I know I'm making a lot of redundant calls.

[gmart7t2]

I made a new wallet and sent just 6 cardinals to it.

Running "ord wallet cardinals" takes 6.2 seconds to return those 6 cardinals. The /output/ endpoint is hit 18 times, 3 for each utxo in the wallet.

This is on regtest on my local machine.

This:

for output in utxos.keys() {
  self.get_output(output)?;
}

at the end of

  pub(crate) fn get_unspent_outputs(&self) -> Result<BTreeMap<OutPoint, TxOut>> {

is responsible for 2/3rds of the calls.