# Benching `TxGraph` queries

### Methods

code: [WIP: Benchmark `TxGraph` queries #1735](https://github.com/bitcoindevkit/bdk/pull/1735)

- Collect data as of 1.0.0-beta.5 (rev: [3bc45b54](https://github.com/bitcoindevkit/bdk/commits/3bc45b5431014af3a8255a201e49227584e953b5))
- Test performance of selected tx-graph queries
- Plot performance in each case under variable workload

### Results

bench: `list_canonical_txs`

plot: Time to list canonical txs vs number of tx descendants

| n descendants | time (ms) |
| --- | --- |
| 25 | 0.589 |
| 50 | 2.3 |
| 100 | 8.8 |
| 250 | 54.5 |
| 500 | 221 |
| 1000 | 876 |


<hr>

bench: `filter_chain_unspents`

plot: Time to filter chain unspents vs number of tx-graph conflicts (single ancestor)

| n conflicts | time (ms) |
| --- | --- |
| 100 | 0.896 |
| 500 | 6.7 |
| 1000 | 18.5 |
| 2500 | 109 |
| 5000 | 326 |


<hr>

bench: `nested_conflicts`

plot: Time to list canonical txs vs total tx-graph conflicts (two ancestors)

| n conflicts | time (ms) |
| --- | --- |
| 200 | 0.979 |
| 500 | 3.07 |
| 1000 | 8.07 |
| 2500 | 37.3 |
| 5000 | 122 |
| 10000 | 447 |